November 17th, 2021 at 2:32 PM ^
Refs taking a TD off the board won't show up in that unfortunately (especially since in the stats it didn't even count as a fumble).
November 17th, 2021 at 3:29 PM ^
Exactly. Much of the "luck" they got won't show up in the box score (and hence won't show up in the exp win %).
November 17th, 2021 at 2:35 PM ^
Believe it was 59% for MSU
November 17th, 2021 at 3:11 PM ^
https://twitter.com/ESPN_BillC/status/1454846066969427969
That's the one I saw (Bill Connelly)
p.s. poor damn Nebraska
November 17th, 2021 at 3:35 PM ^
I wonder how expected turnovers is calculated. I figured it was just applying a 50% recovery rate to fumbles to even out any fumble recovery luck but Michigan had 3 fumbles and MSU had 1 in the box score so I would have thought that'd be 1.5 expected lost fumbles for M and 0.5 for MSU which combined with INTs would be 2.5 expected turnovers for both teams. Wonder if there's also an adjustment to expected INTs based on PBUs or something?
November 17th, 2021 at 3:53 PM ^
You're on the right track. Going off memory, I think you're right that fumble recoveries are assumed to be 50/50. There is an expected INT too, I think it's based on passes defensed (PBU + INT) and maybe an adjustment for QB pressure/sacks (what SP+ calls Havoc). I'm sure Connelly has explained it somewhere, but he's written so much and so many places I'm not going to look for it right now, I do know that I've read it before and it seems reasonable.
November 17th, 2021 at 2:40 PM ^
You can look up post game win expectancies for each team and game here: https://docs.google.com/spreadsheets/d/1q7T5HDF9Kp5SBpOqkDF_1dal3GYdPipaLSn75p2I-rg/edit?usp=sharing
November 17th, 2021 at 3:01 PM ^
Not updated yet for last week... do you know who updates this?
Not sure if they are scraping this or have a personal DB... I have looked for this before.
NP if you don't know.
November 17th, 2021 at 3:13 PM ^
It's Bill Connelly -- keeps all the links in his Twitter bio
https://twitter.com/ESPN_BillC
November 17th, 2021 at 3:52 PM ^
Thanks
Sincerely,
Old guy who deleted all social media (twitter included) after shit went down.
Just barely holding on to MGoBlog account.
November 17th, 2021 at 3:17 PM ^
Post-game win expectancy is a Bill Connelly statistic. I know he tweaked the formula a little mid-season this year. Note that it only has Michigan with 48.2 and 50.4 post-game win expectancy percentages for the Rutgers and Nebraska games.
November 17th, 2021 at 3:48 PM ^
Rutgers surprises me that it's under 50% for M given that M never dropped below a 77% win probability at any point in that game. Like I get that their yards per play ended up being better but Michigan also had the luxury of having a lead and being able to milk it (as frustrating as it was).
I'm a little surprised Nebraska isn't below 50%. Got a huge break in that game that Martinez coughed it up.
November 17th, 2021 at 3:51 PM ^
Yeah, I think Bill C's postgame win expectancy is blind to who is leading or trailing in the game (unless it hits 'garbage time') but rather just looks at the final statistics.
Rutgers is the one game I did not see this year and I haven't had much desire to go back and find it
November 17th, 2021 at 2:47 PM ^
i dont know, but since i didnt watch it live, i tracked it on espn and the win probability was just ridiculous.. almost as bad as the shitty punt game a few years ago..which went from 95%+ to 0 in about 12 seconds.
This time, we were in the 70-80%.. then it sunk.. then we got the ball back up 3.. and it went up again majorly, and then 1 play later it was in the 50s... and yet we still got more cracks at it and failed, but they basically were very low probability (in game) until very late.
Why cant we just lose those games to indiana and purdue, and beat the rivals?!
November 17th, 2021 at 2:51 PM ^
I’d prefer if they just didn’t lose any games if we are choosing
November 17th, 2021 at 3:50 PM ^
That's exactly what Lloyd Carr did: lose to Illinois and Purdue but beat rivals. No one loses to IU though (except in a pandemic year).
November 17th, 2021 at 4:34 PM ^
lloyd didnt beat osu towards the end. i thought IU was a football school now!!!
November 17th, 2021 at 2:56 PM ^
It was high, then the refclowns reviewed it and dropped it to 0%.
November 17th, 2021 at 3:00 PM ^
MSU's explosiveness in that game apparently made up for a lot of other stuff. They were the post-game favorites.
November 17th, 2021 at 3:14 PM ^
We were at around 90% win probability going into the 4th I remember. Last time I look to that to give me peace of mind
November 17th, 2021 at 3:51 PM ^
well, one out of ten of those times...disaster will strike
November 17th, 2021 at 4:00 PM ^
MSU has won 10 of the last 14 or 15? I forget exactly, but it seemed nuts to me to overly confident, 59%, of a Michigan win when both teams were undefeated after 7 games and the game was at MSU. They made the plays to win.
November 17th, 2021 at 4:49 PM ^
Slightly changing the discussion, my expected win % vs Sparty is 75%; 50-50 in EL and every single damn year in Ann Arbor!
November 17th, 2021 at 5:42 PM ^
That's where my mind went with the title as well. In my head it's 2 out of 3 times, so pretty close to yours. Sadly far below that of late.