Attn sys and _FAN (kenpom related)

[Stevesie60]

For yesterday's game, kenpom predicted us to lose by one. Since we won by one, does that system just take into account that it was off by two points, or does the fact that we won count for something more?

For instance, if we were predicted to win by 3, and we instead won by 5, would it be adjusted the same amount?

[sys]

For yesterday's game, kenpom predicted us to lose by one. Since we won by one, does that system just take into account that it was off by two points, or does the fact that we won count for something more?

For instance, if we were predicted to win by 3, and we instead won by 5, would it be adjusted the same amount?

technically, neither.  but the "take point diff. into account, ignore wins and losses" is close enough to not worry about the diff.



btw, ol' pommsey is now predicting 7-9 with a better shot at 8-8 than 6-10.  looks like pullen's last minute heroics impressed someone.

[pike]

For yesterday's game, kenpom predicted us to lose by one. Since we won by one, does that system just take into account that it was off by two points, or does the fact that we won count for something more?

For instance, if we were predicted to win by 3, and we instead won by 5, would it be adjusted the same amount?

technically, neither.  but the "take point diff. into account, ignore wins and losses" is close enough to not worry about the diff.



btw, ol' pommsey is now predicting 7-9 with a better shot at 8-8 than 6-10.  looks like pullen's last minute heroics impressed someone.

We're gonna go dancing!

[Stevesie60]

For yesterday's game, kenpom predicted us to lose by one. Since we won by one, does that system just take into account that it was off by two points, or does the fact that we won count for something more?

For instance, if we were predicted to win by 3, and we instead won by 5, would it be adjusted the same amount?

technically, neither.  but the "take point diff. into account, ignore wins and losses" is close enough to not worry about the diff.



btw, ol' pommsey is now predicting 7-9 with a better shot at 8-8 than 6-10.  looks like pullen's last minute heroics impressed someone.

Yeah, I thought it was closer to the point diff thing, so that's why I was confused that we got bumped up.

[sys]

Yeah, I thought it was closer to the point diff thing, so that's why I was confused that we got bumped up.

he didn't really bump kstate much (they did get a dece bump from the nu win).  but they won two games that has a cumulative win% of somewhere around 110%.  so the conf prediction improved by about 2 - 1.1 = 0.9 wins.

[michigancat]

btw, ol' pommsey is now predicting 7-9 with a better shot at 8-8 than 6-10.  looks like pullen's last minute heroics impressed someone.

[Stevesie60]

Yeah, I thought it was closer to the point diff thing, so that's why I was confused that we got bumped up.

he didn't really bump kstate much (they did get a dece bump from the nu win).  but they won two games that has a cumulative win% of somewhere around 110%.  so the conf prediction improved by about 2 - 1.1 = 0.9 wins.

Got it. Thanks, sys.

[06wildcat]

For yesterday's game, kenpom predicted us to lose by one. Since we won by one, does that system just take into account that it was off by two points, or does the fact that we won count for something more?

For instance, if we were predicted to win by 3, and we instead won by 5, would it be adjusted the same amount?

It's already been answered but I'll throw in some more detail.

Kenpom predictions (the final scores) and forecast (percentage that team x wins) have to be looked at separately and as a whole. So Kenpom's prediction yesterday was a 71-70 ISU win iirc. He forecast ISU to win 51 percent of the time.

A prediction is a statement about a future event. The forecast is the probability of that event coming true. Most people are more focused on the former than the latter, which is how Kenpom judges the veracity of his statistics.

He really doesn't care about the point spreads posted, but if he shows 100 teams with a 51 percent chance of winning and 70 win he's got a problem. If 45-55 of those teams win he's right in the sweet spot, which is true on most occasions.

[Stevesie60]

For yesterday's game, kenpom predicted us to lose by one. Since we won by one, does that system just take into account that it was off by two points, or does the fact that we won count for something more?

For instance, if we were predicted to win by 3, and we instead won by 5, would it be adjusted the same amount?

It's already been answered but I'll throw in some more detail.

Kenpom predictions (the final scores) and forecast (percentage that team x wins) have to be looked at separately and as a whole. So Kenpom's prediction yesterday was a 71-70 ISU win iirc. He forecast ISU to win 51 percent of the time.

A prediction is a statement about a future event. The forecast is the probability of that event coming true. Most people are more focused on the former than the latter, which is how Kenpom judges the veracity of his statistics.

He really doesn't care about the point spreads posted, but if he shows 100 teams with a 51 percent chance of winning and 70 win he's got a problem. If 45-55 of those teams win he's right in the sweet spot, which is true on most occasions.

 

I'll be sure to include you in the title of my next thread in which I inquire about kenpom.

[kso_FAN]

Those guys nailed it.

To be honest, I don't focus on (or care much really) for the predition parts nearly as much as taking the box score numbers and turning into more informative stats like kenpom does.

[Stevesie60]

Those guys nailed it.

To be honest, I don't focus on (or care much really) for the predition parts nearly as much as taking the box score numbers and turning into more informative stats like kenpom does.

Except he still doesn't have 95.

 

[kso_FAN]

Those guys nailed it.

To be honest, I don't focus on (or care much really) for the predition parts nearly as much as taking the box score numbers and turning into more informative stats like kenpom does.

Except he still doesn't have 95.

 

He just hasn't played enough to register even in the "limited roles" list of the roster. He has to get to over 12% possessions used before he makes the bigtime of kenpom stats.

[06wildcat]

Those guys nailed it.

To be honest, I don't focus on (or care much really) for the predition parts nearly as much as taking the box score numbers and turning into more informative stats like kenpom does.

Except he still doesn't have 95.

 

Hey at least Gruds is one of the best role player in the nation.