Pirateboy's NCAAF ML Dogs Experiment - W9

[MyTurn2Raise]

[ QUOTE ]
I started something similar last week. Probably missed a few games. Went 3-5, +3.9U thanks to Pitt and Stanford winning. Missed one by 2 pts and one by 3 pts. 8 games is nothing to judge by but I'm going to stay on it this week at least. I'm intrigued so keep us updated [img]/images/graemlins/tongue.gif[/img]

[/ QUOTE ]

you were likely riding the same games I was

San Diego St's fumble late in the 4th up 4 in New Mexico territory was a killer as the Lobos came back and won in the final minute

gah

[MyTurn2Raise]

oh yeah...your ranges are too large pirateboy

that article is somewhat misleading

like a 3 to 4 point road dog wins SU about 40% of the time
however, 6 to 7 is slightly less than 30%

to put 3.5 to 6 in the same category is not good

13.5 point dogs have only won SU 15.74% of the time in 108 games since 1993

[pirateboy]

MT2R, do you have an easy table for that info made or can link?

[pirateboy]

Tonight we have one play that falls into the experiment:

Air Force +235 @ New Mexico - risking 1u to win 2.35u

Break even ML would have been +166.

[CaptainHook]

You do realize that you're basically just going to be betting the point spreads which fall into the upper half of your bounds when you have sizes like that...right?

Good luck man...

[kdog]

OK first off here's a link to the article.

http://forum.sbrforum.com/college-footba...l-mls-read.html

I took a look at these this week too Pirateboy and, as others have hypothesized, all the dogs fitting the criteria are catching points at the upper ends of the groupings.

Examples (from notes,lines may have changed):
Air Force +6 ML +210
Colorado +13.5 ML +450
C Fla +2.5 ML +125

So is there really value? Maybe. What I've been thinking of doing is placing the BE point in the middle of the range and then calculating a corresponding BE ML for each 1/2 point at the high end of the range utilizing the differences in the BE points from the original chart. It's not exactly linear but it may be closer to actually determining value.
Something like this(home teams)
Range 1-2.5.........3-6.5.......7-9.5
BE +108..........+170........+262
Midpoint 1.75.......4.75........8.25

There is a $62 difference in BE ML's for the first two ranges and there are six possible spreads involved. So we'd need to add ~$10 for each full 1/2 point which would make BE for +2 $113 and for +2.5 it would be $123.

For the next two ranges the difference is $92 and there are seven spread points involved so we'd need to add ~$13 to each full 1/2 point at the upper end. This would set the BE's at $177 for 5, $190 for 5.5, $203 for 6 and $216 for 6.5.

Now admittedly I may be completely off on this but these adjusted BE's seem to match up with posted ML's pretty well.
Feedback anyone?

[pirateboy]

No value on the Fresno ML on Friday night, so here are the Saturday plays:

Connecticut +170

North Carolina +200

Pittsburgh +330

Texas A&M +140

UNLV +10

USC +130

Arizona +150

EMU +175

Georgia +255

Maryland +145

NC State +155

Ohio +240

South Carolina +125

Northern Illinois +145

Penn State +150

Cal +135

ULM +215

Arkansas State +160

North Texas +425

[ThankgodforRB]

[ QUOTE ]
OK first off here's a link to the article.

http://forum.sbrforum.com/college-footba...l-mls-read.html

I took a look at these this week too Pirateboy and, as others have hypothesized, all the dogs fitting the criteria are catching points at the upper ends of the groupings.

Examples (from notes,lines may have changed):
Air Force +6 ML +210
Colorado +13.5 ML +450
C Fla +2.5 ML +125

So is there really value? Maybe. What I've been thinking of doing is placing the BE point in the middle of the range and then calculating a corresponding BE ML for each 1/2 point at the high end of the range utilizing the differences in the BE points from the original chart. It's not exactly linear but it may be closer to actually determining value.
Something like this(home teams)
Range 1-2.5.........3-6.5.......7-9.5
BE +108..........+170........+262
Midpoint 1.75.......4.75........8.25

There is a $62 difference in BE ML's for the first two ranges and there are six possible spreads involved. So we'd need to add ~$10 for each full 1/2 point which would make BE for +2 $113 and for +2.5 it would be $123.

For the next two ranges the difference is $92 and there are seven spread points involved so we'd need to add ~$13 to each full 1/2 point at the upper end. This would set the BE's at $177 for 5, $190 for 5.5, $203 for 6 and $216 for 6.5.

Now admittedly I may be completely off on this but these adjusted BE's seem to match up with posted ML's pretty well.
Feedback anyone?

[/ QUOTE ]

I like it kdog. Not perfect, but seems like a good approximation. I definitely like it much. much better than just using a single number across a wide range (no offense pirateboy).

Will be interested to see your results!

[Austiger]

[ QUOTE ]
No value on the Fresno ML on Friday night, so here are the Saturday plays:

Connecticut +170

North Carolina +200

Pittsburgh +330

Texas A&M +140

UNLV +10

USC +130

Arizona +150

EMU +175

Georgia +255

Maryland +145

NC State +155

Ohio +240

South Carolina +125

Northern Illinois +145

Penn State +150

Cal +135

ULM +215

Arkansas State +160

North Texas +425

[/ QUOTE ]

Maybe I don't understand, but it seems you are going against that article now. You have a lot of bets there on teams that are 3-6.5 point dogs, but you aren't getting the +166 necessary to qualify for your experiment. NC State, Penn State, Cal, Maryland...

[silentbob]

[ QUOTE ]
Maybe I don't understand, but it seems you are going against that article now. You have a lot of bets there on teams that are 3-6.5 point dogs, but you aren't getting the +166 necessary to qualify for your experiment. NC State, Penn State, Cal, Maryland...

[/ QUOTE ]
I think that's the point of his "fine-tuning" the spirit of the article, as explained above. It really doesn't make sense to treat +3 and +6.5 dogs interchangeably and use the same cut-off, so interpolation seems sensible and consistent.

In other words, I'm sure someone could go back and perform the same analysis with a smaller spread range than 3.5 points, and my guess is that the "cut-off" for +3 would be different from the one for +6.5.