coincidence

random. black box, or just bad luck? I've got 14 teams going. out of the 14, 1 team is 2-1 with exactly even dice, 4 teams have positive dice and 9 have negative dice with a couple with over well -100 rolls as the season is ending. 1 team somehow made the playoffs last night with the other 3 having season positive dice. about tired of pissing into the wind

I know

i have similar experience with "expected" wins vs actual

i get screwed about 2/3 of the time, and it's cost me several playoff appearances

i know another GM (won't name out of respect) mentioned they had an unusually long recent streak of playoff teams not winning a title, over 15 i think

somewhere, some place, some GMs must be getting good luck, although that place is so far away and unfamiliar to my teams, we can't even see it nor even imagine it

what is "expected wins " based upon? HAL's Pen management? or any other team setting? or is it based on 50/50 dice?

it's the Black Box!!!!

It's nothing mysterious. It's the same thing you'll see published in any expanded standings for MLB. It's based on the Pythagorean expectation formula. In this case it is Expected Win PCT = RS^2/(RS^2+RA^2). A more accurate exponent for baseball is in the 1.7-1.8 range (other sports use other exponents). In any case it's just a statistic that is correlated with win pct. Some time ago, I measured the Root Mean Square Error for the statistic in SOM baseball over about 1000 leagues and found it similar to the RMSE for the statistic in MLB.

how is this calculated before the 1st game is played?

how is this calculated before the 1st game is played?

It's not. Where are you seeing it calculated before the season starts?

"It's based on the Pythagorean expectation formula. In this case it is Expected Win PCT = RS^2/(RS^2+RA^2). A more accurate exponent for baseball is in the 1.7-1.8 range (other sports use other exponents). In any case it's just a statistic that is correlated with win pct. Some time ago, I measured the Root Mean Square Error for the statistic in SOM baseball over about 1000 leagues and found it similar to the RMSE for the statistic in MLB."

Now I know why I keep getting my ass kicked

Artie

I'm not seeing it B4 the season starts. that's the point.

score......

Hal doesn't look at a GMs career win %, and the makeup of his team compared to others in the league/division, etc, and then make a prediction of "expected" wins....that's not how it works

to my knowledge, it's calculating the aleady played games, and basically the run differential.....that theory others noted, just uses an algebraic equation using the same data, and it comes up with roughly the same analysis

meaning, if your team for example after half the season (81 games), had around a +50 overall run differential, to me, that seems to be around 5 or 6 games over .500 (actual record)....meaning, the "expected wins" for a team having played half the season, and with that +50 differential, would be around 5-6 games over .500 (i'm ballparking here, but think Im fairly close)....take it to a further extent, and if having a +100 differential after that many games, the team would be expected to be about 10-12 games over .500 or thereabouts

so, if your team had the first total, +50 point differerential, and your actual wins in the league was more or less than that number, it's arguably one of the primary factors determining if your team has been lucky, unlucky, or have mostly equal luck (if actual wins are near expected wins)

that was the issue I lamented when first replying to your post

in summary, the expected wins can't be calculated until games have been played, because they are in fact based on those game results...and, the more games played, the more weighty the data, because it's only possible to be a few games up or down in the season's first week or two, so by season's end, those numbers (actual vs. expected) mean the most