creating and odds line

[formula_2002]

So how do you make an odds line when the odds are a function of the final odds you are trying to calculate?

Investigation shows that the win probability of a Factor/Ranking is impacted by the final odds .

Example: On average, Bris_power (the factor), with a 1 ranking , wins 30% of the time.


But when final odds are 1-1 the win rate is 43% and at 3-1 its 23%...at 10-1 its 8%.
-----------------------------------------------------------------------------------
The above example indicates but one of over 100 factors I look at.


The CALCULATED PROBABILITY of all factors is determined from a matrix of factors and rankings. A factor /ranking relationship yields a win% based upon the results over 12,000 races. The factors and ranking are downloaded from the BRIS.com website. Factors and rankings are downloaded prior to the race.
Race results, including the final odds and win, place and show payoffs are used to update the Factors and Ranking file.
The matrix file of win % (CALCULATED PROBABILITY) is then determined from the Factors and Ranking file.

Based on this work, 39778 overlay horses produced a win pool loss of 22%, while 60485 underlay horses also returned a loss of 22%.

I would expect the overlay horses to return better result then underlay horses.

Investigation shows that the win probability of the Factor/Ranking is impacted by the final odds .

[GR1@HTR]

I'm a computer handicapper but if it is very difficult to create an accurate odds line via a single power factor. As you have found, a well bet Prime Power horse might win about 40% of the time for a 22% loss and a so called "overlaid" Prime Power horse will win 20% with a loss as well.

An effective way that I creat an odds line is by either projected figures or manual pace line selection. Since folks are more familiar w/ speed figs I will use the following example. Lets consider an 10 horse field below.

Horse#.....Projected Fig....Odds Line
1..............80
2..............82
3..............100....................5-2
4..............75
5...............98.....................7-2
6...............88
7...............83
8...............95.....................9-2
9...............92.....................7-1
10.............87


Steve Fierros book will give you an excellent breakdown of how to simply create an efective odds line...
Good luck!

[Holy Bull]

All you are saying is that final odds is
a better predictor of victory than the factor you are looking at, which will pretty much always be true.

What you can do is create a line by weightiing the final odds in with the oddsline you create. This sort of ruins
the point of making the oddsline though.

Originally posted by formula_2002
So how do you make an odds line when the odds are a function of the final odds you are trying to calculate?

Investigation shows that the win probability of a Factor/Ranking is impacted by the final odds .

Example: On average, Bris_power (the factor), with a 1 ranking , wins 30% of the time.


But when final odds are 1-1 the win rate is 43% and at 3-1 its 23%...at 10-1 its 8%.
-----------------------------------------------------------------------------------
The above example indicates but one of over 100 factors I look at.


The CALCULATED PROBABILITY of all factors is determined from a matrix of factors and rankings. A factor /ranking relationship yields a win% based upon the results over 12,000 races. The factors and ranking are downloaded from the BRIS.com website. Factors and rankings are downloaded prior to the race.
Race results, including the final odds and win, place and show payoffs are used to update the Factors and Ranking file.
The matrix file of win % (CALCULATED PROBABILITY) is then determined from the Factors and Ranking file.

Based on this work, 39778 overlay horses produced a win pool loss of 22%, while 60485 underlay horses also returned a loss of 22%.

I would expect the overlay horses to return better result then underlay horses.

Investigation shows that the win probability of the Factor/Ranking is impacted by the final odds .

[JustMissed]

I tried to come up with a mechanical method to create my odds line(which is really nothing more than my opinion as to how the order of finish). I gave up when I realized that there are too many factors to consider to come up with anything close to being correct so now I use a 'judgement' method based on the previous recent races final tote odds.

Remembering that the final tote odds are really just a vote count as to how the public voted for a horse last time out, I use that as my starting point. I only calculate my odds on my contender horses only because doing the whole field is just a waste of time.

Let's say my A horse vote was 25% of the publics votes last time out resulting in final odds of 5/1(inverse of 25% = 4 plus 1 =5/1). Let's say my A horse ran a good and even race last time out and finish finished second 1 length behind the winner who went off at 9/5. Let's say this race my horse dropped 3 lbs., dropped in class one level and has the same pilot aboard that won on him the last time he won. If my A horse is, let's say, 4 bris power points higher than my B horse, I might make my A horse a probable winner at 7/5 and any tote odds better than that results in a win bet.

I am a new player and have a lot more to learn than I know, but it is amazing how close I come to the final tote odds when using this method of simply looking at the previous race odds and mentally comparing those performances with how I think the horse will perform today.

Hope this helps and if so, please let us know.

JustMissed

[formula_2002]

Just Missed, a horse that is 3-1 has 25% of the adjusted win pool.

1/(odds+1)=%.

I don't buy into what you are saying as a way to make a profit, simple because I'm certain you have not proved your method over a long period of time...

However if it works for you...go with it.

Regards
Joe M

[Rick]

If you like to make odds lines, you'd be better off by creating a "dumb money" odds line based on obvious factors and then betting against that. Actual odds contain additional information, some of which is "smart money".

[JustMissed]

Glad I could help.

Good luck to you too.

JustMissed
;)

[NormanTD]

I believe that someone here actually uses the morning line odds as part of the calculation for their odds line. While this is not exact (and I don't know how they do it), it would seem this could be a starting point for making your odds line a little more realistic, i.e. making your longshots a little shorter and making your favorites a little longer.

Maybe the PA member doing this can post a note of three on his process. (wish I could remember who it was?!?!?!?!?)

[Rick]

NormanTD,

I'll bet against morning line odds as a pretty good indication of "dumb money" but I wouldn't use it in a true odds calculation. Actual odds would be somewhat useful though since it contains additional significant information. In fact, maybe the difference between the two could be used.

[JustMissed]

Good comment about morning line odds.

If you want to take a look at Gordon Pines article about a mechanical contender selection at icapper.com(I think that is his site), he starts out by throwing out horses 30/1 and then 20/1, etc. to get his contenders down. Of course he takes a second look at the throw outs in order to bring the lives ones back for consideration as longshots or place/show.

From my readings I have determined that most of pro's don't consider MLO except for throw out purposes or like Rick appears to use them for "betting against" purposes.

After all, the best payoffs come when the MLO favorite finishes off the board.

JustMissed
;)

[Rick]

JustMissed,

Best speed rating last race may be an even better throw out if you have a good rating method. I need to do more research on that, but it does appear promising.

[Tom]

Gordon Pine's FigLine is an excellent and cheap program to make odds lines. Fierro's templates are very good. I only look at contenders-say my "power" rating has win rate amoung the top 4 of 80% - I assign 20% to the rest of the field and then make odds on the four contenders.

[andicap]

When I make an odds line, I use Tom's method as well. What I'll do is get my 4 or 5 horses or so, and assign a probability to them. Some races I think its 90% if the other contenders are really woeful. If it's more wide open I'll do 80%.
Say its 4 and 80. That's 20% or 4-1 for each horse and I'll adjust from there. If one horse is a marginal contender he's 6-1. That leaves aboutt 5 points for another horse, etc.
Sure it's a rough estimate, but that's why I demand a big overlay when I do this. (50% at least).
Unlike Karl and I don't have the time or I'm not good enough to pick a single horse with that much certainty.


Gordon Pine is at netcapper.com

[andicap]

I feel like Columbo -- "ah, just one more thing...."

:D

I recall Ragozin a few years back did a small study on his figures that showed taking the top one of his last three showed a small profit at 4-1 odds and up. Not enough of a sample to go to the bank on, but interesting nevertheless.
Here's the kicker: When the top fig was the last race, he kicked that race and his ROI went up. (If someone from Ragozin's office is reading this and I'm dead wrong, please, please correct me -- but I remember reading this somewhere.)

[David McKenzie]

Originally posted by andicap

I recall Ragozin a few years back did a small study on his figures that showed taking the top one of his last three showed a small profit at 4-1 odds and up. Not enough of a sample to go to the bank on, but interesting nevertheless.
Here's the kicker: When the top fig was the last race, he kicked that race and his ROI went up. (If someone from Ragozin's office is reading this and I'm dead wrong, please, please correct me -- but I remember reading this somewhere.)

I recall something like that too. I think he went to Atlantic City with deep pockets and sent it in for a month. Then he wrote an article entitled Neck Not Chopped Off or something to that effect.

If memory serves, he didn't fare too well using a mechanical approach to the excotics, but was able to pocket about seventeen grand on win bets using some kind of rigid system.

I don't remember the specific ROI, but that's not the optimum way to use the sheets regardless, as I'm sure you know.

[Rick]

andicap & David,

I'm not surprised at that at all. I've been looking at my own method and noticing that the price is radically lower when the horse has the best last race speed. BUT, the trick is to have a good enough rating method that the win % is not pulled down too much by eliminating those underlays. Since I need to look at thousands of races in order to decide whether that's better than what I'm doing now, I probably won't change anything soon.