Over the course of a large number of games that stuff generally evens out. It just doesn’t account for very many runs. The eight-innings-for-the-home-team effect only matters if a team has way more home wins than away wins and even then it’s not that many innings. 20 innings a year - 20 wins would be an enormous difference in home vs. road wins - would account for, on average, nine runs.
And not very many games are won by walk-off hits… and games that do partially make up for the eight inning effect.
Pythagorean projections are generally very accurate and teams that over- or under-perform them tend to veer back towards their expected record with remarkable predictability.
Actually, it’s home wins, period, that matter. Barring extra innings, losing team plays nine innings whether it’s home or away, and a winning team plays either eight or nine. In other words, a team that loses more games is going to play more innings, whether they have a significant difference between home and away wins or not.
I hate to crunch more numbers after I’ve been doing it all day, but now I’m curious. Let’s take Average Team at 81-81, with 41 wins at home and 40 away, and a Natinals 162-game projection at 42.5-119.5, with a home/away split of 25.5 wins at home and 17 away. And let’s say for ease of calculation that teams that win at home always play eight innings, and there are no extra-inning games.
Average Team therefore plays 328 innings in home wins, 360 in home losses, 360 in away wins, 369 in away losses: 1,417 innings in total. The Natinals would play 204 innings in home wins, 499.5 in home losses, 153 in away wins, and 576 in away losses: 1,432.5 in total. So, that’s a difference of about 15 innings. Agreed, not that much of a swing.
But that’s only half of the total. See, the Natinals are playing more innings, and therefore have more chances to score. The other half of the equation is that they are facing fewer innings, and therefore the other team has fewer chances to score against them. Running the numbers again, it looks like the Natinals face another 15 fewer innings.
Put it all together and…wait, I did all that and there’s a change of only 30 innings per year? Damn. Let’s see, league average of runs per game is running at 4.65 per game, average length of game this year is 8.91 innings per game (both from this Baseball Reference page), so that gives us a swing of…15.64 runs per year. Applying that to the portion of the season the Natinals have so far played this year (62/162), that’s about 6 runs, which gives them a new Pythagenport winning percentage of…let’s see, X = .45 + 1.5 * log10 ((rs+ra)/g), winning percentage then equal to (rs^x)/(rs^x + ra^x)… .430, as opposed to .434. That’s one less win over the course of a year.
And now, time to sleep, knowing I was wrong, but at least knowing why.
Having said all that, the Natinals still have a chance at antiglory. Their REAL record is brutal, and it’s gotten a little worse since this thread started. They’ve gone 1-4 since my OP, making them now 16-46, so are on a pace for 120 losses! (42-120, not actually as bad as the 40-120 Mets of 1962, but 120 losses is 120 losses.)
I’m optimistic they can play worse and make a run at it.
Let’s see, league average of runs per game is running at 4.65 per game, average length of game this year is 8.91 innings per game (both from