Quote:
Originally Posted by akkopower1
Why is there a point at which early fav bettors start to bet the dog? and how do you know when this points occurs?
Looks to me like you are saying;
9am team A -110, team B-110 (sharps start to hammer team A, team A to win is very profitable)
10am A -120 B+100 (sharps stop hammering, but public continue, team A is now break even)
11am A-130 B+110 (public have over bought Team A, you come along and buy team B before the sharps come back and sell some of A back).
I dont think it possible to buy B at 11am profitably
"Statistical analysis of this sample indicates a profit rate of 5.3 units per 100 wagers with a 95% confidence interval." Are you looking at data and simply getting the best price for team B?
Thanks Akko for your response. Yes, what you described is exactly the scenario I'm talking about. My thought process has been as follows...and please tear me apart where appropriate...I would appreciate it:-).
I've been told for years, by every bookie I've ever spoken with, that the public takes favorites. I make the assumption that oddsmakers, certainly well aware of this, will therefore price favorites higher than they might otherwise to compensate for this expected bias. I assume this because I've also always been told that the less a line moves, the better and less risky it is for the book. If this is reasonable then I assume that an opening line of -150/+130 might otherwise have opened at, say, -145/+125. I continue to assume that as the favorite becomes more expensive in the games in which it moves up, there is what I called the break even point after which there is the opportunity for a no risk wager on the part of early fav bettors. Above the line opens at -150/+130 and if the favorite is bet up I assume that there is a break even point (-170/+150 in this case)after which the early favorite players will have an opportunity to take the dog at +151,+155,+160 etc. and have a no risk wager. (in your example above, the break even point is your 11AM measure; team A-130, team B +110) Back to my example above...if the favorite wins and the dog had moved to +160, a wash..., if the dog wins and the dog had moved to +160, then the favorite loss is the original -150, and the later dog bet of +160, produces a profit.
I've selected 2800 games that have done this in the past and found that, if took the dog at or near its peak (in this case +160), there is an overall profit of 5.3 units per 100 wagers of this type. I measured the closing line and came up with a % that measures how far beyond the break even point the line moved. If, for example, the line closed at 160, then I measured the over lap, in this case 10 ; (difference between -150 fav play and+160 dog play) divided by the closing fav number of -170. So the measure I come up with, for this game, is 10/170 = .058. The database of 2800 games include games for which this measure is between -.01 and +.03. I found an expected profit of 5.3 units per hundred for games in this range. I also found that, if this measure is .04 or higher, then profit plummets and in fact, there is a negative expected mean on the Vassar stats program I use.
I make the assumption that the profit quickly disappears because more and more early bettors take advantage of the larger and larger no risk profit amount available if they make a no risk wager on the dog at +163, +165, 167 etc. if the favorite continues to be bet up..
So far the biggest challenge seem to be the fact that lines sometimes move out of range at the very last few seconds and I'm stuck with a number that was in the range, but is now less than market because the dog jumped even higher, often out of the range and over 4% in to what my data says is no longer in the profit range.
I know that this is amateur to much sharper guys and I will take the advice given to study things like efficient vs. inefficient markets, but, in your opinion, are any of these assumptions reasonable?