Score Prediction Site

mcbaseball10 · 02-18-09, 01:09 PM

Very interesting. Have any tracking info. as far as how the prediction results faired ATS and SU?

coreygman · 02-18-09, 01:14 PM

kewl site, so do you have stats on how the algorithm is working? I could sit down and figure it out, but thought I would ask first

"Sic" · 02-18-09, 01:15 PM

sweet...I like the predictions on Miami FL and Ohio St

KJ90 · 02-18-09, 01:22 PM

tracking info
would be nice.

PittChicGambles · 02-18-09, 02:10 PM

Hey mugu...interesting site...I actually crosschecked my early leans with your system, and out of 12 of my ncaa leans this morning, your system matched 10 of mine. If this system has any type of success, I would love to know. Could mean I may have a good evening..any more info you have on this system would be appreciated! Welcome to the forum

Mugu14 · 02-18-09, 02:31 PM

I've only been logging the predicted scores for a little over a week, so I don't have a lot of historical data to go on. You can see all of the past individual scores by looking at past dates in the scoreboard.

Picking every D1 game straight up since February 10, it has gotten 71% correct. (348 games). I'm actually not sure how bad or good this is, and as I said before only about a week of games so take with a grain of salt.

I don't have any data on old spreads, so I'm not sure how accurate it has been ATS. If someone knows of a site that I could look at to find spreads for old games, I could look

dchecks · 02-18-09, 02:40 PM

There is a site that does this also, it is www.kenpom.com

See if you guys match up!!!

GatorFan · 02-18-09, 02:43 PM

Good stuff Mugu

junkman773 · 02-18-09, 02:45 PM

Nice Site

Mugu14 · 02-18-09, 02:45 PM

I should also add that the site only uses past scores to make predictions, it doesn't take into account injuries, etc. So it is likely to do very poorly picking a team if one of their star players just got injured/suspended or there is some other reason to think they'll be playing a lot differently then they have in the past.

007Fatty · 02-18-09, 03:55 PM

wow nice setup man.
it would be a good one for doing ml too.
because everygame lately has been correct for the winner
even if the score isnt the same
there has only been a couple that were wrong.
nicee

curious · 02-18-09, 04:10 PM

Originally posted by Mugu14

I was working on a project that predicts college basketball games using a computer learning algorithm. I put up a site that has its results and thought that some people here might find it interesting. The site is entirely non-commercial, but I would appreciate any feedback.

Use only for your entertainment, no warranty is implied on any picks.

What would be interesting is to see your predictions posted alongside the spread.

jerseykai · 02-18-09, 04:11 PM

nice site, can you explain to me how it works, to be more precise, what the score predictions are based on?, i think i am going to start to learn programing, it could come in handy to have my own stats, know that the validity is correct, and make more money...muhahahahahahaha

curious · 02-18-09, 04:42 PM

The plays that have value on 2-18-09 according to PA-Ranker

Okay, so I compared PA-Ranker's predictions to the actual line. Assuming Pa-Ranker is right and the lines makers are wrong then these teams have value in the lines set by the books, in order to make this list the line had to be off by 3 or more in our favor:

Old Dominion Win
George Mason Loss
Bowling Green Win
Miami OH Loss
James Madison Win
Towson Loss
Delaware Win
Miami FL Loss
Central Michigan Win
Houston Loss
Georgetown Win
West Virginia Win
Western Michigan Loss
Illinois State Win
Florida Loss
Georgia State Win
Louisville Win
Georgia Lost
Butler Lost
Air Force Lost
Kansas Lost
Va Tech Lost
South Carolina Lost
Mississippi Win
WRight STate Loss
SMU Lost
Green Bay Win
UAB Lost
Ohio State
Nebraska
Boise State
UNLV
Pacific
Long Beach State
CS Northridge
Tennessee Martin Win

One Typo
Indiana State was supposed to be on this list but I had Illinois State listed twice by mistake. Indiana State Win

KJ90 · 02-18-09, 04:50 PM

curious no plays today?

mcbaseball10 · 02-18-09, 04:51 PM

Think this could be a really good tool with the help of somebody like you Curious, that is good at sifting through the numbers. Like your idea here posting all those with value and start monitoring the success...
Cap the games you like, compare it to this list and play the ones you feel confident in.

PittChicGambles · 02-18-09, 04:53 PM

Thats what I did **...I capped this morning, and came up with a list and then compared them to the system and placed my bets...What was great too, was that 10 out of my 12 early leans matched the system as covers...Im going to see how this pans out for me tonight...good luck to you too!

curious · 02-18-09, 04:54 PM

Originally posted by KJ90

curious no plays today?

No, I am busy and it is getting too tough to find value. I'm taking a few days off.

KJ90 · 02-18-09, 04:59 PM

iight

mcbaseball10 · 02-18-09, 05:00 PM

BOL Pitt, I by no means am a good capper so I look for any POTENTIAL edge I can get. Just don't over extend yourself on a new model with a limited sample size.

PittChicGambles · 02-18-09, 05:01 PM

Oh im not..no way did i even play half of my leans...right now I picked 5 solid picks....thanks for the advice **...

MicGan · 02-18-09, 05:04 PM

This site is actually pretty close on the numbers. I'd traced and dated it back past week wagers, they are 90% right on the money. Some of the game I'd wager in and didnt cover and it shown. The only game they are off, are rank teams or national TV games. I can't blame them, National televised game always hard to cap them. We all know that.

Over all this site has great info....... It's a great tool.

SportsLockPicks · 02-18-09, 05:07 PM

Louisville by 37 over Providence, that a lot of points to a team that scores a lot in Providence...lets see what happens on that one.

SportsLockPicks · 02-18-09, 05:32 PM

..

hajune · 02-18-09, 05:57 PM

Originally posted by Mugu14

I've only been logging the predicted scores for a little over a week, so I don't have a lot of historical data to go on. You can see all of the past individual scores by looking at past dates in the scoreboard.

Picking every D1 game straight up since February 10, it has gotten 71% correct. (348 games). I'm actually not sure how bad or good this is, and as I said before only about a week of games so take with a grain of salt.

I don't have any data on old spreads, so I'm not sure how accurate it has been ATS. If someone knows of a site that I could look at to find spreads for old games, I could look

Go to gold sheet.They have 15 years worth

Mugu14 · 02-18-09, 06:18 PM

Originally posted by SportsLockPicks

Louisville by 37 over Providence, that a lot of points to a team that scores a lot in Providence...lets see what happens on that one.

Haha, that does seem like a lot. Like I said before, the algorithm only uses past scores and won't make adjustments that would seem obvious to humans. (Like for example, if Louisville really was up 25, there's a good chance they might call of the dogs a little bit instead of pushing it into the thirties).

Like some other posters have said, I wouldn't take the picks to be gospel. Just an interesting source to look at.

curious · 02-18-09, 06:28 PM

Originally posted by Mugu14

Like some other posters have said, I wouldn't take the picks to be gospel. Just an interesting source to look at.

What? You tell us this AFTER I bet 5 dimes on each one of your picks?

CashMoney71 · 02-18-09, 07:57 PM

How do we know this isn't a hoax? There's no proof that he's using an algorithm other than his word! Do not bet your hard earned money based off of this website! It could just be a game to him!

Hells Kitchen · 02-18-09, 08:20 PM

One MAJOR problem with the site is that game predictions are done for the next 4 games of each team.

What about injuries etcetera?

mcbaseball10 · 02-18-09, 08:22 PM

Originally posted by Mugu14

I should also add that the site only uses past scores to make predictions, it doesn't take into account injuries, etc. So it is likely to do very poorly picking a team if one of their star players just got injured/suspended or there is some other reason to think they'll be playing a lot differently then they have in the past.

Already made this statement earlier

Hells Kitchen · 02-18-09, 08:26 PM

Thx mcbaseball.
Just did a quick scan of the thread and missed your post

mcbaseball10 · 02-18-09, 08:29 PM

No problem

Mugu14 · 02-18-09, 09:32 PM

The future projections will also change everyday to account for games being played. They are locked the day that they are played, so past games reflect the projection that it made that day, but the future ones will continue to change until the day they are played.

curious · 02-18-09, 09:46 PM

Originally posted by CashMoney71

How do we know this isn't a hoax? There's no proof that he's using an algorithm other than his word! Do not bet your hard earned money based off of this website! It could just be a game to him!

Now you tell me after I have $150 large riding on his picks. Oh my gawd, what if his dog told him these plays???

I don't really care as long as we win.