Finding Value in theTrifecta
 

How do I go about finding any sort of value in trifectas with the odds on chance winning most of the time..

I have considered doing my ratings and prices and taking only those that are my price or better, but you have to stick to races of 10 to 12 runners, aything less would be throwing your money away.


Mad Gambler

all about dog trifectas
 

When people wager on Greyhound Trifectas they REALLY get in a rut. There are FAR too many people that use standard wagering forms:

1/23/ALL
1/2345/2345
Box 3 or Box 4
123/1234/12345

Almost all of these simple wager structures are inefficient and lead to trouble. Most have too much play on the best dog to win, and a lot have too much play on the best dog in third, just as a generality. And almost all suffer from the "eliminate some dogs from my bet" mentality.

As a first defence, after you are done handicapping you should look and see if your top selection is the same as the crowd's top selection. If you all agree and you all win, you're going to get small payoffs. That's starkly simple, but people can't seem to get it. If you agree with the crowd favorite and you WIN you don't get much. If you agree with the crowd favorite and LOSE, you don't get anything. Two possible outcomes, neither of which is very satisfying. Much better to disagree with the crowd favorite, and STILL win! That requires a wagering strategy, not a handicapping change.

As a second defence, after you are done handicapping you should look at the Tote board and figure out what you're going to get paid if you win. If your top-ranked dog is 3-1, you are much more likely to have a bigger payoff upon winning, as opposed to times when your dog is 4/6, when you are likely to get a tiny payoff. Maybe think about making a wider bet in 3-1 situations, and making sure you don't include a stronger crowd favorite. On the 4/6 side, if you insist on playing that dog, then narrow up the wager somehow, and hit it 15 times, instead of playing a wide wager with each combination represented only once. If you FORCE me to take a $30 trifecta payoff, I want it 15 times, not once.

The whole issue of playing (or NOT playing) the best dog in a race is important, and people can't face it. The best dog in a race might win 25% of the time. That means he LOSES 75% of the time. Why bet on him to win when he's going to lose three times out of four? Your own handicapping tells you he's going to win only 1 out of 4, but you're going to bet on the 1 out of 4 case instead of the 3 out of 4 case! And on top of it, the 1 out of 4 case pays squat, while the 3 out of 4 case pays anywhere from "better" to "wow"!

If you bet on your top-ranked runner, and your handicapping agrees with the crowd (a very likely situation) then you are going to be continually cashing those little $30 and $60 and $90 trifecta payoffs. If you learn that the favorite doesn't win all that often, and decide to figure out how to make trifecta plays that don't involve that dog, then BY DEFINITION, you will start cashing the bigger trifectas. Remember, SOMEBODY cashed the $250, $450 and $650 trifectas. It just wasn't a guy with his money on the favorite to win in his little trifecta pyramid.

For starters, maybe people could think about it like this:

I can EASILY identify the crowd favorite. It doesn't take much handicapping, and may even take NO handicapping if I'm smart enough to look up at the Tote board once in a while. Okay, now I know which dog to AVOID using in my bets. Now, let me change my handicapping strategy to figure out which OTHER dogs might figure to be in this race, particularly if the crowd favorite ISN'T!

Another important idea to consider... I'm perfectly willing to NOT win the $30 trifecta payouts. When one of those $30 things happens, and I didn't cash it, I can smile and be happy. I didn't waste a bet on a no-ROI wager. I'm perfectly willing to watch SIX of those $30 payoffs go by, if on the next race I cash a $350 trifecta when the favorite is out of the money. Patience. No panic. Adequate bankroll. Maybe people should think, "My strategy is to cash one $350 trifecta tonight" instead of "My strategy is to cash three $30 trifectas tonight." It takes a completely different strategy at the track to accomplish the former. It's no sin to leave the best dog out of your bets, and it may be a profitable virtue!
I ran out a study just now for all greyhounds that went off at 2-1 odds or shorter in QLD for a major part of this year. This happened in 675 races.

- The top-ranked dog (at 2-1 odds or shorter) won 43.85%.
- The top-ranked dog finished in second place 17.63%.
- The top-ranked dog finished in THIRD place only 9.78% of the time.

Compare that third-place performance to the rest of the dogs:
The second-best dog finished third 15.26% versus the fav's 9.78%.
The third-best dog finished third 15.11% versus the fav's 9.78%.
The fourth-best dog finished third 14.96% versus the fav's 9.78%.
The fifth-best dog finished third 13.64% versus the fav's 9.78%.
The sixth-best dog finished third 11.54% versus the fav's 9.78%.
The seventh-best dog finished third 10.50% versus the fav's 9.78%.
The worst dog in the field finished third 9.00% versus the fav's 9.78%.

So in those races where the fav is 2-1 or less, ALL other dogs will finish in third place more often than that favorite, except for the very worst dog, and THAT dog is barely below the fav's third-place percentage.

Now from that point of view, tell me why it makes any sense at all to put a strong favorite in third place in a trifecta wager?

Look at the ROI in the following structures, as measured across all possible races:

1/2345/2345 = -24.73% ROI.
2345/1/2345 = -29.30% ROI.
2345/2345/1 = -33.19% ROI.

12/2345/4567 = -13.62% ROI.

That 11 percentage points better than the BEST of those three strategies, and almost 20 percentage points better than the WORST.

Hope this helps in your trifecta structuring....start playing the trifectas an you start playing against me....

Thanks for the info. It has opened my eyes.

Mad Gambler

Elimination Rules
 

Thanks MG here's a little more to ponder..
A lot thought goes into handicapping, and not enough about wagering. If you stretch your thought processes a little, working on your wagering is very much like working on your handicapping. I've tried to explain a little in the past re my methods, So I thought I would show a simplified idea that pretty much anybody could use, as long as they have some ability to study ROI on various wagers. That might come from database research.

The basic idea is this. First, you study wagering a little bit, and find some basic ideas that ARE NOT PROFITABLE. That means, they have a VERY negative ROI. The WORSE the strategies, the better. Make a list of these "rules" for betting methods or techniques that NOBODY SHOULD USE.

Then, you handicap the races, I use a Monte Carlo simulator. You can get servicable software like this for free on the Internet, so there's no reason why this can't be easily done. You use this software to produce a ranked list of trifecta wagers, with the most likely outcomes at the top, and the least likely at the bottom.

Then you simply go through the list of ranked wagers, eliminating all of the BAD BETS that you identified in step one. This is a simple concept, and only requires a modest investment of effort (time) and some thinking. But it's a MUCH more effective way of structuring your wagers than relying on the old 3-dog box, or the 23/234/2345, or the 1/2345.

Let's look at a simulated example to get the feeling for how this works. Let's say I'm handicapping a nice little race, and I can assume the crowd favorite will be the #1 dog, and the crowd's basic ranking order is this: 14537682. You can pretty much get this off the Tote board, or spend a few minutes "thinking like the crowd" and do it yourself. You don't need specific odds, just the crowd rankings. For most purposes, it doesn't even have to be particularly accurate past identifying the crowd favorite.

Now from those wagering studies you did, here are the bets to avoid. The left side is the description of the wager, and the bets on the right are the actual box numbers the rules generate for the current race we are working.

Wager Elimination Rules

From "Box" players inefficiency... these are the parts of three-dog and four-dog boxes that really are bad investments. Rankings based on crowd odds.

1. 234/1/234 = -31% ROI (345/1/345)
2. 234/234/1 = -43% ROI (345/345/1)

From "Standouts" players inefficiency... the key players overbet the favorite with the best three or four dogs. Rankings based on crowd odds.

3. 1/2345/2345 = -36% ROI (1/3457/3457)

From "Numbers" players inefficiency... people overbet certain boxes, no matter what kind of dog is in them. These are the problem boxes we find from a simple post position study.

4. Box the 123 boxes = -36% ROI (123/123/123)
5. Box the 128 boxes = -31% ROI (128/128/128)
6. Box the 178 boxes = -34% ROI (178/178/178)
7. Box the 678 boxes = -36% ROI (678/678/678)

From "Hot Dogs etc formguides" players. Buy your formguides, and avoid their plays. Or avoid the any tipsters selections

8. Hot Dog = 1-4-7 (147/147/147)
9. The Recorder = 4-3-6 (346/346/346)

From "Factors" players. If you can study factors, you'll find that the crowd overplays some of them. At this track, they SEVERELY overplay the dogs with the best 6-line finish average in the guide. (In my example, the #1 has the best 6-line finish average, followed by the #2 and the #3. Easily calculated from the program page.)

10. 6-line Finish Avg 1/23/23 (1/37/37)
11. 6-line Finish Avg 1/2/ALL (1/3/245678)

Well, those 11 rules are a pretty good start. Here is an example of the simulation results for this race. Here are the simple win percentages as predicted by the simulator, just for your reference:

Sim Results:Box Win Percentages
4 29.2%
1 26.8%
5 10.8%
3 10.5%
7 6.2%
6 5.7%
8 5.6%
2 5.3%

And here are the top 50 Ranked Trifecta Wagers. Note that to the right side of the wagers, I have noted the rule from above that triggered to ELIMINATE this straight trifecta from use for this race. Thus you would want to play the "naked" wagers, the ones left after your eliminations. Here are the first 50 wagers from the simulator, their probabilities, and the trigger rules...

1. 1-4-5 2.08% Rule 3
2. 1-4-3 1.99% Rule 3
3. 4-1-3 1.85% Rule 1
4. 4-1-5 1.85% Rule 1
5. 1-3-4 1.44% Rule 3
6. 4-5-1 1.41% Rule 2
7. 1-5-4 1.41% Rule 3
8. 4-3-1 1.35% Rule 2
9. 1-4-7 1.33% Rule 3
10. 1-4-6 1.30%
11. 4-1-7 1.22% Rule 8
12. 1-4-8 1.17%
13. 4-1-8 1.15%
14. 4-1-2 1.12%
15. 1-4-2 1.12%
16. 4-1-6 1.09%
17. 4-3-5 1.01%
18. 5-4-1 0.94% Rule 2
19. 4-5-3 0.92%
20. 5-1-4 0.86% Rule 1

Hope this gives you some more areas of study

Woof,

This is quality stuff, particularly when contrasted with certain TAB-sponsored material.

Essentially this is consistent with a Dr Z assertion - something like that favoured permutations are overbet in exotics.

But you go further in identifying other mug-plays to avoid like the infernal dog 128 and tipsters' top 3.

However newcomers need to be warned that Trifectas are an extremely tough game.

With very ruthless competitors - particularly those who use the official trend matrix available to only some and not on the Internet for the general betting public.

I think most would be better off playing the new win & place opportunities with sub-5% rakes.

Woof. Good morning. I've read all your stuff in this thread and will consider it, slowly. However, I have a question concerning "handicapping" (be that the dogs or horses or whatever****, which I'm hoping you might clarify for me (or anyone else, of course****. It is this: since all the "facts" are in the guides etc how can the individual handicapper/ratings fan/form student see what the others DON'T??? And regularly enough to consistently win?? What I men is; if we all give importance to times/boxes/finishing positions etc etc then how will we sort ourselves out from the mass of punters/bookies doing likewise?? Please don't take this as criticism. I am REALLY interested in everyone's opinion on this. Thanks.

I think woof's point is that we are not trying to handicap any better than anyone else, even look at the tote board to get our "handicapping" the easy way. His way to profit is statistically sound betting.

KV

Hi KV. I'm not so sure about this. As you know I'm not a stats/ratings fan but Woof does say in both posts that we have to "do our handicapping" and then make comparisons. He also mentions the point that a lot forget; if we ALL agree on the fav etc etc then we are all going to get small reward when THAT fav wins (ie we MUST disagree before skinning them****. Problem is, that when we DISAGREE, we are jumping immediately to the conclusion that it is we who are "closer" to the truth than the rest. The implication being that our assessment of say 3-1 is "better" than the crowd's 5-1 which gives us the edge in taking the crowd's "foolish" $6 offer and also the edge IN REVERSE in Trifectas (ie the crowd are throwing up "unders" on the bad ROI Trifectas, so WE scoop up the "overs" trifectas****. Or am I misreading this?
Plus I'm a little confused about the whole section in the second post. Take "wager 1" to avoid; 234/1/234 is the Public's 2nd 3rd 4th fav to win with Top Fav to come second and 2nd/3rd/4th fav in third place, but what is the (345/1/345**** after it and where does that come from?? Thanks for your patience.

Hi Punter those numbers are from my own sample sims for the race, if you look at the earlier paragraph you will notice how the crowd ranked the race.

question
 

Most ppl don't use all the "facts" within the formguides, my own handicapping is based on LTD Form, it take a major shift in thinking from using so called "Recent" form to using LTD formlines. most "highs an lows" within a runners formline is usually a function of Class, Distance and Box/Barrier draws , in another thread on this forum somewhere, I have detailed my analysis on Recent form as opposed to LTD form.

I only use 4 or 5 Factors/Variables "that the public can read in the formguides" to get my first tier Sims, once I have these probablilties "these are the ones that show the least variation or smoothest variation thru all the rankings when compared to the Public" I then re-handicap using Factors not readily disclosed in the formguides, then compare both sets to find races that might be playable (which is around 600 races approx per week, within aust).

I Classify races based on the Main Factor that the Public identified in each Favourite, then its a matter of doing some cluster analysis on each grouping to find races that have horizontal profit profiles.

I have also mentioned on numerous occasions using Standard Deviation as a TOOL, but most would ignore this valuable tool and so on..