We all want things to be perfect.
Yet rarely in life do things work out that way. This is so universal that we even have an idiom to describe things that fall short of perfection. You’re probably heard of the saying – don’t let the perfect be the enemy of the good. Which brings us to WAR.
WAR is the baseball statistical community’s best attempt to measure the value of players’ contributions to wins, the “W” in the acronym. Is it perfect? Absolutely not – but then it’s never claimed to be that, either. Some people get hysterical because there’s a decimal place and thinks that gives a level of precision that just isn’t there. You can’t claim a player with a 2.6 WAR is better than one with a 2.5 WAR – it’s just not that precise.
We can argue how precise things are on an individual basis and probably not get very far. But when we go from looking at things on an individual basis and move to a team basis, well, we move to something that’s a little easier to measure. And that’s what Devan Fink did back following the 2018 season. Here’s the main takeaway:
As you can see, however, team wins and WAR wins are strongly correlated (R-squared value of 0.86). The line of best fit is y = 1.06 x – 4.61. So, if your team is projected to produce 100 “WAR wins” (47.628 + 52.372 WAR), they would be expected to win about 101 games. While the correlation does create some variance, the line of best fit demonstrates that the ratio of WAR to wins is pretty close to one-to-one, on average.
There’s variance involved, even on the larger scale of a team basis. But it’s not much different from Pythagorean Record. Most teams are going to fall within a win or two of their Pythagorean Record. If a team exceeds (or falls short) of their Pythagorean Record by more than that, well, that’s when we start thinking about luck. Is it the same with WAR? The answer here isn’t clear. It might very well be luck. Or it might be just the percent that the metric falls short of capturing reality.
In the calculation above in the pull quote, 47.628 is what a team of replacement players would produce. In our WAR acronym, the “R” is for replacement. It’s Wins Above Replacement. So, you total up the individual WAR totals for all pitchers and hitters and then add the replacement WAR and you should come up with something pretty close to actual team wins. Let’s use the Mets of the past two seasons, one really good team and one mediocre team, and see how it checks out. This is using the FanGraphs’ version of WAR:
2022 – Batters (29.9) + Pitchers (21.1) + Replacement (47.6) = 98.6
2023 – Batters (18.6) + Pitchers (10.5) + Replacements (47.6) = 76.7
We don’t have decimal places in real wins, so let’s round here. The 2022 Mets exceeded their fWAR total by two wins. Meanwhile, the 2023 Mets fell short of their win total by two wins, assuming they get credited with a win in the suspended game, which by rule they should.
If WAR was perfect, the 2022 Mets should have received either one or two wins fewer than they did. Why one or two and not two? Refer back to the pull quote, where it said that a 100-WAR team would be expected to win about 101 games. But for simplicity’s sake, let’s just say it was off by two wins. In absolute terms, WAR gets us 98% of the way to the real answer.
That’s not perfect. But it’s really hard to argue that it’s not good.
It’s not quite as good with the 2023 team. This year’s fWAR of 77 wins is in absolute terms 97% of the actual total of 75 wins or 96% if MLB gives the Mets a loss (somehow) in the suspended game.
We use OPS and OPS+ a lot here, even if it’s a mathematical sin to add things with different denominators. Why do we do that? Because it’s easy and it gets us extremely close to the right answer. If someone’s OPS+ is 116 and their wRC+ is 118, you’ve pretty much nailed their production using the simpler metric.
Anyway, let’s do a comparison of some players on the Mets who saw significant time in both 2022 and 2023. The first number will be their 2022 fWAR and the second will be their 2023 output:
Pete Alonso – 3.8, 2.8
Francisco Lindor – 6.6, 5.9
Starling Marte – 2.9, (-0.3)
Jeff McNeil – 5.7, 2.5
Brandon Nimmo – 5.2, 4.4
The total for these five players in 2022 was 24.2 and in the just-concluded 2023 season, it was a 15.3 total. We can ballpark it that the drop in production from this quintet cost the Mets around nine wins, with around six of those coming from Marte and McNeil.
At the end of the day, we can hope for a rebound from this quintet. But even if they combine to give exactly what they did in 2022, it still leaves the Mets short of being a playoff team, much less a team to threaten 100 wins. It will be curious to see how much Steve Cohen and David Stearns do to add to the Mets in the offseason to make up the rest of the shortfall.
When your core players, Marte, McNeil and Alonso diminish their fWAR output like this it is very concerning. Two out of the three players have trading value. Maybe they should be traded for other major league players. A 74 win team has to be proactive to improve. Hope is not a strategy.
Why do you prefer fWAR to bWAR? Not a criticism, just wondering what’s the difference and why you use FanGraphs over B-R.
For 2023, bWAR has:
*Lindor at 6.0
*Senga at 4.5
*Nimmo at 4.0
*Alonso at 3.2
*McNeil at 2.3
Close, but a 10% difference for Nimmo.
The main reason to prefer fWAR is because it uses FIP as its main input for pitchers. When we’re trying to determine an individual’s contribution to wins, it makes sense to eliminate factors beyond their control, like the quality of the fielders behind them. B-R attempts to do that elsewhere in their WAR calculation but they’re essentially playing catch=up and never get there, in my opinion.
Another consideration is that when B-R applies park factors to pitchers, it does it to a league-average pitcher, rather than the pitcher in question, like fWAR does Also, fWAR uses 5-year park factors while bWAR uses 3-year.
On the batter side of things, the main difference is in the valuation of fielders. fWAR uses a combination of Statcast metrics and UZR, while bWAR uses DRS. For catchers, fWAR includes a component for framing while bWAR does not.
Finally, fWAR considers pitching to be 43% of the game, while bWAR considers it to be 41%.
Those are the ones that make a difference to me. But there are other differences as well. I’m sure you can Google it and find a list of all the differences between the two systems.
The WAR calculations are really close. That’s pretty cool.
I think Marte is a lost cause. Probably best just to get rid of him, unless it shows that his groin injury really hampered him. He may be a good reserve if that is the case.
The Mets need another big bat in the lineup and one that has better metrics than Alonso.
I am not sold on McNeil, yes he really provides flexibility, but, except for 2022, which was great, 2020, 2021 and 2023 were not good years for him. You would think that he would score more runs than he does. I was surprised to see he is in the 70+ number in runs in several years. He doesn’t steal bases, so, yeah, he is the squirel, but we need more than a squirel. He needs to hit well over .300 to have a real impact. If he doesn’t, he is a liability. We need him to either steal bases which he doesn’t or if he is going to hit .270 he needs to hit more than 20 home runs.
Mike, I really think McNeil should be the lead off hitter. It would allow Nimmo to move down and really make the lineup deeper. My biggest problem with McNeil is the amount of at bats he throws away by just flicking the ball, and not looking to drive it. He is pre-change Murphy and I do not know if he wants to change. McNeil’s production as a second baseman is fine, but lacking as an outfielder. It’s time to put him at second, put Mauricio at 3B, put Baty in LF and stop being cutsie with the player that has been successful to open up spaces for players that might be successful.
Brian:
Outstanding work
But I have come to expect no less