This Has My Attention: New Questions

Actual wins and losses sum to 81*12 = 972

Pythagorean wins and losses do not necessarily sum to 972. They vary from league to league depending on how the runs scored and runs allowed vary from team to team.

Check out baseball reference and you'll see that the pythagorean W/L predictions for the 2022 season also do not sum to 81*15 in either league.

To be clear, I don't support barrmorris in his assertion that his analysis disproves your point. I don't think your unlucky. I think the good teams you put together won't achieve the pythag predictions.

I’m pretty deep into an overly heavy serving of Johnny Walker Black. But I’m pretty sure actual wins and losses sum to 162 x 12 = 1944. The same as the Pythagorean records.

But at this point, I’m open to missing some nuance.

To be clear, I don't support barrmorris in his assertion that his analysis disproves your point.

I never made such an assertion. In fact my first post (which looked at average errors for long-time managers) said this:

"This doesn't prove anything with regard to whether SOM has its thumb on the scales, but it does satisfy me that pythagorean expectations are not the proof that this is occurring."

My more recent post was in response to this statement from J-PAV:

"What I continue to be focused on is (it appears to me) that there was a time when expected win totals fell generally within +/- 3, and there were balanced outliers on either side. Regardless if the pythag exponents are optimized, or if the current formula treats high and low run diffs disproportionately, I think we should still see balanced tails approaching a bell curve."

I was simply showing that the pythagorean estimator errors do appear to have a normal distribution.

That I can agree with.

I’m pretty deep into an overly heavy serving of Johnny Walker Black. But I’m pretty sure actual wins and losses sum to 162 x 12 = 1944. The same as the Pythagorean records.

But at this point, I’m open to missing some nuance.

Here's the first league you posted

https://365.strat-o-matic.com/league/462784

Do the math - or wait until tomorrow morning - pythagorean wins sum to 973 and losses sum to 971.

Okay, that is pretty interesting!

I can’t square in my brain how this is possible. The difference between 973 and 971 remains quite small though, although it’s weird that there even is an inconsistency, if you can call it that.

If we could identify the records of the managers on the tails specifically and not universally, we would have some interesting data. In other words, if the left (negative) tail was filled with outperforming managers, and the right (positive) tail was filled with underperforming managers, then your conclusion that good managers will underperform expected wins would obviously hold true.

Having said that, the number of times good managers have negative run diffs and lesser managers have positive ones could make seeing the forest through the trees a little difficult.

The larger question remains: “Why don’t good teams meet the Pythagorean projection?”

I can’t square in my brain how this is possible.

J-pav, it is because you are viewing these approximations as inviolable laws. Pythag is an approximation. Closed-end stats formulas are approximations. So when you look at the fit of real data to the models, you have to accept when they are "good enough" fits.

If you won't accept it, then to understand how this is possible you will need to dig a little deeper to understand these formulas and how they are derived. There is no way to disprove a random hypothesis, but djmacb and barrmorris both addressed different aspects of your concerns to the limit of what is knowable.

Or like a few others have posted...just enjoy the game without all these sterile stats nerdiness
Actually the larger question is why JW Black

Blue is for special occasions!

I read a Stephen King quote a long time ago that always stuck with me. He said something to the effect of “true horror isn’t when a monster jumps out at you in the dark. It’s when you measure the four corners in a room without mistake and the angles total 359 degrees.”

The larger question remains: “Why don’t good teams meet the Pythagorean projection?”

Actually, the larger question is: “Why have I been so uneasy about how the game appears to be functioning?” I have largely attributed it to Bullpen 3.0, but underperforming pythag is kinda shouting off the page at me. I thought it was worth a look.

I hear what you’re saying and have returned to my default chill mode.

Ha. Hey all good questions No need to avoid that.
And nothing wrong with JW Black...Red would be a different story!
Balvenie double wood is nice for a mainstay sipper too