[Peep]

While I may be a "technical gambler", my knowledge of statistics and math is very limited IMO.

While all a bookie needs to know maybe that 11 is bigger than ten, all I want to know is that I am going to get +120 or better on a game/bet I think is even money. Not a rocket scientist approach, but it keeps me going.

Here is the case in point. Game is Carolina/Philly preseason NFL. Philly -3.5. I am going to bet Carolina in the first half.

1) I have a database of preseason games. I have 302 games where the home favorite is -3, -3.5 or -4.
2) In these 302 games, the HF wins 152, loses 114 and there are 36 ties.
3) With -2.5 at the half, the favorite wins 143 and loses 159, so I like my +120 with +2.5.
4) But 18 games fall on the "3", which is to say that I could give up my plus +120 for +100 and "save" 18 bets out of 302.
5) The current Half time line is +2.5/+120 or +3/+100, or +164 for the dog to win outright.

I am going for the +2.5/+120 since it is easy to see my edge, which is +120 on essentially an even bet (based on my data).

Thoughts on how to think about this? I just seem to get confused....

[tomcowley]

If you assume your numbers are exact and predictive of this game..

If you bet +2.5 +120, you win 52.65% of the time (159/302)

If you bet +3, you win 55.98% (159/284 (302-18 pushes))

I'm sure you have the data on the dog winning outright, but you don't have it in the post, so we can't evaluate that bet.

For every $100 you bet on +2.5 +120, you win $15.83. For every $100 you bet on +3 +100, you win $11.96. Clearly your better bet is the +2 +120. Kelly might suggest staking a small amount on +3, but I'm too lazy to run that right now, and you won't go far wrong betting your intended bet all on the +2.5 +120.

Now.. I'd recommend looking at the particulars of your data (are you weighted heavily towards -3? How would it look if you put in -4.5 and -5 teams?), and the particulars of this game (is philly expected to have an above-average (for the spread) 1H edge because of certain lineup/playtime issues?), to see if you want to be blind-betting it.

[Peep]

Thanks Tom.

It is for sure a "blind bet", I prefer doing it that way, everytime I start thinking, I seem to get into more trouble.

Data on dog winning the half outright is 114/266 (36 games of the 302 were tied at the half).

[Justin7]

Despite what the data has showed, I wouldn't count on winning that bet more than half the time.

[reno cool]

If all your going off is a database of 300 games. And your simply looking at how underdogs do in general against a particular spread it is a blind bet. close to -5%

[LT Profits]

Also a blind bet ignores the fact that the Philly starters are expected to play the entire first half while the Carolina starters will see either limited or no action.

[Peep]

I wasn't saying I made a good bet.

I am more interested in opinions on

1) Given that I have chosen to bet it and given the data I chose to bet it on
2) What in your opinion is the best way to bet it and why?

[LT Profits]

Peep,

Then the obvious choice is +2.5 +120 per Tom's post.

[Data]

Inspite what the data has showed, I wouldn't count on winning that bet more than half the time.

"Typo" corrected.

[Rufus]

Despite what the data has showed, I wouldn't count on winning that bet more than half the time.

I would, but for a different reason, which shows up in my database of 784 preseason games.

(and yes I know Carolina lost)

[Peep]

would, but for a different reason, which shows up in my database of 784 preseason games.

Aw, I only have 770 preseason games, no wonder I can't make any money!

This is the first year I have bet preseason, looks like a different animal for sure. I have broken the data base up so I can sort by week # as well, that looks interesting, seems scoring goes up as the preseason goes along.

BTW, I showed Carolinia winning 10 out of 18 second halves outright under these conditions (3-4 point road dog, up by 10), so there went the big buck 1st half profits and.....

(yes I know Carolina lost).

[Data]

1) I have a database of preseason games. I have 302 games where the home favorite is -3, -3.5 or -4.
2) In these 302 games, the HF wins 152, loses 114 and there are 36 ties.
3) With -2.5 at the half, the favorite wins 143 and loses 159, so I like my +120 with +2.5.

How did you find this split?

Is there a reason behind the numbers that explains why these underdogs should be covering? Did you have an idea about that underlying reason that the numbers confirm? That would be a right approach but I doubt this is true because this looks to me no different than a myriad of other meaningless "9-1 last ten Wednesday games" trends.

This type of "angles" is usually found by looking at many possible splits and noticing which ones are "statistically significant". This is called datamining and it is one of biggest mistakes that many gamblers make.

Aw, I only have 770 preseason games, no wonder I can't make any money!

This is not going to change once you get a bigger database unless you change your methods. Datamining a bigger database will only produce more false angles.

[Peep]

Agree it is datamining, which is at best a tricky method.

But life might be datamining. We all act based on what has happened to us before. I am acting on what numbers have done before. Am I "shaping" them? Maybe.....

Essentially what I am assuming is that the "main" number, the game line is right. I then look to see what the other numbers "split" at. In this case, the halftime line for favorites -3/-3.5/-4 "splits" at a different point than the book's number.

The book does not always split at the right mathematical number, that split does not necessarily maximize their profits. I got cut way back at Gameday for winning 19K on a split that the book KNEW they were doing wrong, BUT, because people would still bet the "wrong" side, increased their overall profit.

[LT Profits]

I have broken the data base up so I can sort by week # as well, that looks interesting, seems scoring goes up as the preseason goes along.

Yes, I made a post on that last month.

Since You Guys Got Me Thinking About Preseason

[BuddyBear]

Don't expect a trend like that to continue. Any descriptive data you have I promise you that linesmakers have it as well. Trend is likely to reverse itself....

[LT Profits]

Don't expect a trend like that to continue. Any descriptive data you have I promise you that linesmakers have it as well. Trend is likely to reverse itself....

Well, the actual scoring will increase and be close to the historical numbers for each week, What I think you are saying is that the posted totals will be higher too, but that doesn't mean the numbers can't be beat with selective shopping.

[Peep]

[

Any descriptive data you have I promise you that linesmakers have it as well.

Probably a different topic all together.

Other than Pinnacle, Oly and CRIS, I have yet to encounter a Caribbean book that I think can set a line.

The above three certainly can, and can do a better job of it than I can. But do they want to post the "right" number (mathematically)? Maybe more money in it for them to post a wrong number.

Was a great thread at Major Wager a few years back, entitled "why fix a game when you can fix a number"?

[durito]

My preseason database has 1544 games