This says the Mets will win 73 games. I say nay.

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.

via The New Expected Final Standings | The LoHud Mets Blog.

So “Pythagorean records” are fun and all but here’s a few things.

– the baseball doesn’t have a memory. So take Pythagoras (and BABIP too) and stick that in your blue cap.

– Baseball is a game where you win and lose. A 10 run loss where you let some poor relief pitcher rot on the mound (aka “take one for the team”) means just as much as a one run loss in which your outfielders can’t yell “I got it.”

So to recap, if the Mets continue to play at a 22-20 pace they will finish 16 games under .500

Got it.

Howard Megdal the author of the article to which I linked posted the below in the comments.

Sorry Shannon, not what it says at all. Mets have outperformed their expected W-L. If all they do is match it from here on, they finish 73-89. If they exceed it the way they have so far- and they’d be doing so better than any MLB team, ever- they’ll keep playing 22-20 ball.

Let’s hope we are seeing that unprecedented miracle. But in the meantime, try not to miss characterize my work, if you wouldn’t mind

I don’t know what else to say other than what I said above. The ball doesn’t have a memory. A coin doesn’t remember that it landed on tails the last 7 times.

I’m happy to give Howard as much space as he’d like. Same for the BABIP crowd that finds Murphy to be lucky.

I don’t know if the Mets will play 22-20 ball all year or not. I would be surprised if they start playing 18 under. Run differential is not a stat that matters to me in baseball as each game is individual. Had Toronto scored 97 runs on Friday would have changed nothing except the Pythagorean record, which MLB has yet to add as a tiebreaker.


Dickey Should Send the Mets Bullpen a Fruit Basket -
Awesome Banner Day Banner Day approaches

6 Replies to “This says the Mets will win 73 games. I say nay.”

  1. Correct.  
    Pythagorean record suggests (note the not 100% definitive word suggest there) that the Mets were lucky over the last 42 games.  It doesn’t suggest very much at all about the next 42 games. 

  2. A win probability-based expectation (average WP over the course of each game, summed up) has the Mets expected winning percentage at .481, not quite as good as their actual .524 winning percentage, but significantly better than the .417 the Pythagorean expectation suggests.

    I don’t believe either is especially predictive, though, so to estimate the remainder of the season, you’re probably going to want to use a Pythagorean with excpected runs & expected runs allowed, rather than actuals or anything win probability-related.

  3. MetsPolice’s and my favorite team may be the ’84 edition.  I want all of you to go to baseball reference and check out their Pythagorean record.  no wait, I’ll do it for you……..

    90-72, Finished 2nd in NL EastManager: Davey Johnson (90-72)Scored 652 runs, Allowed 676 runs. Pythagorean W-L: 78-84

    The lesson?  sometimes the runs theory doesn’t accurately reflect the record – although most of the time it does but there are abberations!  lets tighten up the pen and keep winning the close ones!

Comments are closed.