Understanding Streaks

[sq764]

I am not sure why, but lately my play has been so streaky. Not so much profit/loss for the day, but streaky in the sense that I can lose 8 straight, then hit the next 3. I know the length of the streaks has some dependency on your method of play, but its just odd.

I am not a big believer in ‘luck’, so I wonder if we just go through streaks of where our wagers are sharper than other times, whether we are zoned in better during certain times or whether its just the moons being in alignment.

Personally, I think its that I structure wagers better during certain times, which is when the nice hits come..

[penguinfan]

Check the kind of races your playing, during a losing streak are they maiden races, cheap claimers, routes etc... and when your winning see what kind of races your playing then too, may surprise you what you find out.

Penguinfan

[Dave Schwartz]

SQ764,

We had a user once that noticed this phenomenon during the summer months, especially the negative side. We traced it to his poor performance in maiden special races with lots of firt-time starters. Since many tracks card 2 or 3 of these per day, with a non-maiden in between, it would ruin his pick-three plays. He found that his overall win percentage had only dropped from 51% to around 48% but his pick-three hit rate had dropped from around 13-14% to 9%.

My suggstion? Look at the race types you are playing in whatever system you use to track your bets.


Regards,
Dave Schwartz

[ranchwest]

A couple of things come to mind.

As others have mentioned, the types of races being played may have something to do with it. Remember, the races are written by the racing secretary and there may be a logical explanation why you find several races in a row difficult and then find several in a row to be easy (maybe they really are). The card may be written that way.

The other, as you suggested, can be that you get in a zone at a certain point. If you are usually doing well after a few races go by, try regarding those first few races as a warm-up and just put $2 to win on one horse on those races and once you feel zoned in then get on with your normal pattern.

[sq764]

Maybe I phrased my post wrong, as I don't think I was getting my main point across.

I was trying to see if everyone believes there is a such thing as luck involved. Not so much in winner by DQ or something strange happening in the race that ends up in your favor. I was moreso thinking of getting the close photo in your favor or a trifecta coming back unusually large.

[GameTheory]

Of course there is luck. Just start flipping a coin. You're gonna have a roughly 50% win rate -- probably much higher than your hit rate in racing. But it won't be at all unusual to see many in a row all tails or all heads. With racing the "natural" streaks will be even longer. (Or just sit down with the charts and count how many non-favorites win in a row. The high numbers will surprise you.) If you calculate the length of streak to expect based on your usual performance stats, you'll have a better idea when you're having one that is an anomaly.

[Rick]

There have been a number of studies that show that people don't expect random results to be as streaky as they really are. In fact there's an entertaining trick that's done in math classes sometimes where half of the class creates real random results and the other half makes them up. By just looking at how many streaks there are in the data you can classify them as either random or faked a surprisingly high percentage of the time.

The first pro I ever knew kept years of records and would go back and look at them every time it seemed like the losing streaks were too long. He'd always be able to find something more bizarre that had happened in the past, and just looking at how the streak was eventually terminated in the past gave him more confidence in the present. Of course, just looking at any sequence of truly randomly selected results would show you the same thing.

That being said, there certainly are reasons, mostly seasonal, that would cause the results to be a little more streaky than random. I just wouldn't overemphasize their importance in most cases.

[Rick]

For those who think that pace handicapping is new, it would probably be interesting to read about a group from the 40's called "The Speed Boys" led by a guy named Jule Fink. Damon Runyon helped make them well known to the racing public and they gained legendary status similar to that attributed to the Sartin group. It's surprising how many ideas that you still read about in handicapping books come from some of these guys from way back. Pittsburg Phil was another one of them whose ideas are repeated over and over.

[thoroughbred]

In addition to the reason for the streaks, pointed out by others, which are valid, i.e., different types of races, there is also a far simpler explanation for streaks.

It could be caused by the degree of handicapping skill together with randomness.

To explain what I mean, let me pose a hypothetical. Suppose after you fininsh handicapping a race, your degree of skill is such that the probability of your choice winning is 50%. (We should all be so fortunate :) )

Then, in the same way that streaks occur in the tossing of a coin, so will streaks occur in handicapping.

I used 50% skill level only as an example to relate it to coin tossing. The principle is the same regardless of the degree of skill.

So, in summary, streaks are very natural phenomena, and we needn't look to luck for an explanation.

[Rick]

Oops, that last post wound up in the wrong thread.

As for streakiness, I tested my own handicapping selections and found only a very slight, insignificant amount of streakiness using a statistical test for serial correlation. But my selections require no subjective judgement whatsoever and, from what people are always saying here, that's not the most popular way to handicap.

[David McKenzie]

I'll ride the tangent.

Originally posted by Rick
...But my selections require no subjective judgement whatsoever and, from what people are always saying here, that's not the most popular way to handicap.

It might be more appropros to say it's difficult at best for most to come up with anything cabaple of doing that. It's not that they don't want a system that works. It's more like they can't believe any system actually produces profits over the long run.

Personally, I have no reason to doubt you. We all saw your high Win % when you were trying out for the Interboard contest.

And if it's working for you, who gives a whit what others think anyway?

[Rick]

David,

I don't know anything about the contest you mentioned. Although I've entered a couple of minor things over the years, I don't think anyone would have noticed them. I'm not crazy about having to analyze multiple tracks simultaneously so you probably won't see me entered in much of anything. And in most contests you have to alter your thinking to play all longshots and hope that nobody else has them.

My opinion is that most people would do better if they followed a consistent set of rules. The rules don't necessarily have to be simple but they should be consistent in similar situations. Otherwise, how do you know what works and what doesn't? What good would it do to win some races but not know why you won so you could repeat the process in the future?

But a lot of people are probably more consistent than they think. They just may not know the rules they're following well enough to teach someone else or to program them into software. I have personally never known anyone who wins that doesn't follow a consistent set of rules of some kind, but I can't prove that it isn't possible. I'm no handicapping genius but I am pretty good at analyzing horse racing data without making the typical mistakes.

[Rick]

By the way, my results are not all that amazing. If I played every top rated horse in every race I'd have 29.66% winners and 1.93% loss. But, by eliminating morning line favorites and requiring one other qualifying rule I get 26.77% winners and 20.21% profit on the "overlays" that I bet. I'm positive that some of you have methods that lose less than 5% on all top rated horses that would win at least 10% on selected overlays in 1/3 to 1/2 of the races. It's not that big a deal. Eliminating underlays is the most important part.