Bet Sizing Exotics

[DeanT]

Hey guys,

I am stuck on something, bet sizing exotic betting. I have an account just for win, and I am fine with that. I bet no more than about 4% for a win bet size, slide it based on edge, hit rate, odds level, etc. It's easy and a nice way to keep track, by only betting win in an account.

For another account, I want to exotic bet. Supers, some pick 3's and 4's and all the rest.

The problem? I constantly go broke in the exotic betting account. I assume it is because I am undercapitalized, not playing proper bet sizing. Or alternatively, I suck.

Anyway, if you started with a $2000 bankroll, in an account for just exotics, how would you approach playing it? How would you estimate edge and all the rest on the phenomenon of exotics where low hit rate and high bankroll swings rule the roost? What % max would you make on an exotic bet, like a tri, or super, based on that bankroll size? How do we avoid ruin in such a highly vague wagering environ like exotics, where hit rate and edge are so hard to define?

I am interested in hearing some opinions, so please offer them if you'd like.

[motorhead]

hello

first and foremost, you need (to the best of your ability) to establish a win percentage. a long range win percentage. that way, you will have a handle on just how bad losing streaks can be.

for example, if you have a 20 percent win average, do you know what percent of the time you will lose 10 or more consecutive bets? (take a guess, the answer is down below)

of course, how much you return when you cash a ticket is important, but if you overbet, in the end, it wont matter

the answer? 10.74% of the time, you will lose 10 or more in a row. surprising? i think it is for most players.

if you can find a copy of jim bayle/sport stat's "gambling percentages and probabilities, you will get a better handle on them

good luck

[cmoore]

10 cents supers is where you want to start..Whatever races your best at. Mdns, Claiming, Alw, sprints,routes or dirt and turf..Find the 10 cents super races for your strong point and focus on those. You can make some good size tickets for 10-20 bucks and hit at a good rate..

[Jeff P]

You should be able to look at your records (provided you break them out by bet type) and get answers to questions such as:

Q. What's my (insert name of exotic bet type here) hit rate, avg payoff, and longest losing streak?


Once you know that information, sizing exotic bets isn't really all that different than sizing straight bets.


-jp

.

[Robert Fischer]

Anyway, if you started with a $2000 bankroll, in an account for just exotics, how would you approach playing it? How would you estimate edge and all the rest on the phenomenon of exotics where low hit rate and high bankroll swings rule the roost? What % max would you make on an exotic bet, like a tri, or super, based on that bankroll size? How do we avoid ruin in such a highly vague wagering environ like exotics, where hit rate and edge are so hard to define?

I am interested in hearing some opinions, so please offer them if you'd like.

my approach is to make a safe estimate for my exotics hit% and payout.
then i plug it into these value equations


value = ($2price*hitProbability)/2
edgefactor = hitProbability*value

Lets do an example

superfecta
structure: A x BC x BC x DEF = $6
hitProbability= .40*.50*.40*.40 = 0.032 or 3.2%
$1SFC payout = (at least) $200 ... 2$Price =$200/(6/2) or 200/3= $66.66
value = $66.66*0.032/2 = $1.06:ThmbUp:
edgefactor = 0.032*1.06= $0.03:ThmbDown:

Summary
So what we have here is a Superfecta, with a nice high value. It also comes with a relatively low hit probability. The low hit% is going to drive down the "edgefactor". Edgefactor is simply multiplying Value by the hit% to help weigh value with risk of ruin.
I advise using a minimum edgefactor for most wagers. However with a $2000 bankroll devoted exclusively to exotics, you may choose not to pass the really nice Value here. $1.06 is a nice investment opportunity(remember you made conservative estimates and have the skill to be reasonably accurate on average). One approach is to use "edgefactor" to determine betsize.
For every 10 percent of "edgefactor" use 1% of bankroll
0.1*Edgefactor*BANKROLL=betsize .003*2000=$6
$6 happens to be the minimum $1 wager for SFC A x BC x BC x DEF, so it works out nicely. Round down, so if it had been $11, you still only have 1 $6 ticket to buy.

Say it had been a dream exacta overlay with a high edgefactor of $0.20 = betsize =.02*2000 = $40

I did a more complicated pick3 example here http://www.paceadvantage.com/forum/showpost.php?p=500107&postcount=19


EDIT- I should say that in Low hit rate/ high value situations, if you are really concernede about "edgefactor", there is often some trade off available if you include more horses. What happens is you buy a higher hit% in exchange for slightly less value. The larger your bankroll the less you want to do something like this.

[stu]

Executive Summary: With a $2000 bankroll, you should only bet $1/$2 exactas/doubles and no other exotics.

Explanation:
Here is one approach:

I would recommend calculating bet size based upon your tolerable risk of going broke. Tapping out of your $2000 bankroll. Let's assume that you are willing to tolerate a 1 in 1 million chance of tapping out.

The percentage of your bankroll that you should play on any wager should be equal to or less than the 1 in 1 million risk. I would then use the Poisson distribution to calculate this percentage.

{The math is left as an exercise for you}

Bet size = (-1/6)*log(1-P) where P is the probability of hitting your wager. If you can stomach a 1 in 1000 chance of tapping out replace the 6 with a three.

Let's say that you like a horse to win with a twenty percent chance.

P = 1/5

Betsize = (-1/6)*log(0.8) ~= 1.6% or $32

This way you have a one in a million chance of losing 62 wagers in a row and tapping out assuming that you accurately assessed the chances of the horse.


Let's say that you like an exacta with two horses that you think have a twenty percent chance of winning. The Harville equation says that:

P = 1/5 * ((1/5)/(1-1/5)

or

0.05

then Bet size = (-1/6)*log (0.95) = 0.37%

Thus $7

Let's say that you like a Pick-3 with three horses with a twenty percent chance of winning their respective legs.

P = 1/5 * 1/5 * 1/5 = 0.008

Betsize = 0.0581%

Thus $1

As the probability of each horse decreases so does everything else.

My heuristic simplifies to:

One horse wagers get 1% of bankroll
Two horse wagers get 0.1% of bankroll
Three horse wagers get 0.01% of bankroll
Four horse wagers get 0.001% of bankroll
Five horse wagers get 0.0001% of bankroll
Six horse wagers get 0.00001% of bankroll

Or

You need a minimum of $200 to win wager minimize the risk of tapping out.

You need a min. of $2,000 to bet exactas, quinellas, doubles.

You need a min. of $20,000 to bet trifectas and Pick-3s

You need a min. of $200,000 to bet supers and Pick-4s unless they are dimes then you only need $10K.

Adjust accordingly to your hit rate of exotics and minimum bet sizes.



Just one man's opinion.

[formula_2002]

I look at it this way;

If you could break even in the win pool, your edge is about 17%.

If you could break even in the exotics, your edge is about 25%.

each results in an ROI of 1.00

:rolleyes:

[Niko]

Can't include anything on bet size but a couple questions based on records. Bet size is the second part of the question. Sounds like your fundamental question should be...do you have a positive expectency in your exotic bets? If you don't then it's simple.... bet the smallest amount you're comfortable with losing for the day and have a good job to replenish the bankroll.

If you have a positive expectency disregard the rest of this.

As everyone points out exotics are a very different beast then wins.

There's so many questions....Are you including enough horses in your P3's, P4's? Or not enough? Are you singling in the first leg and spreading in the last leg everytime eroding your edge, hoping to cash? Are you playing into 1 race you know nothing about or two? Can you single in one leg, play 2 strong horses in another leg, spread in the 3rd and show a profit? If you have a longshot as a key are you trying to get too cute going for the big score and leaving out obvious horses causing you to lose over time? Have your read Exotic Betting by Steven Crist?

How about trying exactas or trifectas instead playing into the same race you have an edge in.

Exactas--Are you good enough to key just 1-2 horses under your horse and win? Longer run-outs but higher profits for some, less for others depending on your strengths, personalities and records. Do you need to include 5-6 underneath for a profit and stagger your bet amounts to mirror that of your win bets. For example if your horse is 4-1 bet different amounts on the exacta horses to show a similar or higher return over time. You may bet $1 combo's on the longshots underneath and $2-$3 to stronger combinations?

I don't know if this helps but Exotic Betting covers it pretty good.

[DeanT]

Thanks everyone for the excellent responses. I am going to fire up Bettor Keep track to look at some of the longer term numbers and see if I can do some math.

I am getting tired of betting say $100 win on something, hitting it and then looking up to see I bet $25 in exotics that puked. 25% in exotics for $100 in win is obviously too much, imo, any way I slice it. The problem with me is that I look at my exotic betting account as "fun money that I will lose", because my other account is my "real" account. Your posts are helping me shake that. I have to be a better player.

Thanks again, and I will run these numbers and see what I can come up with.