Sports Betting in General

A chance for all my detractors to blast me.


There are three ways to view a line.


1. It’s accurate.


2. It’s inaccurate because one side is being wrongly over bet.


3. It’s inaccurate because many handicappers are not accounting for intangible factors.


Notice that 2 and 3 could be combined in some or maybe even many cases. Therefore in order to evaluate whether or not a line is accurate you have to have estimated what you think an accurate line is. If you think that the line is seldom accurate then you don’t need to do this. I think that believing the line is seldom accurate is a quick way to the poor house.


How far is the line off is the next consideration. If a football team is laying 19 points and you think that the line should be 21 points you probably don’t have a bet. If a football team is getting 1 point and you think the line should have the team laying 1 point you probably have a bet. Do others agree with this? In both cases the line is off by 2 points but the circumstances are vastly different. Thus I recommend betting on small favorites and small dogs MOST of the time. With the passage of time I try and find games where a big underdog is a good bet. Also in playoff situations I believe that laying some points that would make you feel uncomfortable during the regular season can be profitable often enough.


I recommend flat betting. Especially when your system for determining a line is evolving. I agree with the comments I made by Mr. Malmuth regarding this. Mr. Malmuth wrote:


“I've always been of the opinion that there are a few football bettors who bet too many games. What they seem to do is to use some criterion that identifies candidate games of which a subset of these games are the good bets, but they are not sure which ones are the good bets. Thus they bet them all and have a small edge overall (yet they put a lot of money in action). “


His statement I’m totally in agreement with. I also think that if you started with this approach and kept careful records and reviewed them you would find over what games to pass on thus reducing the number of games bet, increasing the amount of money bet on each game, and increasing the winning percentage.


Most of the post’s I’ve read on this forum seem to try and identify games where the “intangible” factors are not being accounted for. That’s fine but be advised that when you do this you must estimate the probability that you are right. Here are a couple of examples. There was a lot of discussion about the Cleveland vs. Seattle game. One of the intangible factors to consider is the impact of Cleveland’s new coach Butch Davis. One could make a case that he will be an excellent NFL coach but how certain of this could you be especially since he has no experience as an NFL head coach? Another intangible factor is Seattle’s new qb Hasslebeck. A case could be made that he will struggle in his first game as a starting NFL qb but how certain of this can you be? So when you dust off the crystal ball and opine that the loss of a player will greatly effect team chemistry for instance you better make some sort of estimate of the probability that you are right. Keeping records of these types of evaluations and their results is also valuable.


Next Subject


Determining the Line in Pro Football

This was a very interesting post. I hope we more like these.

You have a very interesting post.


Logical. Well written. (Unfortunately) very, very wrong.


1. Any pro would much rather lay -19 on a game that should be -21, than play a team +1 on a game they shoud be -1. Why? I leave you to do the trivial math, but you can use:


Prob pkem type game


a: Home team wins by 1: 2% or so

b: Ties (zero)

c: Road team wins by 1: 2% or


So grabbing +1 on a game that should be -1 is obviously no big deal. You'll win 52%.


Contrast this to a -21 lined game


19: maybe worth 1.5%

20: 3% (ballpark)

21: 3.5% (ballpark)


Obviously the -19 you are dismissing is a BONANZA that will cover 55% of the time, much more than the crummy 52% you grab from capturing 1/2 of two crummy numbers, and from grabbing the entire zero (that is worthless).


2. Obviously thread 1 above 'clobbers' a flat bet theory. You love a favorite is 21 already and are about to bet. It's 21 everywhere. Then the Showboat floats you a 19 for some inexplexicable reason. Do you give them more 1 unit. Oh yeah. Right up the_________!


.......Show me a guy that flat bets, and I'll show you a guy that doesn't understand the mathematics of sports betting and the value of good numbers.


Example: I will play the Jets in a tiny wager tomorrow at 2.5. AS soon as someone gives me the +3, -110, they are going to get the Jets chunk bet.

I am heartened when I read stuff that debunks flat betting.

Compare with Black jack betting, where you should increase your bet size as your advantage increases. The Kelly Criterion shows the optimal way to ballance the risk of going bankrupt from making large bets against betting with a positive expected value.


Of course unlike 21, where you can do exact calculations to know when you have a positive EX, in sports betting you can't know for sure, but there should be times when you find a line that is out of whack, and that's when you should bet more, as Fezzik describes.


On a related note, I managed to get a SB future bet on the Titans at 9-1. This is about double what I have seen in some places, so I was feeling pretty good about it. My opinion has changed after last night's game.

One should estimate an advantage and use a Kelly or fraction-of-Kelly betting scheme. Flat betting would be appropriate if you cannot estimate your edge. But if you cannot even roughly do this, you have no bet. If you are estimating how off a line is, you can easily estimate an edge using historical distributions given the spread. This is not perfect, so you shoudl assume a smaller edge and/or bet a fraction of what the Kelly formula suggests.


As Mason points out in his gambling theory book, flat bettors are losers - in stocks, in sports, in poker, and in life.


Good luck.


Dan Z.

You'd be crazy to bet on every game.


What I wrote was:


____________________________________

I recommend flat betting. Especially when your system for determining a line is evolving."

__________________________________________________ _______


Personally I think this evolution process is very long. If I had a system that was really, really accurate I'd never make it public and it certainly wouldn't be on an internet forum site. Therefore I don't feel that very many are all that accurate in predicting how far the line is off but over time they may get better at it. Further I wrote:


__________________________________________________ ___________

I agree with the comments I made by Mr. Malmuth regarding this. Mr. Malmuth wrote:


“I've always been of the opinion that there are a few football bettors who bet too many games. What they seem to do is to use some criterion that identifies candidate games of which a subset of these games are the good bets, but they are not sure which ones are the good bets. Thus they bet them all and have a small edge overall (yet they put a lot of money in action). “


His statement I’m totally in agreement with. I also think that if you started with this approach and kept careful records and reviewed them you would find over what games to pass on thus reducing the number of games bet, increasing the amount of money bet on each game, and increasing the winning percentage.


__________________________________________________ ____________


Again if you have a system that is extremly accurate I agree that you should size your bets according to your advantage.


You wrote:


Flat betting would be appropriate if you cannot estimate your edge. But if you cannot even roughly do this, you have no bet.


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


I'm sure you'll disagree but I don't think very many have a system that can estimate their edge accurately but I do believe that there is a significant number of bettors that can identify when they do have an edge. Fezzik implies that he can be that accurate. He used the Jets vs. Cots game as an example where .5 points would be the difference between a chunk bet and a small bet. Now either the game was a total fluke and I seriously doubt this or his system wasn't as accurate as he thought in estimating his edge. I know I'm taking the results of 1 game to illustrate a point but I think the point is valid.

Remember the premise of Kelly also is that you will go broke if you overbet your bankroll. When dealing with football and its relative lack of sample size, you are really playing with fire don't you think? I don't know what my exact edge is because there are too many variables in each game. I could make myself sick with all the ways to gauge games (betting home favorites, betting teams on their strong rush D, betting against a team in a flat spot on the schedule, etc). All I know is in the end I have a certain win %. Even that could be dubious because things are always changing in sports betting and the public focuses on different things and its easy to not catch onto that until after its happened somewhat and you can discern it with a step back from the situation. Besides part of the problem is that much of what you do in football is really odds based. Most people won't look at it this way, but thinking you can specify exactly how many points a team is better than another is a fool's game. Think about it, if you say team A is -5.79 points better but the spread is -3.5, how much edge do you really have? How many games end on 5? Don't you still pay attention to key numbers since points are scored in certain sets of amounts? Other sports lend themselves better to more precise betting percentages. They have larger sample sizes for you to get a good idea of where you stand edge wise and are less dependent on key numbers since scoring is not in such specific amounts.

For most suspect post of the month and possibly the year.


You ask how important a crummy half point could be in sizing your advantage for a bet +2.5 vs. +3.0? As "proof" you cite 1 losing game where it did not matter. The weekend after my post of the great importance to UNLOAD when you capture a good number like a 3.5 on a 3 game you like, 4 (four!!!) games lined at around 3 (2.5 to 3.5) landed squarely on the 3.


Sea vs. Cle. Sea 9-6.

Oak vs. KC Oak 27-24

SF vs. Atl SanFran 16-13

Rams vs Phi Rams 20-17.


This was one of the greatest aberration NFL weeks of all time with over half of the games lined 2.5-3.5 landing on the number. And of the 7 games, you took 1 of the 3 that didn't middle and cited it as evidence that it was no big deal?

I don't know what else we can do but run a semi into your bedroom with a "3" written on it to get you to wake up on this one.


Ps. Tbay -3 is my best bet of the weekend. Obviously, if a shop floats a -2.5, I'm not asking you to use differential equations to determine optimal fractional Kelly bet sizing. All I'm looking for is an obvious conclusion: BET MORE ........the same bet size, IMO, ranks right up there in bad intuition with having a "1st card is an Ace coupon" and flat betting your normal unit size.