How Complex Should Your Method Be?

[Capper Al]

Is picking a horse with a simple method inferior to picking a horse with a complex method? Here's a story about a man who realizes he forgot keys when returning home. The only way for him to get into his house when locked out(without breaking a window) is to go into the garage get the ladder, put the ladder under his second story bathroom window, and giggle the screen out. After doing this he crawled through the window and into his house. When he came down to the main floor and into the living room, he noticed that his front door had been unlocked all the time. And same goes for many cappers who feel more confident with an E=MC^2 equation than a horse dropping in class and getting a top jockey today or any other straight forward approach. Balancing simple vs. complex is difficult and, might I add, more difficult than just complex itself.

Let's discuss how we handle or should handle, what I call, the handicapper's dilemma.

[classhandicapper]

I don't think it matters if your method is simple or complex. What matters is that there is value in the selections it produces.

Most people intuitively assume that finding value will involve greater complexity, but sometimes complexity adds small errors that can compound themselves into larger errors as you add layers.

[Pick 'em Charlie]

The simples are generally what we call angles. It all comes down to knowing your chances verses the post time odds on which way to go. Let's say both horse A and B are going off at 5/1. If your angle hits at 2/1 on A and your complex formula hits at 3/1 on B, the greater value will be with your angle on A. Normally, it won't be that clear cut, but this is the idea.

[Dave Schwartz]

Everybody wants simple solutions.

The reality is that most things that work well - whether in horse racing, business or science - are usually at least somewhat complex solutions.

[thaskalos]

"Everything should be made as simple as possible...but not simpler."-- Albert Einstein

[Track Collector]

IMHO, to be successful (i.e. profitable) one must be using predictive factors in a virtually unique way. Simpler ways are obviously more identifiable by more people, which is why they are almost guaranteed not to be profitable. Consider that a LOT of sharp folks have studied racing for years and years. Can it really be possible that that some successful simple methodology has been overlooked. I think not!

Uniqueness pretty much requires complexity when it comes to successful handicapping. Consider that one has a Win wager that yields a 5% profit even after one's $100 wager. Having as little as $100 to $200 more wagered on that horse by others who are familiar with your specific wagering methodology could cause your plays to lose money! See how hard handicapping is?

Perhaps one can find something simple that works for a short time, but long term it is doomed to failure.

[thaskalos]

IMHO, to be successful (i.e. profitable) one must be using predictive factors in a virtually unique way. Simpler ways are obviously more identifiable by more people, which is why they are almost guaranteed not to be profitable. Consider that a LOT of sharp folks have studied racing for years and years. Can it really be possible that that some successful simple methodology has been overlooked. I think not!

Uniqueness pretty much requires complexity when it comes to successful handicapping. Consider that one has a Win wager that yields a 5% profit even after one's $100 wager. Having as little as $100 to $200 more wagered on that horse by others who are familiar with your specific wagering methodology could cause your plays to lose money! See how hard handicapping is?

Perhaps one can find something simple that works for a short time, but long term it is doomed to failure.

Simple doesn't mean easy though.

I think we can all agree that losing weight is "simple"...and yet, the majority of the public is still walking around unhealthily overweight. "Success" or "failure" in something reaches far beyond our common definitions of "simple" and "complex".

[MJC922]

The 'method' I'm doing real-world testing with at the moment is so minimalist I'd be embarrassed to tell anyone what it is. It back tested and forward tested to positive ROI but there's also a minimum acceptable odds requirement which knocks it down to around 500 plays per month. For now I'm seeing what happens with $2 win bets and if that holds up I'll let you know. I'm not overly optimistic due to the odds requirement but anyway... the fact that something very minimalist can play a lot of races and neutralize the takeout seems to indicate that it's worth considering. What's key to this (I've found) is knowing precisely what a factor is worth. So many factors in racing are multicollinear that it requires quite a bit of effort to find out exactly what something is really 'worth' relative to everything else.

[BELMONT 6-6-09]

IMHO, to be successful (i.e. profitable) one must be using predictive factors in a virtually unique way. Simpler ways are obviously more identifiable by more people, which is why they are almost guaranteed not to be profitable. Consider that a LOT of sharp folks have studied racing for years and years. Can it really be possible that that some successful simple methodology has been overlooked. I think not!

Uniqueness pretty much requires complexity when it comes to successful handicapping. Consider that one has a Win wager that yields a 5% profit even after one's $100 wager. Having as little as $100 to $200 more wagered on that horse by others who are familiar with your specific wagering methodology could cause your plays to lose money! See how hard handicapping is?

Perhaps one can find something simple that works for a short time, but long term it is doomed to failure.

Spot on! Brings up the point that you must stray from the herd when handicapping...BUT NOT TOO FAR as to place the percentages against you.

[JJMartin]

There's already many simple ways to handicap a horse race. You could bet the morning line favorite. You could bet the top speed figure. You could bet the best finish. The only problem is that any of those methods will yield a negative ROI. In Horse Racing the real issue is that the general public already knows too much. You have to be willing to study the aspects that the typical player is unwilling to do. It takes a lot of work, there is no way around it.

[Robert Fischer]

"Everything should be made as simple as possible...but not simpler."-- Albert Einstein

That's basically the essence of it.

Similar to structuring a sentence. You don't want to make the sentence unnecessarily wordy, but a certain amount of words and structure are necessary to get your point across within the appropriate style.

[Robert Fischer]

You've got an understanding of multiple models that help us predict which horses have about how much of an advantage for different reasons.

You have a way of seeing which models are significant in this race.

You've got an understanding of how the public should perceive and bet the race.

You've got information on how the public is actually betting the race.

When you've got enough of the right models operating in the same direction, you get a 'critical mass' like effect. (sounds fancy/complicated, actually more simple)

More often you have conflicting tradeoffs. (difficult to accurately estimate)

[Secondbest]

:ThmbUp: "Everything should be made as simple as possible...but not simpler."-- Albert Einstein

[whodoyoulike]

If you think about it, everything is really simple if you understand it. It doesn't matter which subject you're discussing.

[Capper Al]

Even with my complex system, I will do a kind of Spring cleaning at times.

[Pick 'em Charlie]

This post should be in 'What's up with the Capper?' thread, but since we are talking about complexity here I'll post it here.

The rewrite is at a turning point. I have about a dozen ways to look at class and another dozen ways to look at speed and another dozen or more pace variables plus dozens of other variables. Then multiple all these variables by different methods of aggregation, such as best 2 of last three etc. I do follow these variable's outcomes, but find it is mind boggling to architect a single comprehensive approach. So it is time to narrow the process down. I've decided just to study the best 2 of 3 for aggregation and pick my top 3 elements for each category to continue to research. It never ends, but the rewrite should be an improvement with many more new elements incorporated.

And this is a familiar process, keep adding until I need to narrow down the scope again to what works the best- over and over again. winning isn't easy.

[Sinner369]

Charlie...........the only relevant criteria is which one works best for you or your system.

[Capper Al]

Charlie...........the only relevant criteria is which one works best for you or your system.

There's difficulty to getting down to the one that works best. Lots of R&D.

[PaceAdvantage]

Why are you posting under two different user names?

[JohnGalt1]

My results are worse when I consider or add more factors to my normal handicapping process, and/or when I play too many races.

One would think that pace/speed, class, form, trainer win percentages with specific conditions, and other positive factors would lead to more wins.

I HAVE to keep things simple for me to win.

For the last few years I have added trainer stats for certain categories. First and second after claim, winner last race, 90+ days lay off, and first time blinkers. Trainer must win at 30% with at least 4 wins. I also made note of jockey with trainer that wins at 30%.

I've been reviewing my results. Most of the time when I included horses in the above categories, they did no better than my normal handicapping. And, in fact, led me to include horses I shouldn't have, or skipped races that I should have played, whether I would've won or lost

I now note who trainers and jockeys are, but I will place more emphasis on the HORSE'S data.

Another issue I sometimes have is after many good days in a row selectively
playing only a few races per card, or skipping cards altogether, I play too many races with horses who are only marginally better than their competitors.

I get the God complex. And worse is when I WIN after playing too many races, because it doesn't last.

So simple method, simple bet structure works best for me, unless I stray off the reservation.

[Capper Al]

My results are worse when I consider or add more factors to my normal handicapping process, and/or when I play too many races.

One would think that pace/speed, class, form, trainer win percentages with specific conditions, and other positive factors would lead to more wins.

I HAVE to keep things simple for me to win.

For the last few years I have added trainer stats for certain categories. First and second after claim, winner last race, 90+ days lay off, and first time blinkers. Trainer must win at 30% with at least 4 wins. I also made note of jockey with trainer that wins at 30%.

I've been reviewing my results. Most of the time when I included horses in the above categories, they did no better than my normal handicapping. And, in fact, led me to include horses I shouldn't have, or skipped races that I should have played, whether I would've won or lost

I now note who trainers and jockeys are, but I will place more emphasis on the HORSE'S data.

Another issue I sometimes have is after many good days in a row selectively
playing only a few races per card, or skipping cards altogether, I play too many races with horses who are only marginally better than their competitors.

I get the God complex. And worse is when I WIN after playing too many races, because it doesn't last.

So simple method, simple bet structure works best for me, unless I stray off the reservation.


Have you ever played on only one factor?