Friday, January 11, 2008

The Bettor: Optimal Bet Size and More

I am the king of procrastination today. I work for myself. I should fire me for being lazy, but since I am so lazy right now I will wait to fire myself Monday.

I have chatted with Jeff Platt, founder of jcapper, a betting software. He is a sharp bettor, and has been working honing his craft for many, many years. His software is commercially available if anyone is interested.

Today he was chatting about optimal bet size, and how smaller pools and watered down product hurt the business. It is an area we somewhat explored by chatting with Ian Meyers of Premier Turf Club here.

Jeff posted up his thoughts and I asked if I could reproduce them here. He said "sure". It is a decent look at how players - serious players - approach the game, and why when pools get small, wagering gets smaller. It is a pillar in the building block belief that when a track loses handles it can snowball, and why rebating can make tracks handles' grow. It is a brilliant illustration as well that when players are handcuffed, the incentive is to bet less and that is something we obviously do not want.

It also stresses that our pools not being big enough can make people play other games, in many cases the stock market. I played fairly seriously at that game for years and I have some thoughts about that versus racing, and how I think their business plan (online trading) has hurt racing immensely (they did what we should have done, imo). I will do a post on that soon.

In addition, there seems to be this obsession at times in "getting people back to the track". Often times it is pure fallacy with big players. Jeff lets his thoughts be known on that.

For a background illustration on rebates and lower rake, check our post here.

Anyway, thanks to Jeff for these thoughts. It is a very enlightening and a nice way to glimpse into a serious players mind; and it teaches us things. Education is a good thing, very few people want to share it in betting.

Here it is:

I agree with you that rebates (or lower takeout) causes large bankroll players to approach the game differently and bet drastically more. And IMHO that's good for the game.

Why racing execs fail to see this is completely beyond me. If guys like Daruty and Coutow got their wish and managed to make the concept of ADWs and rebates evaporate overnight and force guys like me back to the track... one thing they fail to understand is that there is no way in the world I can bet anything close to my current handle on location at a track compared to what I now bet online. They need to face facts. Pandora's box is open. I'm not going back to the track. Force me back to the track and instead they will force me out of the game. I'm already dabbling successfully in swing trading... a game I consider to be far and away easier to beat than parimutuel betting. Take away rebates and where do they think I'll focus my attention?

I'm going to make a guess:

At 20k+ betting a night it isn't that hard to overload the pools at many of the smaller pooled tracks to the point where you are actually betting against yourself. As you you no doubt know there is an optimal or max bet size that each pool will support.

For those who may not be aware I'm going to take a stab at illustrating this.

Optimal bet size is the bet size that produces not max roi but max profits for the player. It's based on pool size, takeout, and the strength of your play. In a perfect world you could calculate it ahead of time. But in real life, because you can't know the exact ending size of the pool with your bet in it before you bet, you can only make an educated guess as to what it actually turns out to be.

I'll use the following chart to illustrate my point. Here I am simulating the performance of WIN bets at track with a win pool similar in size to that found at GGX or BM. Using a UDM or spot play that has on paper produced a 1.15 roi over a race meet with 200 hypothetical plays where takeout and the average pool size dictates each additional $100 bet lowers roi by 2 percent, here's what happens:

Bet Amt Net
Size Plays Bet ROI Profit
$10 200 $2000 1.15 $300
$50 200 $10000 1.14 $1400
$100 200 $20000 1.12 $2400
$200 200 $40000 1.10 $4000
$300 200 $60000 1.08 $4800
$350 200 $70000 1.07 $4900
$400 200 $80000 1.06 $4800
$500 200 $100000 1.04 $4000
$600 200 $120000 1.02 $2400

Notice that max roi happens with tiny bets. As the player increases the size of his bet his roi drops. I say this shouldn't bother the player one bit. Why? Because as bet size increases so does net profit - but only up to a point - the point of optimal bet size. However, when bet size is increased beyond optimal bet size then net profit starts to decrease. In this example optimal bet size happens somewhere between $300 and $400. The same UDM (note: UDM is an acronym for "user defined model", something Jcapper allows you to do) played at a different track with a larger pool (or a lower takeout) would of course offer a larger optimal bet size and more net profit across a sample of 200 plays.

Depending on the pool and the venue it may be entirely possible for the large player to take home more by betting less.

For more thoughts from Jeff, or to try a demo of his software (thoroughbred software only), please visit him at Jcapper.com.

1 comment:

Ryan said...

Did you end up doing a post on comparing online trading's strategy? I would be interested in reading that one, but cant seem to find it. Thanks!