i just lost 45 minutes worth of response, and I don’t have the energy to type it again. Anyone know a way to make that “feature” go away?

Anyway, I suggested a way to try to remove luck over games played so far from the “future wins” calculations by looking at current record*strength of schedule (if you think SoS matters) + (future wins determined using something like FIP and baseruns). Or take it a step further and derive a FIP adjusted to account for team defensive skill using DER (someone has probably already done this).

Your formula attempts to do this by looking at run differential, but RS and RA are both functions of luck…

Also my gut feel is that “recent play” means diddly, but I could easily be totally wrong about that. If anything, recent play might be a negative, since there’s a decent chance it means your ace just pitched and won’t again for 4 days.

the problem is that too many online power rankings take in recent performance. A bad week should definitely have an impact, but too much of the time it just slaughters a team’s power rankings.

because it makes us look more like underdogs, and we usually play better as underdogs

First off, I made a typo in the chart. San Diego’s percentage should be .397, everything else is the same.

Washington is not listed as the worst team because they’d won 6 of 10 and they are underperforming their pythagorean record. Arizona and San Diego had each lost 8 of 10. Washington was trending up while those two were trending down, and that’s where the results come from.

At the time of the posting, the Cards were in a 3-way tie for first, so it’s not outrageous that they aren’t ranked higher. Is 87 wins too far off from what we have expected?

Recent performance is what may give some indication of what the team will do in the immediate future. Recent performance also may take into account recent injuries to key players. I used the teams’ records in their last 10 games so it’s not like teams will get bumped up or down considerably from a 3-4 game winning or losing streak when it’s weighted at 10%. However, teams that slump badly will fall down the rankings, as they probably should.

Strength of schedule is not that important. The reason I weight it relatively heavily is that SOS clusters around .500 for every team. This means that it will bring slumping teams’ overall ratings up and hot teams’ overall ratings down.

As for the games played not mattering in predicting future performance, I disagree. Obviously just because X team has won at Y percent so far doesn’t mean that they will continue at that pace, but it’s fairly safe to say that the proportion won’t be too far off in most cases this far into the season. The run differential and SOS are both there to help balance out the numbers for the outliers.

I think it might be a good idea to add something that takes into account batting luck to what I already have. Perhaps I could take the mean team BABIP and compare it to each individual team’s BABIP. Something like (Power now) + (.2)(Mean BABIP – Team BABIP). This would raise the score for unlucky teams and lower it for lucky teams. The only question is, what multiplier should I use? Looking at both ends of the spectrum, a .2 multiplier would lower the Dodgers win total from 98.4 to 97.4, and raise the Padres win total from 64.3 to 65.3. I like this idea, but what multiplier do you think I should use? And should the multiplier get smaller as the season goes on?

Honestly, I don’t know how to account for pitching or defense. Pitching BABIP won’t really work because that’s a product of defense. The pythagorean win expectancy is about as close as I can get without some help. I don’t necessarily think that anything needs to be adjusted for this, but if there is some relatively painless way to do so, I’m all ears.