Michael Salfino of the Wall Street Journal (full disclosure: Salfino is a terrific writer) has a very useful piece today, discussing run differential and how it hasn’t seemed to apply to the New York Mets.
Writes Salfino: “The New York Mets are off to the most cognitively dissonant start in baseball… The Mets are on pace to finish 87-75 while being outscored by 122 runs. This would be a rather historic achievement: All-time, the worst run differential by a winning team belonged to the 1905 Detroit Tigers (minus-90), who went 79-74.”
In other words: either the Mets are an outlier more than a century in the making, or they’ll need to play significantly better to keep winning games at their current rate.
The other problem with even doing that is, winning at their current rate only has them at 22-20, good for fourth place in the National League East. That misrepresents their position somewhat; fourth place is just one game out of third, which is the final National League playoff spot. But the point is, significant over-achievement of their run differential doesn’t have them in a playoff spot, with a cushion against regression. It has left them with more work to do.
The Mets, for instance, are 22-20. But their Pythagorean record, based on runs scored and runs allowed, is 18-24, a .429 winning percentage. So if they play at their current level over the last 120 games, they’ll be likely to finish with a 51-69 record, for a total record of 73-89.
But let’s go ahead and extrapolate from the division’s run differential to determine, if the five NL East teams simply play at the levels they have for the first 42 games, how they will finish. We’ll take the current records, which are established, and add the winning percentage those teams should expect over the final 120 games, given their runs scored and runs allowed.
Team Record to Date Expected Record, Rest of Season Final Record
Braves 26-17 69-50 95-67
Nationals 25-17 66-54 91-71
Marlins 23-19 60-60 83-79
Mets 22-20 51-69 73-89
Phillies 21-22 61-58 82-80
What do we learn from this? Well, the quarter of a season playing above their expected win total has been kind to the Mets. Had they simply been playing at the 18-24 mark their runs scored/runs allowed would have predicted, they’d be on pace to finish 69-93. So those are four wins they don’t have to give back.
And if Pythagorean record isn’t picking up some aspect of this team that allows it to consistently outperform its run differential, the good times may continue.
Unfortunately, there’s well over a century of baseball that says the Mets are still a good bet to finish last in the division, and by a good margin.
4 Comments
Salfino might suffer from cognitive dissonance but this theory is preposterous, I’d guess that poor old Pythagoras is spinning in his grave somewhere in ancient Alexandria. The Mets lost 14-5, 8-0, 8-1, 18-9 in 2012 to date. That’s -33 run deferential – and only 4 losses. Granted, they aint putting up big numbers but these figures can be so easily skewed in just a few games and only a few losses. I’d argue that a good team with generally consistant good pitching can win more games than it loses and still have a nasty negative run deferential. I’d guess Salfino’s talking about probabilities here but even that doesn’t hold water.
Or as Studs Terkel would say, “It just don’t cut the mustard”. JMO
I mean, I love Studs Terkel. But his sample size was every team, ever, which presumably includes teams with consistent good pitching. And no one’s ever done it.
OK. But what exactly does “on pace to finish 87-75 while being outscored by 122 runs”
really mean? To my mind it’s nonsensical to think in terms of “on pace” where infinite factors
and conditions affect the outcome every day. There’s nothing to support a universe where baseball is played in a scoring continuum.
I just did the arithmetic and as of today they’ve been outscored by 32 runs. The 4 games I mentioned were a -33. Tomorrow they might win by 20. Then what?
Then their Pythagorean record would change. Let’s cross that bridge when they win by 20.