What to do?

[cj]

I always find myself caught in a dilemma when I face the situation I encountered below. What do some of my fellow handicappers do:

Aqu 8th yesterday, I had the following line before the race:

6-103 Bet Odds 5-2 (went off at 7-2)
1-101 Bet Odds 4-1 (went off at 5-1)
9- 98 Bet Odds 9-1 (went off at 4-5)

I bet the 6 (and lost) while watching the 1 horse gallop at a nice price. This seems to happen to me more often than not when I have two or more overlays in the same race.

HELP!!!

CJ

[Jake]

Easiest is to bet both horses. Determine what you think their combined win probability is, and your advantage relative to their combined odds. Adjust your bet size accordingly.



Originally posted by cjmilkowski
I always find myself caught in a dilemma when I face the situation I encountered below. What do some of my fellow handicappers do:

Aqu 8th yesterday, I had the following line before the race:

6-103 Bet Odds 5-2 (went off at 7-2)
1-101 Bet Odds 4-1 (went off at 5-1)
9- 98 Bet Odds 9-1 (went off at 4-5)

I bet the 6 (and lost) while watching the 1 horse gallop at a nice price. This seems to happen to me more often than not when I have two or more overlays in the same race.

HELP!!!

CJ

[gumbo]

Agree with Jake. Particularly when neither horse goes off at 2-1 or less.

It becomes more difficult to bet both when one is 2-1 (or less). Now your betting $4 to potentially win $6 (assuming 2- $2 bets and the low odds horse wins) and your win percentage needs to be up around 60% (not 67% because somethime the higher odds horse wins).

I don't buy the argument that betting 2 horses is merely a way to minimize daily losses or protect your bankroll against long losing streaks (though it does have that benefit). The fact of the matter is, if you can find 2 horses in a given race that you have as your first and second choices and they both go off at 3-1, your break even pick percentage is 50%.

How many of you think you can pick 50% winners if you get 2 horses?

gumbo

[Tee]

Bet the horse you think is going to get the better trip!!

[GameTheory]

Originally posted by cjmilkowski

Aqu 8th yesterday, I had the following line before the race:

6-103 Bet Odds 5-2 (went off at 7-2)
1-101 Bet Odds 4-1 (went off at 5-1)
9- 98 Bet Odds 9-1 (went off at 4-5)



In my opinion you made a mistake in your choice of bets -- if you are only going to bet *one* horse, then I would always choose the one at longer ones, so I would have bet #1 instead of #6.

But in this particular case, you should have bet both horses. Adding an extra win bet (of the same amount) is the same as knocking one odds point off of the bets you do make -- so you would have been getting your exact "Bet Odds": 5/2 on #6, and 4-1 on #1.

I'm assuming "Bet Odds" means "odds at which you will bet", but if you require the odds to be *above* that level, then betting two horses would have knocked the value out of both. In a case like that, I would bet the higher priced horse only.

[BillW]

CJ,

I hope you weren't looking for a concensus here.

In the spirit of confusion, I'll add my $0.02.

I think you made the right play. I would play the higher odds if I evaluated 2 horses to be equal (you did not). Otherwise you played the larger edge 7:2 / 5:2 > 5 / 4 (slightly). That may be nitpicking but that's my ( academic!) answer and I'm sticking to it. ;)

Bill W.

[GameTheory]

I agree with BillW in theory but not in practice, because I don't think anyone's estimations of "edge" are that accurate.

Go for the bigger price in a single bet, but bet both if possible.

[cj]

Excellent, well thought advice from all. That's why I've come to really like this board! I used to go with my highest rated horse who was at or above my bet odds, but I think I'll try the two horse bet when it still leaves both with acceptable odds. I'll let you know how it goes.

Thanks all,
CJ

[anotherdave]

Your edge on horse 6 was about 28% and your edge on horse one was 20%. Now those are just from your odds line.

I have often been just as frustrated when facing such situations. 15 years ago, when I was in graduate school, I made one of my projects to figure out an optimal dutch formulas for multiple overlays (as well as the win/place situation). The results were that mathematically you should always bet both - now that may be disputable in real life! I still use these formula as a guide when faced such a dilemma.

In the situation stated the formula say :

Bet 62% of my wager on the 6 and 38% on the 1.

Take it for what it is worth! Hey, it is the only paper that I have published in my academic career, so I got someone to believe it!

[GameTheory]

I would love to see an elaboration on those formulas, if they're not too complicated (maybe a new thread is needed)...

[anotherdave]

Originally posted by GameTheory
I would love to see an elaboration on those formulas, if they're not too complicated (maybe a new thread is needed)...

I'll try to explain it for the 2 horse case. Still might be a bit confusing but I'll give it a shot. (They get more complicated for 3 or more horses - need matrices.) I've tried to simplify the concepts.

The basic setup is:

Horse 1 assume we know the edge, and actual odds

I will call the edge E1 and the odds Q1

Similarly for horse 2 the edge will be E2 and the odds Q2.


Now we figure out the following:

For horse 1 :

E1*Q2 + E2

or horse 2:

E2*Q1 + E1

In CJ's example his first choice (#6) had a 28.6% edge and odds of 7-2. (E1=.286, Q1=3.5)

His second choice had a 20% edge and odds of 5-1.(E2=.2, Q2=5)

so for the 1st choice we get .286*5 +.2= 1.63

for the 2nd choice .2*3.5 +.286= 0.99

Now to determine the proportion on each we divide each number by the sum of the two (1.63+.99)=2.62

so 1st choice (#6) = 1.63/2.62 = 0.62

and 2nd choice (#1) = 0.99/2.62 = 0.38

so 62% on 6 and 38% on 1. My usual bet is $100 so I would bet $62 on #6 and $38 on #1.


There are a few extra things you can do too (like figure optimal % of bankroll on each of the two bets like the Kelly Criterion, but I won't get into that. I don't use Kelly as such)

Sorry if I gave a lousy explanation - it is easier to calculate when you set it up on Excel!

Also as was mentioned by Game Theory, if CJs estimates of the fair odds are off (even if both horses are still overlays), then these percentages will likely become wrong (Garbage in, garbage out). So much for math!

[GameTheory]

Interesting...

I will file those formulas away for future reference.

Your use of the word "edge" is different than what I am used to, however. You are using it as simply my (or cj's) estimation of the winning chances of the horse (5/2 = 28.6%, 4-1 = 20%, etc. assuming those are the "fair odds" levels), wheras I would normally think of it as these numbers compared to the actual odds (7/2 is greater than 5/2, therefore a positive edge).

One of us bright people ought to make a horseracing faq with all the technical terms we're always throwing around strictly defined, along with a list of useful formulas for horse bettors like the one above.

You don't have to explain it here, but is there some way you could relate to me the more complicated formulas for 3 or more horses (the ones with matrices)? I'm making my own "personal toteboard" that grabs the odds/pool info off the internet and saves it to analyze/detect/predict toteboard patterns, etc. and building into it the ability to make automatic dutching calculations in real-time would be pretty cool...

[anotherdave]

Originally posted by GameTheory
You don't have to explain it here, but is there some way you could relate to me the more complicated formulas for 3 or more horses (the ones with matrices)? I'm making my own "personal toteboard" that grabs the odds/pool info off the internet and saves it to analyze/detect/predict toteboard patterns, etc. and building into it the ability to make automatic dutching calculations in real-time would be pretty cool...


I usually use the concepts for 2 horses and occasionally 3. Once I get to 3 horses I am making 3 guesstimates about "fair odds" and I think mistakes become far more likely. And as you said, if those "fair odds" are off, we've got trouble.

I could tell you the matrix set up if you like. You would need to be able to find the inverse of a matrix and then do a multiplication. I know you can do it on Quattro, can't remember if Excel does matrices? If you are interested still, I'll send it to you. (Unless there are others interested in the math , then I'll post it here.)

AD

[cj]

Keep one thing in mind with all those calculations...they work good on paper, but waiting until the last minute to bet due to fluctuating odds is hard enough without thowing an extra egg in the basket just before post. Perhaps a simple constant ratio such as 60%, 40% for two horses or 50%, 30%, 20% for three horses would suffice.

CJ

[GR1@HTR]

Play a tri

AB/ABC/ABCDE...
AB/ABCD/ABCDEF...

Where:
A=top ranked horse
B=2nd ranked horse
and so on...

Only if you can find value in either A or B selection. Looks like you had good value in your 2nd pick @ 5-1

[anotherdave]

Originally posted by cjmilkowski
Keep one thing in mind with all those calculations...they work good on paper, but waiting until the last minute to bet due to fluctuating odds is hard enough without thowing an extra egg in the basket just before post. Perhaps a simple constant ratio such as 60%, 40% for two horses or 50%, 30%, 20% for three horses would suffice.

CJ

You are right. The fluctuations in odds and the (sometimes) shakiness of the odds line I set can cause trouble. But usually the odds end up pretty close to what I figure on the 2 horses. A 60-40 rule isn't a bad idea. But you'd be amazed how many times I've ended up doing something more like 80-20, especially when I have a lower odds overlay that I really like and a promising longshot (breeding, trainer play, etc).

I should say that I probably play one horse to win 75% of the time and 2 horses the rest.

I hate to admit it but last year I was looking at a horrible 11 horse race - absolute dogs. I hated all the horses about equally. But the favourites had no reason to be 8-5 and 3-1 etc. I thought they were the worst of the bunch. My stupidest bet of 2001:
I took my $100 bet and bought 5 double-digit horses to win at $20 apiece(no I didn't use my formulas!). One of them won and paid $38.40. Still it was an insane thing to do.

AD

[Bob Harris]

Originally posted by cjmilkowski
I always find myself caught in a dilemma when I face the situation I encountered below. What do some of my fellow handicappers do:

Aqu 8th yesterday, I had the following line before the race:

6-103 Bet Odds 5-2 (went off at 7-2)
1-101 Bet Odds 4-1 (went off at 5-1)
9- 98 Bet Odds 9-1 (went off at 4-5)

I bet the 6 (and lost) while watching the 1 horse gallop at a nice price. This seems to happen to me more often than not when I have two or more overlays in the same race.

HELP!!!

CJ


CJ,

To really use a betting line properly there needs to be some sense of expectation...what approx. odds did you expect the three horses to go off at?

If the 1 horse had some general handicapping flaws which the public usually disdains and you expected it to go off around 8-1, the fact that the horse was drawing more money than anticipated could actually be a very positive sign.

The 9 horse going off at 4/5 after you made it a weak 3rd choice is also a big red flag. Was this a horse with huge Beyers which you expected to draw lots of attention or did the heavy betting catch you by surprise?

There really shouldn't be alot of surprises on the board in races which you bet into.

Have a great week,

Bob

[cato]

In theory I would have probably passed looking at those rankings and odds. Two very minor overlays and a 4-5 horse. Would be a pass for me.

Plan B would take the higher odds horse but at 7-2 versus 5-1, the horse withthe highest odds could be different every time there's a tote board flash...

Plan C would be to dutch the tow, but the odds just seem too low to justify the risk, unless you think the 4-5 horse had little chance. I mean, jeez, lets say you bet $2 on eah, you are risking $4 to net $8 or $7.

Pass

Cato :confused:

[cj]

I did think the 4-5 shot had little chance, and yes, I did think she would be heavily bet. That is why I was interested in the race. I use a mechanical program to come up with the odds line once I have a figure for all the horses. That is why the 9 was given a line at all. I knew she would never be a bet, and that nearly all others would be an overlay. These are the races I try to hit hardest, just missed this one!

CJ

cato...I don't consider them minor overlays, I already build in an edge over fair odds to establish the bet line...(25% first choice, 50% second choice, 75% third choice, etc.)

[Bob Harris]

I really don't see a big problem with what you did then, CJ. You evaluated the favorite (nice call by the way) and determined that you would likely have overlays with your other two selections. It looks to me like you made a good bet which just had a bad outcome.

Bob

[justin]

I distinctly disagree with the the post saying to pass. A heavy favorite when your own odds say 9-1 is a dream situation, especially when said horse is hung out and a known front runner. Judging by the charts, a hot pace should have been predictable judging by the very contentious opening quarter. I'm familiar with Dat You, so maybe I'm biased here. If you pass on situations where your numbers say the heavy favorite is a fraud, why make an odds line at all?

Justin