Forecaster's Toolbox: Gaming and Strategy

Click here to return to the Education page.

Click here to return to the Forecaster's Toolbox home page.

General Concepts | Batting | Pitching | Minor League Toolbox | Strategy Toolbox | Team Toolbox

 

Gaming

Standard Rules and Variations

Rotisserie Baseball was invented as an elegant confluence of baseball and economics. Whether by design or accident, the result has lasted for more than three decades. But what would Rotisserie and fantasy have been like if the Founding Fathers knew then what we know now about statistical analysis and game design? You can be sure things would be different.

The world has changed since the original game was introduced yet many leagues use the same rules today. New technologies have opened up opportunities to improve elements of the game that might have been limited by the capabilities of the 1980s. New analytical approaches have revealed areas where the original game falls short.

As such, there are good reasons to tinker and experiment; to find ways to enhance the experience.

Following are the basic elements of fantasy competition, those that provide opportunities for alternative rules and experimentation. This is by no means an exhaustive list, but at minimum provides some interesting food-for-thought. 

Player pool

Standard: American League-only, National League-only or Mixed League. 

AL/NL-only typically drafts 8-12 teams (pool penetration of 49% to 74%). Mixed leagues draft 10-18 teams (31% to 55% penetration), though 15 teams (46%) is a common number.

Drafting of reserve players will increase the penetration percentages. A 12-team AL/NL-only league adding six reserves onto 23-man rosters would draft 93% of the available pool of players on all teams’ 25-man rosters.

The draft penetration level determines which fantasy management skills are most important to your league. The higher the penetration, the more important it is to draft a good team. The lower the penetration, the greater the availability of free agents and the more important in-season roster management becomes. 

There is no generally-accepted optimal penetration level, but we have often suggested that 75% (including reserves) provides a good balance between the skills required for both draft prep and in-season management.

Alternative pools: There is a wide variety of options here. Certain leagues draft from within a small group of major league divisions or teams. Some competitions, like home run leagues, only draft batters.

Bottom-tier pool: Drafting from the entire major league population, the only players available are those who posted a Rotisserie dollar value of $5 or less in the previous season. Intended as a test of an owner’s ability to identify talent with upside. Best used as a pick-a-player contest with any number of teams participating.

Positional structure

Standard: 23 players. One at each defensive position (though three outfielders may be from any of LF, CF or RF), plus one additional catcher, one middle infielder (2B or SS), one corner infielder (1B or 3B), two additional outfielders and a utility player/designated hitter (which often can be a batter who qualifies anywhere). Nine pitchers, typically holding any starting or relief role.

Open: 25 players. One at each defensive position (plus DH), 5-man starting rotation and two relief pitchers. Nine additional players at any position, which may be a part of the active roster or constitute a reserve list.

40-man: Standard 23 plus 17 reserves. Used in many keeper and dynasty leagues.

Reapportioned: In recent years, new obstacles are being faced by 12-team AL/NL-only leagues thanks to changes in the real game. The 14/9 split between batters and pitchers no longer reflects how MLB teams structure their rosters. Of the 30 teams, each with 25-man rosters, not one contains 14 batters for any length of time. In fact, many spend a good part of the season with only 12 batters, which means teams often have more pitchers than hitters.

For fantasy purposes in AL/NL-only leagues, that leaves a disproportionate draft penetration into the batter and pitcher pools:

                       BATTERS    PITCHERS
On all MLB rosters      195         180
Players drafted         168         108
Pct.                    86%         60%

These drafts are depleting 26% more batters out of the pool than pitchers. Add in those leagues with reserve lists—perhaps an additional six players per team removing another 72 players —and post-draft free agent pools are very thin, especially on the batting side.

The impact is less in 15-team mixed leagues, though the FA pitching pool is still disproportionately deep.

                 BATTERS     PITCHERS
On all rosters     381          369
Drafted            210          135
Pct.               55%          37%

 One solution is to reapportion the number of batters and pitchers that are rostered. Adding one pitcher slot and eliminating one batter slot may be enough to provide better balance. The batting slot most often removed is the second catcher, since it is the position with the least depth.

Beginning in the 2012 season, the Tout Wars AL/NL-only experts leagues opted to eliminate one of the outfield slots and replace it with a “swingman” position. This position could be any batter or pitcher, depending upon the owner’s needs at any given time during the season. 

Selecting players 

Standard: The three most prevalent methods for stocking fantasy rosters are:

Snake/Straight/Serpentine draft: Players are selected in order with seeds reversed in alternating rounds. This method has become the most popular due to its speed, ease of implementation and ease of automation.

In these drafts, the underlying assumption is that value can be ranked relative to a linear baseline. Pick #1 is better than pick #2, which is better than pick #3, and the difference between each pick is assumed to be somewhat equivalent. While a faulty assumption, we must believe in it to assume a level playing field.

Auction: Players are sold to the highest bidder from a fixed budget, typically $260. Auctions provide the team owner with more control over which players will be on his team, but can take twice as long as snake drafts.

The baseline is $0 at the beginning of each player put up for bid. The final purchase price for each player is shaped by many wildly variable factors, from roster need to geographic location of the draft. A $30 player can mean different things to different drafters. 

One option that can help reduce the time commitment of auctions is to force minimum bids at each hour mark. You could mandate $15 openers in hour #1; $10 openers in hour #2, etc.

Pick-a-player / Salary cap: Players are assigned fixed dollar values and owners assemble their roster within a fixed cap. This type of roster-stocking is an individual exercise which results in teams typically having some of the same players.

In these leagues, the “value” decision is taken out of the hands of the owners. Each player has a fixed value, pre-assigned based on past season performance and/or future expectation. 

Hybrid snake-auction: Each draft begins as an auction. Each team has to fill its first seven roster slots from a budget of $154. Opening bid for any player is $15. After each team has filled seven slots, it becomes a snake draft.

This method is intended to reduce draft time while still providing an economic component for selecting players.

Stat categories

Standard: The standard statistical categories for Rotisserie leagues are:

4x4:  HR, RBI, SB, BA, W, Sv, ERA, WHIP

5x5:  HR, R, RBI, SB, BA, W, Sv, K, ERA, WHIP

6x6: Categories typically added are Holds and OPS.

7x7, etc.: Any number of categories may be added.

In general, the more categories you add, the more complicated it is to isolate individual performance and manage the categorical impact on your roster. There is also the danger of redundancy; with multiple categories measuring like stats, certain skills can get over-valued. For instance, home runs are double-counted when using the categories of both HR and slugging average. (Though note that HRs are actually already triple-counted in standard 5x5—HRs, runs, and RBIs)

If the goal is to have categories that create a more encompassing picture of player performance, it is actually possible to accomplish more with less:

Modified 4x4: HR, (R+RBI-HR), SB, OBA, (W+QS), (Sv+Hld), K, ERA

This provides a better balance between batting and pitching in that each has three counting categories and one ratio category. In fact, the balance is shown to be even more notable here:

                                      BATTING       PITCHING
Pure skill counting stat                HR             K
Ratio category                         OBA            ERA
Dependent upon managerial decision      SB        (Sv+Hold)
Dependent upon team support         (R+RBI-HR)      (W+QS)

Replacing saves: The problem with the Saves statistic is that we have a scarce commodity that is centered on a small group of players, thereby creating inflated demand for those players. With the rising failure rate for closers these days, the incentive to pay full value for the commodity decreases. The higher the risk, the lower the prices. 

We can increase the value of the commodity by reducing the risk. We might do this by increasing the number of players that contribute to that category, thereby spreading the risk around. One way we can accomplish this is by changing the category to Saves + Holds.

Holds are not perfect, but the typical argument about them being random and arbitrary can apply to saves these days as well. In fact, many of the pitchers who record holds are far more skilled and valuable than closers; they are often called to the mound in much higher leverage situations (a fact backed up by a scan of each pitcher’s Leverage Index).

Neither stat is perfect, but together they form a reasonable proxy for overall bullpen performance. 

In tandem, they effectively double the player pool of draftable relievers while also flattening the values allotted to those pitchers. The more players around which we spread the risk, the more control we have in managing our pitching staffs.

Replacing wins: Using reasons similar to replacing Saves with Saves + Holds, some have argued for replacing the Wins statistic with W + QS (quality starts). This method of scoring gives value to a starting pitcher who pitches well, but fails to receive the win due to his team’s poor offense or poor luck.

Keeping score

Standard:  These are the most common scoring methods:

Rotisserie: Players are evaluated in several statistical categories. Totals of these statistics are ranked by team. The winner is the team with the highest cumulative ranking.

Points: Players receive points for events that they contribute to in each game. Points are totaled for each team and teams are then ranked.

Head-to-Head (H2H): Using Rotisserie or points scoring, teams are scheduled in daily or weekly matchups. The winner of each matchup is the team that finishes higher in more categories (Rotisserie) or scores the most points. 

Hybrid H2H-Rotisserie: Rotisserie’s category ranking system can be converted into a weekly won-loss record. Depending upon where your team finishes for that week’s statistics determines how many games you win for that week.  Each week, your team will play seven games.

    *Place      Record    *Place      Record
    1st          7-0        7th        3-4
    2nd          6-1        8th        2-5
    3rd          6-1        9th        2-5
    4th          5-2       10th        1-6
    5th          5-2       11th        1-6
    6th          4-3       12th        0-7

 * Based on overall Rotisserie category ranking for the week.

At the end of each week, all the statistics revert to zero and you start over. You never dig a hole in any category that you can’t climb out of, because all categories themselves are incidental to the standings.

The regular season lasts for 23 weeks, which equals 161 games. Weeks 24, 25 and 26 are for play-offs.

Free agent acquisition

Standard: Three methods are the most common for acquiring free agent players during the season. 

First to the phone: Free agents are awarded to the first owner who claims them.

Reverse order of standings: Access to the free agent pool is typically in a snake draft fashion with the last place team getting the first pick, and each successive team higher in the standings picking afterwards. 

Free agent acquisition budget (FAAB): Teams are given a set budget at the beginning of the season (typically, $100 or $1000) from which they bid on free agents in a closed auction process.

Vickrey FAAB: Research has shown that more than 50% of FAAB dollars are lost via overbid on an annual basis. Given that this is a scarce commodity, one would think that a system to better manage these dollars might be desirable. The Vickrey system conducts a closed auction in the same way as standard FAAB, but the price of the winning bid is set at the amount of the second highest bid, plus $1. In some cases, gross overbids (at least $10 over) are reduced to the second highest bid plus $5. 

This method was designed by William Vickrey, a Professor of Economics at Columbia University. His theory was that this process reveals the true value of the commodity. For his work, Vickrey was awarded the Nobel Prize for Economics (and $1.2 million) in 1996. 

Double-Bid FAAB: One of the inherent difficulties in the current FAAB system is that we have so many options for setting a bid amount. You can bid $47, or $51, or $23. You might agonize over whether to go $38 or $39. With a $100 budget, there are 100 decision points. And while you may come up with a rough guesstimate of the range in which your opponents might bid, the results for any individual player bidding are typically random within that range.

The first part of this process reduces the number of decision points. Owners must categorize their interest by bidding a fixed number of pre-set dollar amounts for each player. In a $100 FAAB league, for instance, those levels might be $1, $5, $10, $15, $20, $30, $40 or $50. All owners would set the general market value for free agents in these eight levels of interest. (This system sets a $50 maximum, but that is not absolutely necessary.)

The initial stage of the bidding process serves to screen out those who are not interested in a player at the appropriate market level. That leaves a high potential for tied owners, those who share the same level of interest.

The tied owners must then submit a second bid of equal or greater value than their first bid. These bids can be in $1 increments. The winning owner gets the player; if there is still a tie, then the player would go to the owner lower in the standings.

An advantage of this second bid is that it gives owners an opportunity to see who they are going up against, and adjust. If you are bidding against an owner close to you in the standings, you may need to be more aggressive in that second bid. If you see that the tied owner(s) wouldn’t hurt you by acquiring that player, then maybe you resubmit the original bid and be content to potentially lose out on the player. If you’re ahead in the standings, it’s actually a way to potentially opt out on that player completely by resubmitting your original bid and forcing another owner to spend his FAAB.

Some leagues will balk at adding another layer to the weekly deadline process; it’s a trade-off to having more control over managing your FAAB.

The season

Standard: Leagues are played out during the course of the entire Major League Baseball season.

Split-season: Leagues are conducted from Opening Day through the All-Star break, then re-drafted to play from the All-Star break through the end of the season.

50-game split-season: Leagues are divided into three 50-game seasons with one-week break in between.

Monthly: Leagues are divided into six seasons or rolling four-week seasons.

The advantages of these shorter time frames:

•    Shorter time frames can help to maintain interest. There would be fewer abandoned teams.

•    There would be more shots at a title each year.

•    Given that drafting is considered the most fun aspect of the game, these splits multiply the opportunities to participate in some type of draft. Leagues may choose to do complete re-drafts and treat the year as distinct mini-seasons. Or, leagues might allow teams to drop their five worst players and conduct a restocking draft at each break.

Daily games:  Participants select a roster of players from one day’s MLB schedule. Scoring is based on an aggregate points-based system rather than categories, with cash prizes awarded based on the day’s results. The structure and distribution of that prize pool varies across different types of events, and those differences can affect roster construction strategies. Although scoring and prizes are based on one day’s play, the season-long element of bankroll management provides a proxy for overall standings. 

In terms of projecting outcomes, daily games are drastically different than full-season leagues. Playing time is one key element of any projection, and daily games offer near-100% accuracy in projecting playing time: you can check pre-game lineups to see exactly which players are in the lineup that night. The other key component of any projection is performance, but that is plagued by variance in daily competitions. Even if you roster a team full of the most advantageous matchups (for instance, Mike Trout facing Franklin Morales at Coors Field), Trout will sometimes go 0-for-4 on that one night.

Single game (Quint-Inning): A game that drafts from the active rosters of two major league teams in a single game. The rules: 

1. Start with five owners.

2. Prior to first pitch, conduct a simple snake draft where each owner selects five players. If you’re ambitious, auction off the 25 players giving each owner a budget of $50 of real or fake money.

3. Scoring is simple. For batters, singles, walks, hit-by-pitches and stolen bases are one point each. Doubles are 2 points. Triples are 3 points. Home runs are 4 points. Pitchers get one point for each complete inning pitched but lose one point for every run they allow. 

4. At the beginning of the 5th inning, each owner has the option of doubling any future points for one player on his roster. We call that player the Quint. Points for all batters are doubled beginning in the 9th inning. That means the Quint’s points would be quadrupled.

5. At the end of each inning, you can cut players, claim players from the free agent pool or trade players. You must maintain five players at all times, so all adds, drops and trades must keep your roster square. Free agent claims are done in reverse order of the standings. If two teams are tied and both want the same player, it can be helpful to have a deck of cards handy - the owner who draws high card would get the player.

6. Quint-Inning is a betting game (which makes it technically illegal). Owners need to ante up to play, typically $5, though if you’re using a $50 auction budget, that works fine. It then costs $1 per inning to stay in the game for the second through fourth innings. Beginning in the 5th inning, the stakes increase to $2 per inning to stay in the game. You can use higher or lower stakes if you prefer.

7. Owners can fold at any time, forfeiting any monies they contributed to the pot. Their players are released into the free agent pool and are available to the remaining owners in reverse order of the standings.

8. The owner with the most points at the end of the game wins the pot.

Post-season league: Some leagues re-draft teams from among the MLB post-season contenders and play out a separate competition. It is possible, however, to make a post-season competition that is an extension of the regular season.

Start by designating a set number of regular season finishers as qualifying for the post-season. The top four teams in a league is a good number. 

These four teams would designate a fixed 23-man roster for all post-season games. First, they would freeze all of their currently-owned players who are on MLB post-season teams. 

In order to fill the roster holes that will likely exist, these four teams would then pick players from their league’s non-playoff teams (for the sake of the post-season only). This would be in the form of a snake draft done on the day following the end of the regular season. Draft order would be regular season finish, so the play-off team with the most regular season points would get first pick. Picks would continue until all four rosters are filled with 23 men.

Regular scoring would be used for all games during October. The team with the best play-off stats at the end of the World Series is the overall champ.

Snake Drafting

Snake draft first round history

The following tables record the comparison between pre-season projected player rankings (using Average Draft Position data from Mock Draft Central and National Fantasy Baseball Championship) and actual end-of-season results. The 12-year success rate of identifying each season’s top talent is only 34%, and getting worse. During the first six years of this period, our success rate was 38%; in the six years since, it has dropped to 31%.

2008      ADP                        ACTUAL = 7    
1         Alex Rodriguez        1      Albert Pujols (10)
2         Hanley Ramirez        2      Jose Reyes (4)
3         David Wright          3      Hanley Ramirez (2)
4         Jose Reyes            4      Manny Ramirez
5         Matt Holliday         5      Matt Holliday (5)
6         Jimmy Rollins         6      David Wright (3)
7         Miguel Cabrera        7      Lance Berkman
8         Chase Utley           8      Dustin Pedroia
9         Ryan Howard           9      Roy Halladay
10        Albert Pujols        10      Josh Hamilton
11        Prince Fielder       11      Alex Rodriguez (1)
12        Ryan Braun           12      C.C. Sabathia
13        Johan Santana        13      Carlos Beltran
14        Carl Crawford        14      Grady Sizemore
15        Alfonso Soriano      15      Chase Utley (8)

2009      ADP                         ACTUAL = 5    
1         Hanley Ramirez        1      Albert Pujols (2)
2         Albert Pujols         2      Hanley Ramirez (1)
3         Jose Reyes            3      Tim Lincecum
4         David Wright          4      Dan Haren
5         Grady Sizemore        5      Carl Crawford
6         Miguel Cabrera        6      Matt Kemp
7         Ryan Braun            7      Joe Mauer
8         Jimmy Rollins         8      Derek Jeter
9         Ian Kinsler           9      Zach Greinke
10        Josh Hamilton        10      Ryan Braun (7)
11        Ryan Howard          11      Jacoby Ellsbury
12        Mark Teixeira        12      Mark Reynolds
13        Alex Rodriguez       13      Prince Fielder
14        Matt Holliday        14      Chase Utley (15)
15        Chase Utley          15      Miguel Cabrera (6)

2010      ADP                     ACTUAL = 5    
1         Albert Pujols         1   Carlos Gonzalez
2         Hanley Ramirez        2   Albert Pujols (1)
3         Alex Rodriguez        3   Joey Votto
4         Chase Utley           4   Roy Halladay
5         Ryan Braun            5   Carl Crawford (15)
6         Mark Teixeira         6   Miguel Cabrera (9)
7         Matt Kemp             7   Josh Hamilton
8         Prince Fielder        8   Adam Wainwright
9         Miguel Cabrera        9   Felix Hernandez
10        Ryan Howard          10   Robinson Cano
11        Evan Longoria        11   Jose Bautista
12        Tom Lincecum         12   Paul Konerko
13        Joe Mauer            13   Matt Holliday
14        David Wright         14   Ryan Braun (5)
15        Carl Crawford        15   Hanley Ramirez (2)

2011     ADP                   ACTUAL = 6    
1         Albert Pujols         1      Matt Kemp
2         Hanley Ramirez        2      Jacoby Ellsbury
3         Miguel Cabrera        3      Ryan Braun (10)
4         Troy Tulowitzki       4      Justin Verlander
5         Evan Longoria         5      Clayton Kershaw
6         Carlos Gonzalez       6      Curtis Granderson
7         Joey Votto            7      Adrian Gonzalez (8)
8         Adrian Gonzalez       8      Miguel Cabrera (3)
9         Robinson Cano         9      Roy Halladay (15)
10        Ryan Braun           10      Cliff Lee
11        David Wright         11      Jose Bautista 
12        Mark Teixeira        12      Dustin Pedroia
13        Carl Crawford        13      Jered Weaver
14        Josh Hamilton        14      Albert Pujols (1)
15        Roy Halladay         15      Robinson Cano (9)

2012     ADP          ACTUAL = 4    
1    Matt Kemp                 1    Mike Trout
2    Ryan Braun                2    Ryan Braun (2)
3    Albert Pujols             3    Miguel Cabrera (4)
4    Miguel Cabrera            4    Andrew McCutchen
5    Troy Tulowitzki           5    R.A. Dickey
6    Jose Bautista             6    Clayton Kershaw
7    Jacoby Ellsbury           7    Justin Verlander (8)
8    Justin Verlander          8    Josh Hamilton
9    Adrian Gonzalez           9    Fernando Rodney
10    Justin Upton            10    Adrian Beltre
11    Robinson Cano           11    Alex Rios
12    Joey Votto              12    David Price
13    Evan Longoria           13    Chase Headley
14    Carlos Gonzalez         14    Robinson Cano (11)
15    Prince Fielder          15    Edwin Encarnacion

2013     ADP    ACTUAL = 5    
1    Ryan Braun               1    Miguel Cabrera (2)
2    Miguel Cabrera           2    Mike Trout (3)
3    Mike Trout               3    Clayton Kershaw (15)
4    Matt Kemp                4    Chris Davis
5    Andrew McCutchen         5    Paul Goldschmidt
6    Albert Pujols            6    Andrew McCutchen (5)
7    Robinson Cano            7    Adam Jones
8    Jose Bautista            8    Jacoby Ellsbury
9    Joey Votto               9    Max Scherzer
10    Carlos Gonzalez        10    Carlos Gomez
11    Buster Posey           11    Hunter Pence
12    Justin Upton           12    Robinson Cano (7)
13    Giancarlo Stanton      13    Alex Rios
14    Prince Fielder         14    Adrian Beltre
15    Clayton Kershaw        15    Matt Harvey

2014     ADP    ACTUAL = 4    
1    Mike Trout               1    Jose Altuve
2    Miguel Cabrera           2    Clayton Kershaw (6)
3    Paul Goldschmidt         3    Michael Brantley
4    Andrew McCutchen         4    Mike Trout (1)
5    Carlos Gonzalez          5    Johnny Cueto
6    Clayton Kershaw          6    Felix Hernandez
7    Chris Davis              7    Victor Martinez
8    Ryan Braun               8    Jose Abreu
9    Adam Jones               9    Giancarlo Stanton
10    Bryce Harper           10    Andrew McCutchen (4)
11    Robinson Cano          11    Miguel Cabrera (2)
12    Hanley Ramirez         12    Carlos Gomez
13    Jacoby Ellsbury        13    Jose Bautista
14    Prince Fielder         14    Dee Gordon
15    Troy Tulowitzki        15    Anthony Rendon

2015     ADP    ACTUAL = 4    

1    Mike Trout               1    Jake Arrieta
2    Andrew McCutchen         2    Zack Greinke
3    Clayton Kershaw          3    Clayton Kershaw (3)
4    Giancarlo Stanton        4    Paul Goldschmidt (5)
5    Paul Goldschmidt         5    A.J. Pollock
6    Miguel Cabrera           6    Dee Gordon
7    Jose Abreu               7    Bryce Harper
8    Carlos Gomez             8    Josh Donaldson
9    Jose Batista             9    Jose Altuve (12)
10    Edwin Encarnacion      10    Mike Trout (1)
11    Felix Hernandez        11    Nolan Arenado
12    Jose Altuve            12    Manny Machado
13    Anthony Rizzo          13    Dallas Keuchel
14    Adam Jones             14    Max Scherzer
15    Troy Tulowitzki        15    Nelson Cruz

This is the first time since we’ve been keeping ADP records (2004) that a pitcher has finished ranked #1. It is also the first time that pitchers have occupied the top three spots in the rankings. It is the third time that five pitchers made it into the Top 15.

ADP attrition

Why is our success rate so low in identifying what should be the most easy-to-project players each year? We rank and draft players based on the expectation that those ranked higher will return greater value in terms of productivity and playing time, as well as being the safest investments. However, there are many variables affecting where players finish.

Earlier, it was shown that players spend an inordinate number of days on the disabled list. In fact, of the players projected to finish in the top 300, the number who lost playing time due to injuries, demotions and suspensions has been extreme: 

Year    Pct. of top-ranked 300 players who lost PT
2009    51%
2010    44%
2011    49%
2012    45%
2013    51%
2014    53%
2015    47%

When you consider that about half of each season’s very best players had fewer at-bats or innings pitched than we projected, it shows how tough it is to rank players each year. 

The fallout? Consider: It is nearly a foregone conclusion that players like A.J. Pollock and Manny Machado—players who finished in the top 15 for the first time last year—will rank as first round picks in 2016. The above data provide a strong argument against them returning first-round value.

Yes, they are excellent players, two of the best in the game, in 2015 anyway. But the issue is not their skills profile. The issue is the profile of what makes a worthy first rounder. 

Since 2004:

•    Two-thirds of players finishing in the Top 15 were not in the Top 15 the previous year. There is a great deal of turnover in the first round, year-to-year.

•    Of those who were first-timers, only 14% repeated in the first round the following year. 

•    Established superstars who finished in the Top 15 were no guarantee to repeat.

As such, the odds are against Pollock and Machado repeating in the first round, as counter-intuitive as it may seem. In past years, sudden stars like Hunter Pence, Curtis Granderson and Dustin Pedroia have failed to repeat. As talented as these players are, it’s not just about skill; it’s also about skill relative to the rest of a volatile player pool.

Importance of the Early Rounds  (Bill Macey)

It’s long been said that you can’t win your league in the first round, but you can lose it there. An analysis of data from actual drafts reveals that this holds true—those who spend an early round pick on a player that severely under-performs expectations rarely win their league and seldom even finish in the top 3.

At the same time, drafting a player in the first round that actually returns first-round value is no guarantee of success. In fact, those that draft some of the best values still only win their league about a quarter of the time and finish in the top 3 less than half the time. Research also shows that drafting pitchers in the first round is a risky proposition. Even if the pitchers deliver first-round value, the opportunity cost of passing up on an elite batter makes you less likely to win your league. 

What is the best seed to draft from?

Most drafters like mid-round so they never have to wait too long for their next player. Some like the swing pick, suggesting that getting two players at 15 and 16 is better than a 1 and a 30. Many drafters assume that the swing pick means you’d be getting something like two $30 players instead of a $40 and $20.  

Equivalent auction dollar values reveal the following facts about the first two snake draft rounds: 

In an AL/NL-only league, the top seed would get a $44 player (at #1) and a $24 player (at #24) for a total of $68; the 12th seed would get two $29s (at #12 and #13) for $58. 

In a mixed league, the top seed would get a $47 and a $24 ($71); the 15th seed would get two $28s ($56). 

Since the talent level flattens out after the 2nd round, low seeds never get a chance to catch up: 

                         Dollar value difference between
                        first player selected and last player selected
    Round        12-team      15-team
     1            $15           $19
     2             $7            $8
     3             $5            $4
     4             $3            $3
     5             $2            $2
     6             $2            $1
    7-17           $1            $1
   18-23           $0            $0

The total value each seed accumulates at the end of the draft is hardly equitable: 

    Seed         Mixed      AL/NL-only
    1            $266          $273
    2            $264          $269
    3            $263          $261
    4            $262          $262
    5            $259          $260
    6            $261          $260
    7            $260          $260
    8            $261          $260
    9            $261          $258
    10           $257          $260
    11           $257          $257
    12           $258          $257
    13           $254   
    14           $255   
    15           $256    

Of course, the draft is just the starting point for managing your roster and player values are variable. Still, it’s tough to imagine a scenario where the top seed wouldn’t have an advantage over the bottom seed. 

Using ADPs to determine when to select players  (Bill Macey) 

Although average draft position (ADP) data provides a good idea of where in the draft each player is selected, it can be misleading when trying to determine how early to target a player. This chart summarizes the percentage of players drafted within 15 picks of his ADP as well as the average standard deviation by grouping of players. 

    ADP      % within        Standard 
    Rank      15 picks       Deviation 
    1-25       100%            2.5 
    26-50       97%            6.1 
    51-100      87%            9.6 
    100-150     72%           14.0 
    150-200     61%           17.4 
    200-250     53%           20.9 

As the draft progresses, the picks for each player become more widely dispersed and less clustered around the average. Most top 100 players will go within one round of their ADP-converted round. However, as you reach the mid-to-late rounds, there is much more uncertainty as to when a player will be selected. Pitchers have slightly smaller standard deviations than do batters (i.e. they tend to be drafted in a narrower range). This suggests that drafters may be more likely to reach for a batter than for a pitcher. 

Using the ADP and corresponding standard deviation, we can to estimate the likelihood that a given player will be available at a certain draft pick. We estimate the predicted standard deviation for each player as follows: 

Stdev = -0.42 + 0.42*(ADP - Earliest Pick) 

(That the figure 0.42 appears twice is pure coincidence; the numbers are not equal past two decimal points.) 

If we assume that the picks are normally distributed, we can use a player’s ADP and estimated standard deviation to estimate the likelihood that the player is available with a certain pick (MS Excel formula): 

=1-normdist(x,ADP,Standard Deviation,True) 

where «x» represents the pick number to be evaluated. 

We can use this information to prepare for a snake draft by determining how early we may need to reach in order to roster a player. Suppose you have the 8th pick in a 15-team league draft and your target is 2009 sleeper candidate Nelson Cruz. His ADP is 128.9 and his earliest selection was with the 94th pick. This yields an estimated standard deviation of 14.2. You can then enter these values into the formula above to estimate the likelihood that he is still available at each of the following picks: 

            Likelihood 
    Pick    Available 
    83        100% 
    98         99% 
    113        87% 
    128        53% 
    143        16% 
    158         2% 

ADPs and scarcity (Bill Macey) 

Most players are selected within a round or two of their ADP with tight clustering around the average. But every draft is unique and every pick in the draft seemingly affects the ordering of subsequent picks. In fact, deviations from “expected” sequences can sometimes start a chain reaction at that position. This is most often seen in runs at scarce positions such as the closer; once the first one goes, the next seems sure to closely follow. 

Research also suggests that within each position, there is a correlation within tiers of players. The sooner players within a generally accepted tier are selected, the sooner other players within the same tier will be taken. However, once that tier is exhausted, draft order reverts to normal. 

How can we use this information? If you notice a reach pick, you can expect that other drafters may follow suit. If your draft plan is to get a similar player within that tier, you’ll need to adjust your picks accordingly. 

Mapping ADPs to auction value (Bill Macey) 

Reliable average auction values (AAV) are often tougher to come by than ADP data for snake drafts. However, we can estimate predicted auction prices as a function of ADP, arriving at the following equation: 

y = -9.8ln(x) + 57.8
where ln(x) is the natural log function, x represents the actual ADP, and y represents the predicted AAV. 

This equation does an excellent job estimating auction prices (r2=0.93), though deviations are unavoidable. The asymptotic nature of the logarithmic function, however, causes the model to predict overly high prices for the top players. So be aware of that, and adjust.

The value of mock drafts (Todd Zola)

Most assume the purpose of a mock draft is to get to know the market value of the player pool. But even more important, mock drafting is general preparation for the environment and process,  thereby allowing the drafter to completely focus on the draft when it counts. Mock drafting is more about fine-tuning your strategy than player value. Here are some tips to maximize your mock drafting experience.

1. Make sure you can seamlessly use an on-line drafting room, draft software or your own lists to track your draft or auction.The less time you spend looking, adding and adjusting names, the more time you can spend on thinking about what player is best for your team. This also gives you the opportunity to make sure your draft lists are complete, and assures all the players are listed at the correct position(s).

2. Alter the positions from which you mock. The flow of each mock will be different, but if you do a few mocks with an early initial pick, a few in the middle and a few with a late first pick, you may learn you prefer one of the spots more than the others. If you’re in a league where you can choose your draft spot, this helps you decide where to select. Once you know your spot, a few mocks from that spot will help you decide how to deal with positional runs.

3. Use non-typical strategies and consider players you rarely target. We all have our favorite players. Intentionally passing on those players not only gives you an idea when others may draft them but it also forces you to research players you normally don’t consider. The more players you have researched, the more prepared you’ll be for any series of events that occurs during your real draft.

Auction Value Analysis

Auction values (R$) in perspective

R$ is the dollar value placed on a player’s statistical performance in a Rotisserie league, and designed to measure the impact that player has on the standings.  

There are several methods to calculate a player’s value from his projected (or actual) statistics. 

One method is Standings Gain Points, described in the book, How to Value Players for Rotisserie Baseball, by Art McGee (2nd edition available at BaseballHQ.com). SGP converts a player’s statistics in each Rotisserie category into the number of points those stats will allow you to gain in the standings. These are then converted back into dollars.

Another popular method is the Percentage Valuation Method. In PVM, a least valuable, or replacement performance level is set for each category (in a given league size) and then values are calculated representing the incremental improvement from that base. A player is then awarded value in direct proportion to the level he contributes to each category.  

As much as these methods serve to attach a firm number to projected performance, the winning bid for any player is still highly variable depending upon many factors:

•    the salary cap limit

•    the number of teams in the league

•    each team’s roster size

•    the impact of any protected players

•    each team’s positional demands at the time of bidding

•    the statistical category demands at the time of bidding

•    external factors, e.g. media inflation or deflation of value

In other words, a $30 player is only a $30 player if someone in your draft pays $30 for him.

Roster slot valuation (John Burnson)

When you draft a player, what have you bought? 

“You have bought the stats generated by this player.” 

No. You have bought the stats generated by his slot. Initially, the drafted player fills the slot, but he need not fill the slot for the season, and he need not contribute from Day One. If you trade the player during the season, then your bid on Draft Day paid for the stats of the original player plus the stats of the new player. If the player misses time due to injury or demotion, then you bought the stats of whoever fills the time while the drafted player is missing. At season’s end, there will be more players providing positive value than there are roster slots.

Before the season, the number of players projected for positive value has to equal the total number of roster slots—after all, we can’t order owners to draft more players than can fit on their rosters. However, the projected productivity should be adjusted by the potential to capture extra value in the slot. This is especially important for injury-rehab cases and late-season call-ups. For example, if we think that a player will miss half the season, then we would augment his projected stats with a half-year of stats from a replacement-level player at his position. Only then would we calculate prices. Essentially, we want to apportion $260 per team among the slots, not the players.

Average player value by draft round
    Rd     AL/NL       Mxd 
    1       $34        $34
    2       $26        $26
    3       $23        $23
    4       $20        $20
    5       $18        $18
    6       $17        $16
    7       $16        $15
    8       $15        $13
    9       $13        $12
    10      $12        $11
    11      $11        $10
    12      $10         $9
    13       $9         $8
    14       $8         $8
    15       $7         $7
    16       $6         $6
    17       $5         $5
    18       $4         $4
    19       $3         $3
    20       $2         $2
    21       $1         $2
    22       $1         $1
    23       $1         $1

Benchmarks for auction players: 

•    All $30 players will go in the first round. 

•    All $20-plus players will go in the first four rounds. 

•    Double-digit value ends pretty much after Round 11. 

•    The $1 end game starts at about Round 21.

Dollar values by lineup position (Michael Roy)  

How much value is derived from batting order position? 

    Pos     PA        R     RBI      R$
    #1      747      107     72    $18.75
    #2      728      102     84    $19.00
    #3      715       95    100    $19.45
    #4      698       93    104    $19.36
    #5      682       86     94    $18.18
    #6      665       85     82    $17.19
    #7      645       81     80    $16.60
    #8      623       78     80    $16.19
    #9      600       78     73    $15.50

So, a batter moving from the bottom of the order to the clean-up spot, with no change in performance, would gain nearly $4 in value from runs and RBIs alone.

Dollar values: expected projective accuracy  

There is a 65% chance that a player projected for a certain dollar value will finish the season with a final value within plus-or-minus $5 of that projection. That means, if you value a player at $25, you only have about a 2-in-3 shot of him finishing between $20 and $30.

If you want to get your odds up to 80%, the range now becomes +/- $9. You have an 80% shot that your $25 player will finish somewhere between $16 and $34. 

How likely is it that a $30 player will repeat? (Matt Cederholm)

From 2003-2008, there were 205 players who earned $30 or more (using single-league 5x5 values). Only 70 of them (34%) earned $30 or more in the next season. 

In fact, the odds of repeating a $30 season aren’t good. As seen below, the best odds during that period were 42%. And as we would expect, pitchers fare far worse than hitters.

                         Total>$30      # Repeat      % Repeat
Hitters                    167              64           38%
Pitchers                    38               6           16%
Total                      205              70           34%

High-Reliability*
Hitters                     42              16           38%
Pitchers                     7               0            0%
Total                       49              16           33%

100+ BPV
Hitters                     60              25           42%
Pitchers                    31               6           19%
Total                       91              31           19%

High-Reliability and 100+ BPV*
Hitters                     12               5           42%
Pitchers                     6               0            0%
Total                       18               5           28%

    *Reliability figures are from 2006-2008

For players with multiple seasons of $30 or more, the numbers get better. Players with consecutive $30 seasons, 2003-2008:

                Total>$30      # Repeat      % Repeat
Two Years          62              29           55%   
Three+ Years       29              19           66%

Still, a player with two consecutive seasons at $30 in value is barely a 50/50 proposition. And three consecutive seasons is only a 2/3 shot. Small sample sizes aside, this does illustrate the nature of the beast. Even the most consistent, reliable players fail 1/3 of the time. Of course, this is true whether they are kept or drafted anew, so this alone shouldn’t prevent you from keeping a player.

How well do elite pitchers retain their value? (Michael Weddell)

An elite pitcher (one who earns at least $24 in a season) on average keeps 80% of his R$ value from year 1 to year 2.  This compares to the baseline case of only 52%.  

Historically, 36% of elite pitchers improve, returning a greater R$ in the second year than they did the first year.  That is an impressive performance considering they already were at an elite level. 17% collapse, returning less than a third of their R$ in the second year. The remaining 47% experience a middling outcome, keeping more than a third but less than all of their R$ from one year to the next.

Valuing closers

Given the high risk associated with the closer’s role, it is difficult to determine a fair draft value. Typically, those who have successfully held the role for several seasons will earn the highest draft price, but valuing less stable commodities is troublesome.

A rough rule of thumb is to start by paying $10 for the role alone. Any pitcher tagged the closer on draft day should merit at least $10. Then add anywhere from $0 to $15 for support skills. 

In this way, the top level talents will draw upwards of $20-$25. Those with moderate skill will draw $15-$20, and those with more questionable skill in the $10-$15 range.

Profiling the end game

What types of players are typically the most profitable in the end-game?  First, our overall track record on $1 picks:

Avg Return    %Profitable      Avg Prof     Avg. Loss
   $1.89          51%           $10.37       ($7.17)

On aggregate, the hundreds of players drafted in the end-game earned $1.89 on our $1 investments. While they were profitable overall, only 51% of them actually turned a profit. Those that did cleared more than $10 on average. Those that didn’t—the other 49%—lost about $7 apiece.

Pos    Pct.of tot      Avg Val       %Profit    Avg Prof     Avg Loss
CA        12%          ($1.68)         41%        $7.11      ($7.77)
CO         9%           $6.12          71%       $10.97      ($3.80)
MI         9%           $3.59          53%       $10.33      ($4.84)
OF        22%           $2.61          46%       $12.06      ($5.90)
SP        29%           $1.96          52%        $8.19      ($7.06)
RP        19%           $0.35          50%       $11.33     ($10.10) 

These results bear out the danger of leaving catchers to the end; only catchers returned negative value. Corner infielder returns say leaving a 1B or 3B open until late.

Age    Pct.of tot      Avg Val       %Profit    Avg Prof     Avg Loss
< 25      15%          ($0.88)         33%        $8.25      ($8.71)
25-29     48%           $2.59          56%       $11.10      ($8.38)
30-35     28%           $2.06          44%       $10.39      ($5.04)
35+        9%           $2.15          41%        $8.86      ($5.67)

The practice of speculating on younger players—mostly rookies—in the end game was a washout. Part of the reason was that those that even made it to the end game were often the long-term or fringe type. Better prospects were typically drafted earlier.

               Pct.of tot      Avg Val       %Profit    Avg Prof     Avg Loss
Injury rehabs     20%           $3.63          36%       $15.07      ($5.65)

One in five end-gamers were players coming back from injury. While only 36% of them were profitable, the healthy ones returned a healthy profit. The group’s losses were small, likely because they weren’t healthy enough to play.

Realistic expectations of $1 endgamers  (Patrick Davitt)

Many fantasy articles insist leagues are won or lost with $1 batters, because “that’s where the profits are.” But are they?

A 2011 analysis showed that when considering $1 players in deep leagues, managing $1 endgamers should be more about minimizing losses than fishing for profit. In the cohort of batters projected $0 to -$5, 82% returned losses, based on a $1 bid. Two-thirds of the projected $1 cohort returned losses. In addition, when considering $1 players, speculate on speed.

Advanced Draft Strategies

Stars & Scrubs v. Spread the Risk

Stars & Scrubs (S&S): A Rotisserie auction strategy in which a roster is anchored by a core of high priced stars and the remaining positions filled with low-cost players.

Spread the Risk (STR): An auction strategy in which available dollars are spread evenly among all roster slots. 

Both approaches have benefits and risks. An experiment was conducted in 2004 whereby a league was stocked with four teams assembled as S&S, four as STR and four as a control group. Rosters were then frozen for the season. 

The Stars & Scrubs teams won all three ratio categories. Those deep investments ensured stability in the categories that are typically most difficult to manage. On the batting side, however, S&S teams amassed the least amount of playing time, which in turn led to bottom-rung finishes in HRs, RBIs and Runs. 

One of the arguments for the S&S approach is that it is easier to replace end-game losers (which, in turn, may help resolve the playing time issues). Not only is this true, but the results of this experiment show that replacing those bottom players is critical to success. 

The Spread the Risk teams stockpiled playing time, which led to strong finishes in many counting stats, including clear victories in RBIs, wins and strikeouts. This is a key tenet in drafting philosophy; we often say that the team that compiles the most ABs will be among the top teams in RBI and Runs. 

The danger is on the pitching side. More innings did yield more wins and Ks, but also destroyed ERA/WHIP. 

So, what approach makes the most sense? The optimal strategy might be to STR on offense and go S&S with your pitching staff. STR buys more ABs, so you immediately position yourself well in four of the five batting categories. On pitching, it might be more advisable to roster a few core arms, though that immediately elevates your risk exposure. Admittedly, it’s a balancing act, which is why we need to pay more attention to risk analysis and look closer at strategies like the Portfolio3 Plan. 

The LIMA Plan

The LIMA Plan is a strategy for Rotisserie leagues (though the underlying concept can be used in other formats) that allows you to target high skills pitchers at very low cost, thereby freeing up dollars for offense. LIMA is an acronym for Low Investment Mound Aces, and also pays tribute to Jose Lima, a $1 pitcher in 1998 who exemplified the power of the strategy. In a $260 league:

1.    Budget a maximum of $60 for your pitching staff. 

2.    Allot no more than $30 of that budget for acquiring saves. In 5x5 leagues, it is reasonable to forego saves at the draft (and acquire them during the season) and re-allocate this $30 to starters ($20) and offense ($10).

3.    Ignore ERA. Draft only pitchers with:

    • Command ratio (K/BB) of 2.5 or better.

    • Strikeout rate of 7.0 or better.

    • Expected home run rate of 1.0 or less.

4.    Draft as few innings as your league rules will allow. This is intended to manage risk. For some game formats, this should be a secondary consideration.

5.    Maximize your batting slots. Target batters with:

    • Contact rate of at least 80% 

    • Walk rate of at least 10% 

    • PX or Spd level of at least 100

Spend no more than $29 for any player and try to keep the $1 picks to a minimum.

The goal is to ace the batting categories and carefully pick your pitching staff so that it will finish in the upper third in ERA, WHIP and saves (and Ks in 5x5), and an upside of perhaps 9th in wins. In a competitive league, that should be enough to win, and definitely enough to finish in the money. Worst case, you should have an excess of offense available that you can deal for pitching.

The strategy works because it better allocates resources. Fantasy leaguers who spend a lot for pitching are not only paying for expected performance, they are also paying for better defined roles—#1 and #2 rotation starters, ace closers, etc.—which are expected to translate into more IP, wins and saves. But roles are highly variable. A pitcher’s role will usually come down to his skill and performance; if he doesn’t perform, he’ll lose the role.

The LIMA Plan says, let’s invest in skill and let the roles fall where they may. In the long run, better skills should translate into more innings, wins and saves. And as it turns out, pitching skill costs less than pitching roles do. 

In snake draft leagues, don’t start drafting starting pitchers until Round 10. In shallow mixed leagues, the LIMA Plan may not be necessary; just focus on the peripheral metrics. In simulation leagues, build your staff around those metrics

Variations on the LIMA Plan

LIMA Extrema: Limit your total pitching budget to only $30, or less. This can be particularly effective in shallow leagues where LIMA-caliber starting pitcher free agents are plentiful during the season.

SANTANA Plan: Instead of spending $30 on saves, you spend it on a starting pitcher anchor. In 5x5 leagues where you can reasonably punt saves at the draft table, allocating those dollars to a high-end LIMA-caliber starting pitcher can work well as long as you pick the right anchor. 

Total Control Drafting (TCD)

On Draft Day, we make every effort to control as many elements as possible. In reality, the players that end up on our teams are largely controlled by the other owners. Their bidding affects your ability to roster the players you want. In a snake draft, the other owners control your roster even more. We are really only able to get the players we want within the limitations set by others. 

However, an optimal roster can be constructed from a fanalytic assessment of skill and risk combined with more assertive draft day demeanor.

Why this makes sense 

1. Our obsession with projected player values is holding us back. If a player on your draft list is valued at $20 and you agonize when the bidding hits $23, odds are about two chances in three that he could really earn anywhere from $15 to $25. What this means is, in some cases, and within reason, you should just pay what it takes to get the players you want. 

2. There is no such thing as a bargain. Most of us don’t just pay what it takes because we are always on the lookout for players who go under value. But we really don’t know which players will cost less than they will earn because prices are still driven by the draft table. The concept of “bargain” assumes that we even know what a player’s true value is.  

3. “Control” is there for the taking. Most owners are so focused on their own team that they really don’t pay much attention to what you’re doing. There are some exceptions, and bidding wars do happen, but in general, other owners will not provide that much resistance.  

How it’s done

1. Create your optimal draft pool.

2. Get those players.

Start by identifying which players will be draftable based on the LIMA or Portfolio3 criteria. Then, at the draft, focus solely on your roster. When it’s your bid opener, toss a player you need at about 50%-75% of your projected value. Bid aggressively and just pay what you need to pay. Of course, don’t spend $40 for a player with $25 market value, but it’s okay to exceed your projected value within reason.

From a tactical perspective, mix up the caliber of openers. Drop out early on some bids to prevent other owners from catching on to you. 

In the end, it’s okay to pay a slight premium to make sure you get the players with the highest potential to provide a good return on your investment. It’s no different than the premium you might pay for a player with position flexbility or to get the last valuable shortstop. With TCD, you’re just spending those extra dollars up front to ensure you are rostering your targets. As a side benefit, TCD almost asssures that you don’t leave money on the table.  

Mayberry Method

The foundation of the Mayberry Method (MM) is the assertion that we really can’t project player performance with the level of precision that advanced metrics and modeling systems would like us to believe. 

MM is named after the fictional TV village where life was simpler. MM evaluates skill by embracing the imprecision of the forecasting process and projecting performance in broad strokes rather than with hard statistics. 

MM reduces every player to a 7-character code. The format of the code is 5555 AAA, where the first four characters describe elements of a player’s skill on a scale of 0 to 5. These skills are indexed to the league average so that players are evaluated within the context of the level of offense or pitching in a given year.

The three alpha characters are our reliability grades (Health, Experience and Consistency) on the standard A-to-F scale. The skills numerics are forward-looking; the alpha characters grade reliability based on past history.

Batting

The first character in the MM code measures a batter’s power skills. It is assigned using the following table:

Power Index      MM   
0 - 49            0  
50 - 79           1  
80 - 99           2  
100 - 119         3  
120 - 159         4  
160+              5   

The second character measures a batter’s speed skills. RSpd takes our Statistically Scouted Speed metric (Spd) and adds the elements of opportunity and success rate, to construct the formula of RSpd = Spd x (SBO + SB%).

RSpd               MM   
0 - 39              0    
40 - 59             1    
60 - 79             2    
80 - 99             3    
100 - 119           4    
120+                5    

The third character measures expected batting average. 

xBA Index     MM
0-87           0
88-92          1
93-97          2
98-102         3
103-107        4
108+           5

The fourth character measures playing time.

Role                        PA           MM
Potential full-timers       450+          5
Mid-timers               250-449          3
Fringe/bench             100-249          1
Non-factors                 0-99          0

Pitching

The first character in the pitching MM code measures xERA, which captures a pitcher’s overall ability and is a proxy for ERA, and even WHIP. 

xERA Index       MM
0-80             0
81-90            1
91-100           2
101-110          3
111-120          4
121+             5

The second character measures strikeout ability.

K/9 Index    MM
0-76         0
77-88        1
89-100       2
101-112      3
113-124      4
125+         5

The third character measures saves potential.

Description                                   Saves est.    MM
No hope for saves; starting pitchers            0           0
Speculative closer                            1-9           1
Closer in a pen with alternatives           10-24           2
Frontline closer with firm bullpen role       25+           3
The fourth character measures playing time. 

Role                                 IP     MM
Potential #1-2 starters             180+     5
Potential #3-4 starters          130-179     3
#5 starters/swingmen              70-129     1
Relievers                           0-69     0

Overall Mayberry Scores

The real value of Mayberry is to provide a skills profile on a player-by-player basis. I want to be able to see this…

Player A    4455 AAB

Player B    5245 BBD

Player C    5255 BAB

Player D    5155 BAF

…and make an objective, unbiased determination about these four players without being swayed by preconceived notions and baggage. But there is a calculation that provides a single, overall value for each player.

This is the calculation for the overall MM batting score:

MM Score = 

(PX score + Spd score + xBA score + PA score) x PA score

An overall MM pitching score is calculated as:

MM Score = 

((xERA score x 2) + K/9 score + Saves score + IP score) x (IP score + Saves score)

The highest score you can get for either is 100. That makes the result of the formula easy to assess.

BaseballHQ.com analyst Patrick Davitt did some great research about using Reliability Grades to adjust the Mayberry scores. His research showed that “higher-reliability players met their Mayberry targets more often than their lower-reliability counterparts, and players with all “D” or “F” reliability scores underperform Mayberry projections far more often. Those results can be reflected by multiplying a player’s MM Score by each of three reliability bonuses or penalties:”

I’ve taken his work a minor step further and applied slightly different multipliers to each Reliability element. 

            Health      Experience    Consistency
A           x 1.10       x 1.10          x 1.10   
B           x 1.05       x 1.05          x 1.05
C           x 1.00       x 1.00          x 1.00
D           x 0.90       x 0.95          x 0.95
F           x 0.80       x 0.90          x 0.90

So, let’s perform the overall calculations for Player A above, using these Reliability adjustments.

Player A: 4455 AAB 

= (4+4+5+5) x 5 

= 90 x 1.10 x 1.10 x 1.05

= 114.3

The Portfolio3 Plan (P3)

When it comes to profitability, all players are not created equal. Every player has a different role on your team by virtue of his skill set, dollar value/draft round, position and risk profile. When it comes to a strategy for how to approach a specific player, one size does not fit all.

We need some players to return fair value more than others. A $40/first round player going belly-up is going to hurt you far more than a $1/23rd round bust. End-gamers are easily replaceable.

We rely on some players for profit more than others. First-rounders do not provide the most profit potential; that comes from players further down the value rankings.

We can afford to weather more risk with some players than with others. Since high-priced early-rounders need to return at least fair value, we cannot afford to take on excessive risk. Our risk tolerance opens up with later-round/lower cost picks.

Players have different risk profiles based solely on what roster spot they are going to fill. Catchers are more injury prone. A closer’s value is highly dependent on managerial decision. These types of players are high risk even if they have great skills. That needs to affect their draft price or draft round.

For some players, the promise of providing a scarce skill, or productivity at a scarce position, may trump risk. Not always, but sometimes. The determining factor is usually price.

In the end, we need a way to integrate all these different types of players, roles and needs. We need to put some structure to the concept of a diversified draft approach. Thus:

The Portfolio3 Plan provides a three-tiered structure to the draft. Just like most folks prefer to diversify their stock portfolio, P3 advises to diversify your roster with three different types of players. Depending upon the stage of the draft (and budget constraints in auction leagues), P3 uses a different set of rules for each tier that you’ll draft from. The three tiers are:

1. Core Players

2. Mid-Game Players 

3. End-Game Players

Mayberry scores can be used as proxies for the skills filters. When planning your draft, pretty much all you need to remember is the number “3”. That essentially represents “just over league average” and makes it easy to set your targets.

TIER 1: CORE PLAYERS 

General Roster Goals

Auction target: Budget a maximum of $160. Any player purchased for $20 or more should meet the Tier 1 skills criteria

Snake draft target: 5-8 players, with an emphasis on those drafted in the earlier rounds

Reliability grades: No worse than “B” for each variable (Health, Experience and Consistency)

Playing time: No restrictions, however, pricier early round players should have more guaranteed playing time 

Batter skills: Minimum MM scores of 3 in xBA plus either PX or RSpd 

Pitcher skills: Minimum MM scores of 3 in xERA and K/9 

Tier 1 players provide the foundation to your roster. These are your prime contributors and where you will invest the largest percentage of your budget or early round picks. There is no room for risk here, so the majority of these players should be batters.

TIER 2: MID-GAME PLAYERS

General Roster Goals

Auction target: Budget between $50 and $100; players should be under $20 

Snake draft target: 7-13 players

Reliability grades: No worse than “B” for Health, no worse than “C” for Experience and Consistency

Playing time: Must have a MM score of 5 for batters (meaning full-time batters) and minimum 3 for pitchers (meaning at least mid-rotation starting pitchers)  

Batter skills: Minimum MM scores of 3 in xBA or PX or RSpd

Pitcher skills: Minimum MM score of 3 in xERA or K/9

Tier 1 players are all about skill. Tier 2 is all about accumulating playing time, particularly on the batting side, with lesser regard to skill. This is where you can beef up on runs and RBI. If a player is getting 500 AB, he is likely going to provide positive value in those categories just from opportunity alone. And given that his team is seeing fit to give him those AB, he is probably also contributing somewhere else.

For pitchers, we use Tier 2 to accumulate arms whose innings provide some level of positive support, either by stockpiling strikeouts or by building your ERA foundation.

TIER 3: END-GAME PLAYERS

General Roster Goals

Auction target: Budget up to $50; players should be under $10

Snake draft target: 5-10 players

Reliability grades: No restrictions, except no “F” Health grades.

Playing time: No restrictions 

Batter skills: Minimum MM scores of 3 in xBA plus either PX or RSpd (same as Tier 1)

Pitcher skills: Minimum MM score of 3 in xERA

Tier 3 players are your gambling chips, but every end-gamer must provide the promise of upside. For that reason, the focus must remain on skill and conditional opportunity. MP3 drafters should fill the majority of their pitching slots from this group.

By definition, end-gamers are typically high risk players, but risk is something you’ll want to embrace here. If a a Tier 3 player does not pan out, he can be easily replaced.

As such, the best Tier 3 options should possess the MM skill levels noted above, and at least one of the following:

•    playing time upside as a back-up to a risky front-liner 

•    an injury history that has depressed his value (but not chronically injured players) 

•    solid skills demonstrated at some point in the past 

•    minor league potential even if he has been more recently a major league bust

A complete list of players in each tier appears in the back of the book starting on page 269. One of the major benefits of the MP3 process is that any player failing to find a home in one of the tiers can be safely ignored. Either his skills are not draft-worthy or his risk-profile too dangerous, regardless of skill. By shrinking the draftable player pool, it makes the roster planning and construction process easier. 

Category Targets

The the final task is to set MM targets for each category. 

If you are in a league with good trading activity, this may not be important—you can always deal away excesses to beef up weak categories. But for those in leagues with little or no trading, drafting a balanced team is critical.  

For skills budgeting purposes, here are targets for several standard leagues:

BATTING               PX      RSpd     xBA       PA    
12-team mixed         41       28       40       66   
15-team mixed         41       26       39       64   
12-team AL/NL         37       23       32       54   

PITCHING             xERA*    K/9*      Sv       IP   
12-team mixed         23       33        7       29   
15-team mixed         20       30        6       30       
12-team AL/NL         17       27        5       25    

* Make sure the majority of these points come from starting pitchers.

As you draft, track each MM score and keep a running total of all the categories. With the above goals will allow you to shift your in-draft targets if you see you are falling behind in any area. 

Building a Homogeneous Head-to-Head Team (David Martin)

Though variety is the spice of life, it has no place in the type of players rostered on head-to-head teams. Teams in head-to-head leagues need players cut from the same cloth—players that are completely homogenous. 

Focusing on certain metrics helps build a homogenous team; Drafting a homogenous team inherently builds consistency into your roster. Our filters for such success are: 

•    Contact rate = minimum 80% 

•    xBA = minimum .280

•    PX (or Spd) = minimum 120

•    RC/G = minimum 5.00 

Research shows that a homogeneous team based on these metrics is more likely to be a consistent team, which is the roster holy grail for head-to-head players.

Ratio Insulation in Head-to-Head Leagues (David Martin)

On a week-to-week basis, inequities are inherent in the head-to-head game. One way to eliminate your competitor’s advantage in the pure numbers game is to build your team’s foundation around the ratio categories. 

One should normally insulate at the end of a draft, once your hitters are in place. To obtain several ratio insulators, target players that have:

•    Cmd greater than 3.0

•    Dom greater than 7.5

•    xERA less than 3.30

While adopting this strategy may compromise wins, research has shown that wins come at a cost to ERA and WHIP. Roster space permitting, adding two to four insulators to your team will improve your team’s weekly ERA and WHIP.

In-Season Analyses

The efficacy of streaming  (John Burnson)

In leagues that allow weekly or daily transactions, many owners flit from hot player to hot player. But published dollar values don’t capture this traffic—they assume that players are owned from April to October. For many leagues, this may be unrealistic.

We decided to calculate these “investor returns.” For each week, we identified the top players by one statistic—BA for hitters, ERA for pitchers—and took the top 100 hitters and top 50 pitchers. We then said that, at the end of the week, the #1 player was picked up (or already owned) by 100% of teams, the #2 player was picked up or owned by 99% of teams, and so on, down to the 100th player, who was on 1% of teams. (For pitchers, we stepped by 2%.) Last, we tracked each player’s performance in the next week, when ownership matters.

We ran this process anew for every week of the season, tabulating each player’s “investor returns” along the way. If a player was owned by 100% of teams, then we awarded him 100% of his performance. If the player was owned by half the teams, we gave him half his performance. If he was owned by no one (that is, he was not among the top players in the prior week), his performance was ignored. A player’s cumulative stats over the season was his investor return. 

The results... 

•    60% of pitchers had poorer investor returns, with an aggregate ERA 0.40 higher than their true ERA. 

•    55% of batters had poorer investor returns, but with an aggregate batting average virtually identical to the true BA. 

Sitting stars and starting scrubs  (Ed DeCaria)

In setting your pitching rotation, conventional wisdom suggests sticking with trusted stars despite difficult matchups. But does this hold up? And can you carefully start inferior pitchers against weaker opponents? Here are the ERAs posted by varying skilled pitchers facing a range of different strength offenses:

                OPPOSING OFFENSE (RC/G)
Pitcher(ERA)     5.25+    5.00    4.25    4.00    <4.00
3.00-            3.46     3.04     3.04    2.50    2.20
3.50             3.98     3.94     3.44    3.17    2.87
4.00             4.72     4.57     3.96    3.66    3.24
4.50             5.37     4.92     4.47    4.07    3.66
5.00+            6.02     5.41     5.15    4.94    4.42

Recommendations:

1.    Never start below replacement-level pitchers. 

2.    Always start elite pitchers.

3.    Other than that, never say never or always. 

Playing matchups can pay off when the difference in opposing offense is severe. 

Two-start pitcher weeks  (Ed DeCaria)

A two-start pitcher is a prized possession. But those starts can mean two DOMinant outings, two DISasters, or anything else in between, as shown by these results:

PQS Pair      % Weeks     ERA       WHIP      Win/Wk      K/Wk
DOM-DOM         20%       2.53      1.02       1.1        12.0
DOM-AVG         28%       3.60      1.25       0.8         9.2
AVG-AVG         14%       4.44      1.45       0.7         6.8 
DOM-DIS         15%       5.24      1.48       0.6         7.9
AVG-DIS         17%       6.58      1.74       0.5         5.7
DIS-DIS          6%       8.85      2.07       0.3         5.0

Weeks that include even one DISaster start produce terrible results. Unfortunately, avoiding such disasters is much easier in hindsight. But what is the actual impact of this decision on the stat categories?

ERA and WHIP: When the difference between opponents is extreme, inferior pitchers can be a better percentage play. This is true both for 1-start pitchers and 2-start pitchers, and for choosing inferior one-start pitchers over superior two-start pitchers.

Strikeouts per Week: Unlike the two rate stats, there is a massive shift in the balance of power between one-start and two-start pitchers in the strikeout category. Even stars with easy one-start matchups can only barely keep pace with two-start replacement-level arms in strikeouts per week.

Wins per week are also dominated by the two-start pitchers. Even the very worst two-start pitchers will earn a half of a win on average, which is the same rate as the very best one-start pitchers.

The bottom line:  If strikeouts and wins are the strategic priority, use as many two-start weeks as the rules allow, even if it means using a replacement-level pitcher with two tough starts instead of a mid-level arm with a single easy start. But if ERA and/or WHIP management are the priority, two-start pitchers can be very powerful, as a single week might impact the standings by over 1.5 points in ERA/WHIP, positively or negatively. 

Six Tips on Category Management (Todd Zola)

1.    Disregard whether you are near the top or the bottom of a category; focus instead on the gaps directly above and below your squad.

2.    Prorate the difference in stats between teams.

3.    ERA tends to move towards WHIP.

4.    As the season progresses, the number of AB/IF do not preclude a gain/loss in the ratio categories.

5.    An opponent’s point lost is your point gained.

6.    Most important! Come crunch time, forget value, forget names, and forget reputation. It’s all about stats and where you are situated within each category.

Consistency  (Dylan Hedges)

Few things are as valuable to H2H league success as filling your roster with players who can produce a solid baseline of stats, week in and week out. In traditional leagues, while consistency is not as important—all we care about are aggregate numbers—filling your team with consistent players can make roster management easier.

Consistent batters have good plate discipline, walk rates and on base percentages. These are foundation skills. Those who add power to the mix are obviously more valuable, however, the ability to hit home runs consistently is rare. 

Consistent pitchers demonstrate similar skills in each outing; if they also produce similar results, they are even more valuable.

We can track consistency but predicting it is difficult. Many fantasy leaguers try to predict a batter’s hot or cold streaks, or individual pitcher starts, but that is typically a fool’s errand. The best we can do is find players who demonstrate seasonal consistency; in-season, we must manage players and consistency tactically.

Consistency in points leagues (Bill Macey)

Previous research has demonstrated that week-to-week statistical consistency is important for Rotisserie-based head-to-head play. But one can use the same foundation in points-based games. A study showed that not only do players with better skills post more overall points in this format, but that the format caters to consistent performances on a week-to-week basis, even after accounting for differences in total points scored and playing-time. 

Therefore, when drafting your batters in points-based head-to-head leagues, ct% and bb% make excellent tiebreakers if you are having trouble deciding between two players with similarly projected point totals. Likewise, when rostering pitchers, favor those who tend not to give up home runs.

Other Diamonds

Cellar value

The dollar value at which a player cannot help but earn more than he costs. Always profit here.

Crickets

The sound heard when someone’s opening draft bid on a player is also the only bid.

Scott Elarton List

Players you drop out on when the bidding reaches $1.

End-game wasteland

Home for players undraftable in the deepest of leagues, who stay in the free agent pool all year. It’s the place where even crickets keep quiet when a name is called at the draft. 

FAAB Forewarnings 

1.    Spend early and often.

2.    Emptying your budget for one prime league-crosser is a tactic that should be reserved for the desperate.

3.    If you chase two rabbits, you will lose them both.

Fantasy Economics 101

The market value for a player is based on the aura of past performance, not the promise of future potential. Your greatest advantage is to leverage the space between market value and real value.

Fantasy Economics 102

The variance between market value and real value is far more important than the absolute accuracy of any individual player projection.

Hope

A commodity that routinely goes for $5 over value at the draft table.

Professional Free Agent (PFA)

Player whose name will never come up on draft day but will always end up on a roster at some point during the season as an injury replacement.

Mike Timlin List

Players who you are unable to resist drafting even though they have burned you multiple times in the past. 

Seasonal Assessment Standard

If you still have reason to be reading the boxscores during the last weekend of the season, then your year has to be considered a success.

The Three Cardinal Rules for Winners

If you cherish this hobby, you will live by them or die by them...

1. Revel in your success; fame is fleeting.

2. Exercise excruciating humility.

3. 100% of winnings must be spent on significant others.