Computer/System Day 6

[Cubbiefan16]

1-0 yesterday on a 15+, below are the overall totals.

15+ differences: 3-2
10-14 differences: 5-2
7-9 differences: 3-4

Several 7+ plays for Saturday; I will only be playing the top 5 plays, but here they all are. If anyone wants the other scores from today, just let me know and I'll be happy to post them. Not guessing anyone does, but just in case..


Tennessee 92 @ Temple 64 --> Tennessee -7 (21 point diff)

Seton Hall 82 @ St Peter's 51 --> Seton Hall -12 (19 point diff)

Eastern Kentucky 76 @ Ball St 63 --> Eastern Kentucky +4 (17 point diff)

Northern Iowa 66 @ Wyoming 84 --> Wyoming -3 (15 point diff)

San Diego 57 @ Oregon 79 --> Oregon -7.5 (14.5 point diff)

Mississippi 83 at New Mexico 73 --> Mississippi +4 (14 point diff)

NC Greensboro 59 @ East Carolina 86 --> East Carolina -14.5 (12.5 point diff)

Samford 52 @ St Louis 55 --> Samford +13 (10 point diff)

Saint Mary's CA 75 @ San Diego St 62 --> St Mary's -3.5 (9.5 point diff)

Oral Roberts 60 @ North Carolina 105 --> North Carolina -36 (9 point diff)

Utah 75 @ Oklahoma 79 --> Utah +12.5 (8.5 point diff)

Virginia Commonwealth 80 @ Richmond 69 --> Va. Commonwealth -3.5 (7.5 point diff)

Nebraska 70 @ Oregon St 55 --> Nebraska -7.5 (7.5 point diff)


That's it.. Should be a fun Saturday. Good luck to whoever you play

[mikkii10]

hey cubbiefan how about ohio st. vs butler?

[The Frozen Pond]

Let's do this.

[gandolph]

Good Luck today CF!

[Cubbiefan16]

Sorry mikki...I had Butler 59 @ Ohio St 62

[Cubbiefan16]

Frozen Pond - Let's go man...Tennessee needs to pick it up

gandolph - Thanks man

[The Frozen Pond]

Originally Posted by Cubbiefan16

Frozen Pond - Let's go man...Tennessee needs to pick it up

gandolph - Thanks man

One of the sloppiest, most undisciplined teams I have ever watched. And what's really gonna piss me off is they'll win the game, but not cover.

Hate the Vols. With a passion.

[ScreaminPain]

Originally Posted by Cubbiefan16

Northern Iowa 66 @ Wyoming 84 --> Wyoming -3 (15 point diff)

You may want to be careful with your numbers on Wyoming. They have played a brutally easy schedule which has skewed their performance figures. Using figures in a computer, as you're doing, can be misleading on occasion.

Just sayin'......

Northern Iowa is nothing special and I don't have a dog in this fight, however I looked closely at this matchup last night before chucking it in.

Keep up the effort, I think it will pay off
good Luck on your action

[The Frozen Pond]

Ha, nevermind. They suck so bad they can't even win outright against a crappy Temple team.

Nice start to the day.

[Cubbiefan16]

Originally Posted by The Frozen Pond

One of the sloppiest, most undisciplined teams I have ever watched. And what's really gonna piss me off is they'll win the game, but not cover.

Hate the Vols. With a passion.

Way too many forced outside shots in the first half...that's what kept them from scoring like they should..

Originally Posted by ScreaminPain

You may want to be careful with your numbers on Wyoming. They have played a brutally easy schedule which has skewed their performance figures. Using figures in a computer, as you're doing, can be misleading on occasion.

Just sayin'......

Northern Iowa is nothing special and I don't have a dog in this fight, however I looked closely at this matchup last night before chucking it in.

Keep up the effort, I think it will pay off
good Luck on your action

Thanks for the concern SP, it really is appreciated. The system takes into account all of that though, adjusting offensive statistics and defensive statistics to opponents strengths and weaknesses. It is actually almost doubly accounted for, with the second being a lot less severe than the first...I found that that was the best way to integrate those statistics.. It still has a little tweaking needed, but I am confident with the scores it produces..

[reggie02]

What variables do you use for the model? Does it understand home court or do you just assign a number like 3 or something.

[revnecro1273]

Originally Posted by reggie02

What variables do you use for the model? Does it understand home court or do you just assign a number like 3 or something.

i remember reading in one of his threads that home court is taken into consideration

[Cubbiefan16]

Originally Posted by reggie02

What variables do you use for the model? Does it understand home court or do you just assign a number like 3 or something.

It understands home court well. It's amalgamated in all the scores; and if a team seems to play significantly better at home or on the road, the system gives them an increased or decreased advantage depending on where the game is played.

It uses almost every variable imaginable in the best form that I have found possible. It even takes a team's oppenent's opponent's defensive and offensive efficiency into consideration as well. It took me a long time to design but right now I'm happy with it. But like I said, it still needs a few tweaks and there is definitely some room for improvement.


Update today:

15+ plays: 2-1
10-14.5 plays: 0-0
7-9.5 plays (not recommended): 0-1

Good luck the rest of day guys

[reggie02]

So is it a regression model based on a set of games already played? I built something similar for ncaa football where I use 3000 games and also opponents opponents. I standardize all the variables and use a simulator that I bought a license for years ago. If you are using aregression or something similar with games played, I would use a variable for home field depicted by a 0 or 1, that way the model will understand home field as its seperate variable.

I tried building something for ncaa Basketball back in the day and was very disappointed with the results. I am planning to build another one now that I actually know what im doing and was just curious as to what stats you use. I will be testing for significance, but if you can tell what variables seem to be the most important, that would be great.

Good luck

[Cubbiefan16]

Originally Posted by reggie02

So is it a regression model based on a set of games already played? I built something similar for ncaa football where I use 3000 games and also opponents opponents. I standardize all the variables and use a simulator that I bought a license for years ago. If you are using aregression or something similar with games played, I would use a variable for home field depicted by a 0 or 1, that way the model will understand home field as its seperate variable.

I tried building something for ncaa Basketball back in the day and was very disappointed with the results. I am planning to build another one now that I actually know what im doing and was just curious as to what stats you use. I will be testing for significance, but if you can tell what variables seem to be the most important, that would be great.

Good luck

I'll put together an explanation of some of the important factors and give you what I know some time later, maybe tomorrow. Don't have the time right now..


revnecro - How's what you were working on using these plays going today?