[hkarmy_kenny]

December 13, 2012

Day 2 of integrating a new mathematical model into the plays. Model is based off of possession algorithms and factoring in the four main factors of defense/offense to output scores. These scores are also are adjusted through SOS factors, pace variance and then compared with power ratings/vegas implied/decision tree.

Model Record: 7-4
Best Bet: 1-1

------------------------------------------------------------

Wichita St. -1 at Tennessee [LOSS] [SCORE 60-69]
[BEST BET] Wichita St. Wins By 6

Washington atSeattle +8 [LOSS] [SCORE 87-74]
Washington wins by 7

Middle Tennessee at Belmont -1.5 [WIN] [SCORE 49-64]
Belmont wins by 2

Idaho State at CSU Fullerton -15.5 [LOSS] [SCORE 53-66]
CSU Fullerton wins by 20+


*These plays do not factor in injuries


Yesterday's thread: http://www.sportsbookreview.com/forum/ncaa-baske...12-2012-a.html

Good luck

[bane]

kenny, any reason you're not just keeping one main thread? I'd subscribe if you did but I know some folks just like to post randomly.

[hkarmy_kenny]

kenny, any reason you're not just keeping one main thread? I'd subscribe if you did but I know some folks just like to post randomly.

I'm going to start keeping a main one after today, thanks

[Louisvillekid1]

Well looks like your going to go 4-4 in beating the closing number, so You already won in my book.

[Louisvillekid1]

Don't mean to double post your thread, but i just read this more closely and have a few questions?

I noticed you predict Belmont to win by 2 and bet them -1.5 as well as playing Seattle +8, while predicting a 7 point Huskie win.

My question is, 0.5 - 1 point edge from your system is enough reason to make a play?

I'd think if you are trusting this system, you'd look for a bigger edge then a point.

Anyway Good Luck

Also: If you trust you system to the point, then a play on Wch St or Cal St Full, should be bigger plays than the other too, because you see a much bigger edge...

[hkarmy_kenny]

Don't mean to double post your thread, but i just read this more closely and have a few questions?

I noticed you predict Belmont to win by 2 and bet them -1.5 as well as playing Seattle +8, while predicting a 7 point Huskie win.

My question is, 0.5 - 1 point edge from your system is enough reason to make a play?

I'd think if you are trusting this system, you'd look for a bigger edge then a point.

Anyway Good Luck

Also: If you trust you system to the point, then a play on Wch St or Cal St Full, should be bigger plays than the other too, because you see a much bigger edge...

I appreciate your questions

This is the 2nd day of trying out this model and I honestly didn't do nearly enough backtesting to warrant this as a profitable model. I truly believe that simple techniques, common sense and understanding the players/staff behind the teams may be a more efficient way of choosing plays.

Of course the point prediction is just that, a prediction. But I think the fact that Seattle has the statistical factors to make this a close game, rather than an 8 point Huskie win, would warrant this as a decent play for the underdog.

For example, yesterday's play I had Oregon to win by 11, so I chose Portland +12.5. I think final was Oregon won by 5. Same with DePaul+6.5. I had Arizona to win by 4, and DePaul won significantly....My main rule of thumb when playing is to trust my gut. My gut feeling says if it's going to be close (based on the model), the underdog would be the lean...

But this could all just be beginner luck. This could all be a win streak that is setting up for the worse - I have nothing to back up the legitimacy of this model and perhaps may turn into a 100% fade.

As for finding the edge :

The way I come up with the best bet is when my model predicts a margin that is vastly larger than the book spread. Once I find one, I compare predictive power margins from other stat websites, run tempo-free simulations and then simulate probability of each teams total game possessions / scoring ability per possession.

When ALLLL of those items produce a margin that is higher than current book odds, I feel confident with the edge the play has. Colorado yesterday was -3, I had them winning by 9. Before I made those plays, Santa Clara the day before passed all the tests and they won by a considerable margin.

Today I have Wichita at 6 and in at -1. Out of the shitty games on the card today I would feel they were best bet of the day.

The model does not factor in home team advantage. Since each team would have a unique and varying home team advantage assigned, I would leave that it up to everyone and our gut instincts. If you like to give every home team a 3-point advantage then allow Wichita winning by 3... no more edge.

Cal State Fullerton had a really significant advantage but I also don't trust large point spreads. Great example was VCU/ODU last week when VCU up by 20 points at the half and just coasted through the rest of the game winning by 13. Spread was -15.5. I think it's out of our predictive control to really decide if a team will continue to go full force 2nd half especially when up by 20.

But like I said, I have no backing to even say this model is profitable whatsoever. It may go 0-4 tonight and make me look like the biggest jackass in the world. It's the 2nd night of actual play so if it keeps going well I'll continue to post the plays

Your questions are appreciated! Thanks LKid and best of luck tonight

[hkarmy_kenny]

I know im writing a penetrating novel but I'd just like to add....

Predictions are predictions. When a team doesn't show up to play, they don't show up to play. All the knowledge in the world couldn't help you.

(Unless you were getting drunk with the team the previous night then you have the edge)

[Ron29301]

I know im writing a penetrating novel but I'd just like to add....

Predictions are predictions. When a team doesn't show up to play, they don't show up to play. All the knowledge in the world couldn't help you.

(Unless you were getting drunk with the team the previous night then you have the edge)

I respect the choice of words used in this post. You couldn't explain it any differently.