The Secret Sauce
Key Features/Components
- For position players the proposed figure of merit (Hall of Maximum Achievement Score- HOMAS) is primarily based on a combination of peak period bWAR- Wins Against Replacement (<a href="https://www.baseball-reference.com/about/war_explained_position.shtml">Baseball-Reference position WAR explained</a>) and career OPS+ stats with the OPS+ figure scaled for primary positions played and with a base stealing component added. Additional points are awarded for extended peak longevity.
- For pitchers the proposed figure of merit is based on a combination of peak period pitching pWAR (<a href="https://www.baseball-reference.com/about/war_explained_pitch.shtml">Baseball-Reference pitching WAR explained</a>) and career ERA+ stats, again with points awarded for extended peak longevity. Some adjustments are made for relief pitchers. Only pitchers pitching WAR is considered. Offensive contributions are ignored. If the pitcher was also a position player then his bWAR was considered separately.
- A 20% penalty is levied in the Peak WAR and Adjusted OPS+ and ERA+ calculations for confirmed or strongly suspected PED users.
- In accordance with their recognition as a true major league, performance statistics of players in those recognized Negro Leagues are fully integrated into all analyses.
- To account for shorter NGL seasons while still trying to allow for meaningful performance comparisons certain methodological adjustments are made.
- Reasonable assumptions are made regarding missing statistics in order to complete merit score calculations. This occurs only in certain limited cases.
Calculating the Various Selected Metrics
The major criterion in evaluating exceptional player performance is finding a measurement that relates to long term peak performance. The WAR metric is almost ideal as it combines offense, base-running and defense for a position player and a pure pitching component for a pitcher to give a measure of a player’s contribution/worth. The higher a player’s score over a longer period defines excellence. In determining the actual score for this first component for traditional MLers, the number of seasons that a player achieved a WAR score of three or more is determined. The total WAR score during that period is summed and divided by the number of peak seasons to determine a Peak War Score. To simplify WAR tallies, a player is awarded the middle score for each integer value for each peak season. That is a score between 3.0 and 3.9 is awarded a 3.5 while that between 4.0 and 4.9 is awarded a 4.5, etc. In addition an extra half point is awarded for each of the peak seasons exceeding eight seasons. Thus the formula for this metric is:
Peak War Score = Accumulated Peak War/Peak War Seasons + .5 x (Peak War Seasons – 8)
If it was determined that a player regularly used PED’s at some point in his career or was highly suspected a 20% penalty was subtracted from the first term of the equation.
As certain measured players had less than eight peak WAR seasons (that is seasons with a WAR greater than 3) then a recheck was done for possible value 2 WAR seasons (indicative of average ability for that season) to fill part of the remaining 8 year peak time frame. A player who did not have eight seasons of WAR scores of two or more wound up receiving a negative .5 for the longevity score for each of those seasons when calculating a final peak WAR score.
To determine a relevant peak score for Negro Leaguers, modifications to how seasonal WAR and number of peak seasons were determined had to be made. NGL seasons were typically less than 50% of the length of MLB seasons. If the same formulas were used NGLers scores would be massively deflated as WAR is a cumulative statistic. To correct for this, a substitute average WAR over a 162 game season as calculated by The Baseball Reference was first determined. This was then modified to an expected seasonal Peak WAR value. The “best fit” conversion was determined by comparing the relation between 7 year peak WAR average and career 162 game WAR average for several hundred, top ranking, retired, “live ball” era players as used by the Jaffe WAR Scores system (JAWS).
To determine a longevity factor a player’s total NGL games were converted to an equivalent total 140 game MLB seasons. This value was was based on the premiss that any typical everyday player’s season would be about 10% less than the standard 154 games for that era (1928-1948). As the average number of 140 games season for the evaluated NGLers was around 5, six was used as a reference number for minimum peak seasons (as compared to the traditional MLers eight seasons). The peak WAR score formula for pure NGLers then is:
Peak WAR Score = (Peak War to 162 Game WAR Slope Conversion Factor) x (162 Game WAR) + Peak War to 162 Game WAR Constant Conversion Factor +.5 x (Converted 140 Game NGL seasons – 6)
In general, for players who played in both the traditional Major Leagues and the Negro Leagues a weighted average of the two peak scores is calculated with eight seasons again used as the longevity reference. In the unique case of several NGL players who arrived late in their careers to the Major Leagues, the longevity reference used was 6 years. These players so evaluated were Satchel Paige, Monte Irvin, Sam Jethroe, and Luke Easter. Also note that no NGL best-fit peak WAR correction was used for those combo players that played less than six seasons in the Negro Leagues.
The use of “best fit” conversion factors to derive a Peak WAR value might be considered a fairly radical statisical manipulation to determine NGL player value. However, correlations in the .8 range for the sampled Major League players indicates this to be a valid means to project a peak WAR value.
Despite these attempts to reconcile NGLer performance, there does remain a continuing sense that their careers are being statistically short-changed, particularly as many had outstanding careers in Cuban and Mexican leagues which do not count in determining their overall performance. One player who particularly suffered from this limitation was Martinn Dihigo. As a compromise, games played in these leagues (not actual statistical performances) are included in the calculation of a player’s peak longevity if their Latin leagues OPS+/ERA+ is at least 110 or at least 10 points more than their Negro League record.
The second criterion selected for HOMAS score is a rate based metric that is principally related to a position players overall ability to hit (OPS+) and pitchers overall ability to pitch (ERA+). The strictly OPS+ statistic measures comparative offensive production across eras while ERA+ does the same for pitchers. To add other components to better get a sense for overall performance, the OPS+ is adjusted by a position multiplier according to a player’s primary position. For shortstops and catchers it was a 1.3; for second basemen it was 1.25, for centerfielders and third basemen it was a 1.2. Right fielders are at 1.15; leftfielders and first basemen are at 1.1; designated hitters are at 1.0. Adjustments to this multiplier are made for significant time at other positions. Also a longevity scaling factor is included in the metric. The scaling factor is the ratio between peak WAR seasons and a minimum baseline. The baseline used is the same as that for Peak WAR longevity (8 for traditional MLB and 6 for total career NGL players or combo league players that began their traditional MLB career late in their entire career).
Finally a stolen base component is considered. For traditional MLB players at least 25 steals during any 5 seasons qualifed that player for evaluation. As an eight season longevity reference was again used, seasons with 20 or more steals were considered as needed to fill in an eight season total period. Included in this metric’s formula is an average peak steals component, an efficiency measurement, a .2 bonus for each season beyond eight with 25 or more steals and a 20% deduction for confirmed or highly suspected PED use. A stolen base efficiency rate of 72% was set as a reference standard.
For entire career NGL position players, career stolen bases was used instead of calculating a result for seasons with a minimum of 25 or 20 steals. These players had to achieve at least 100 career stolen bases.
As some data for caught stealing is missing for some traditional MLers and completely missing for NGLers, some assumptions were made to complete the data. For traditional MLers caught stealing data ratios from available high steal seasons were used to fill in the blanks. For NGLers an 80% efficiency rate, that matches them to the median stolen base rate of the top 50 modern era players, was used.
The formula for this component is:
Steals Score = (Peak Average Steals x (Peak Steal Efficiency - .72) + .2 x (Peak Steals Seasons – 8 (MLB) or 6 (NGL)) – 20% of the first term (if PED use))/2
The basic total adjusted OPS+ score formula is:
Adjusted OPS+ = (Position multiplier x OPS+ - 100)/10 * (1 + .1 (Peak Seasons - Peak Seasons Longevity Baseline) + Peak Steals Score.
Like Total Peak War Score a 20% penalty is subtracted for any player regularly involved or suspected of using PEDs. This penalty is 20% of the first term of this equation.
The final HOMAS (Hall of Maximum Achievement Score) metric for position players is then the sum of Total Peak Score and the Adjusted OPS+. For pitchers it is the sum of Total Peak Score and Adjusted ERA+. The modifications for ERA+ for pitchers are a 1.3 scaling factor for starting pitchers versus a 1.1 scaling factor for relief pitchers. Similar to position players, there is a scaling ratio for peak longevity with the baseline denominator value set at 8 yrs for starting pitchers and 5 years for relief pitchers. However, for every season where a reliever's role was primarily as a starting pitcher, those years are added to to the minimum longevity baseline.
Correlations between the two major HOMA score factors indicate a very close similarity in how performance is measured by the two factors, except in the case of relief pitchers. If one or the other measurement scale is valid then so is the other. The less strict relationship for relief pitchers again indicates how difficult it is to develop consistent measures for this particular baseball role.
NGL Position Players .93
MLB Position Players .84
NGL Starting Pitchers .94
MLB Starting Pitchers .80
MLB Relief Pitchers .49