Yankees’ Expected Run Differential
Jul 3, 2014; Minneapolis, MN, USA; New York Yankees shortstop Derek Jeter (2) celebrates with second baseman Brendan Ryan (17) after beating the Minnesota Twins at Target Field. The Yankees won 7-4. Mandatory Credit: Jesse Johnson-USA TODAY Sports
Bill James sported the idea that a teams run differential is a better measure of its true talent level than its actual team record. He devised a formula–(runs scored^2)/((runs scored^2)+(runs allowed^2))– to calculate an expected winning percentage based on the run differential and called it the Pythagorean win-loss because the “squares” reminded him of the Pythagorean Theorem. The idea is that two teams can have the same number of runs allowed and scored but distribute them in a way that allows one team to win more games than the other. This distribution is random, not a skill. Joe Peta, author of “Trading Bases”, took this a step further and identified cluster luck, the idea that two teams can have (allow) the same number and type of hits but scatter them in a way that they score (surrender) a different amount of runs. This clustering is also random. Dave Cameron explained these two ideas and highlighted their inclusion on fangraphs. The first section of the standings page is the normal win-loss record. The Yankees are 42-42 with a -34 run differential (4.01 runs scored/game, 4.42 runs allowed/game). The second section of the standings page is PythagPat which is a variation of Bill James’ Pythag that derives the exponent based on the run environment. The Yankees actual run differential record, based on normal distribution of runs, is 38-46. This is better than looking at just win-loss records, but still doesn’t go far enough because it doesn’t strip out the sequencing and cluster luck involved in offensive and defensive events. As Dave says, “It removes the timing aspects of converting runs into wins, but ignores the timing aspects of converting baserunners into runs.” The 3rd section of the standings page is based on BaseRuns and gives us an expected run differential that can be translated into a record. Dave points out that this is a better attempt at predicting runs scored and allowed based on offensive events while stripping out sequencing than using wOBA differential. BaseRuns shows that the Yankees expected run differential and record should be -27 (4.13 runs scored/game, 4.45 runs allowed/game) and 39-45, respectively. That is 3 wins worse than actuality. The Yankees expected run differential luck up to this point has saved them from an unsightly 39-45 record. Another feature on fangraphs, the playoff odds, echo what the current standings, PythagPat, and BaseRuns conclude. This Yankees team, as currently constructed will have a hard time making the playoffs. Fangraphs projection mode from the playoff odds page says they have a 11.0% chance of winning the division and a 4.4% chance of winning a wild card spot. Their combined chance of making the playoffs, even just the coin-flip game, is 15.4%. Small moves like starting Brendan Ryan at shortstop and moving Derek Jeter down in the order would help, but the team needs upgrades before the trade deadline to move the needle. A starting pitcher, 3rd baseman, 2nd baseman, rightfielder, and reliever are all spots that could be upgraded. The Yankees certainly have the ability to take on money and the prospects to get most deals done, but there might just be too many holes to fill in one trading season.