Handicapping average: What's possible in terms of % ?

[the_fat_man]

We know for certain that one can win at the races using computer generated models that look for overlays. (Benter et al)

Now,the majority on this forum seem to favor implimenting a particular strategy (computer generated, databases, etc.) over many race tracks-- the more bets, the better.

There are others (the minority?) that consider handicapping an art. The idea being that the development of one's skills are directly proportional to the time/effort invested. This approach would involve trip notes, trainer stats, paddock/prerace warmups, workouts, etc. Rather than betting as many races as possible, this handicapper would be limited to best bets only --say 10-20 bets max per week, per track. The player

The baseball player, for example, reaches a consistent level of proficiency --a 300 hitter can be counted on to hit 300 most years. Why wouldn't the same apply to the handicapper?

So, what can someone following the handicapping as art approach, which would probably limit one to a single or at most 2 tracks, expect in terms of win percentage?

is 30-35% winners with $7-$8 average payout realistic?

and if so, why isn't everyone playing this way?

[GameTheory]

and if so, why isn't everyone playing this way?You can make more money with the volume approach. Much much more. Especially with rebates...

[the_fat_man]

You can make more money with the volume approach. Much much more. Especially with rebates...

Let's work this out, assuming, from your statement, that you believe the win percentages and return rate given above are attainable.

A single track

15 bets per week x 50 weeks = 750 win bets

let's set the standard wager at $500 (x 750) = $375,000 wagered per year

let's use 33% winners at $7.50 average payout

750 x .33 = 248 x 7.50 = 1860 x 250 = $465,000

thus $90,000 profit

and the rebate can be applied as well

pocket change to Benter and his type

pocket change to the average 'pro'?

[Fastracehorse]

We know for certain that one can win at the races using computer generated models that look for overlays. (Benter et al)

Now,the majority on this forum seem to favor implimenting a particular strategy (computer generated, databases, etc.) over many race tracks-- the more bets, the better.

There are others (the minority?) that consider handicapping an art. The idea being that the development of one's skills are directly proportional to the time/effort invested. This approach would involve trip notes, trainer stats, paddock/prerace warmups, workouts, etc. Rather than betting as many races as possible, this handicapper would be limited to best bets only --say 10-20 bets max per week, per track. The player

The baseball player, for example, reaches a consistent level of proficiency --a 300 hitter can be counted on to hit 300 most years. Why wouldn't the same apply to the handicapper?

So, what can someone following the handicapping as art approach, which would probably limit one to a single or at most 2 tracks, expect in terms of win percentage?

is 30-35% winners with $7-$8 average payout realistic?

and if so, why isn't everyone playing this way?

As a longshot player - and a spot player - your average payout is conservative but your win % is bang on.

fffastt

[sjk]

I would have no idea how to win 30% of races bet with good results. I do fine with a win pct of 11-13%. Horses (or combinations) that I think offer a positive expectation just do not win 30% of the time.

What percentage of races are won by horses that are "overlays" given the usual information set? GT can tell us if that is an answerable question.

[twindouble]

I just can't phantom the idea of betting many races, many tracks in one day. To me it would take everything I enjoy out of the game. Those that do, I trust that they can make money doing it as they say. Mainly because I know how important it is to have a reasonably good bankroll regardless of how good anyone is or the software.

Your win percentage can vary a lot from day to day week to week, even when it's low you could very well hit some huge scores that bring up your ROI so I don't concern myself with it. I've had many periods where I did a lot damage to the bankroll but was able to pull it out eventually. Then there the times when I would have had to do a lot dumb ass things to give it all back. Not being that dumb ass enables me to stay in the game.


Good luck,

T.D.

[hurrikane]

I'm not so sure you can do 1 or 2 tracks and get that many bets, in that range, at a track that can handle $500 win bet, in a year. The people I know that play the way you say and bet that kind of money are lucky to get 6-7 win bets down a week.

So far this year I avg about 90-100 bets a month, that is every track, every race, every day I am mining looking for opportunity. I'm sure you can do many many more than that but I have some pretty strick rules to play by.

1 or 2 tracks. I could not find 15 playable races a week that the fav is not going to win. Back before simulcast when I was playing 1 trk a day I would be doing well to find 1 or 2 prime bets a day.

[GameTheory]

I suggest watching Dave Schwartz play the next time he has one of his video conferences. He is betting 2-4 horses in practically every race he looks at in the win & exacta pools. And he doesn't even have to look at the race until post time...

[Fastracehorse]

I'm not so sure you can do 1 or 2 tracks and get that many bets, in that range, at a track that can handle $500 win bet, in a year. The people I know that play the way you say and bet that kind of money are lucky to get 6-7 win bets down a week.

So far this year I avg about 90-100 bets a month, that is every track, every race, every day I am mining looking for opportunity. I'm sure you can do many many more than that but I have some pretty strick rules to play by.

1 or 2 tracks. I could not find 15 playable races a week that the fav is not going to win. Back before simulcast when I was playing 1 trk a day I would be doing well to find 1 or 2 prime bets a day.

6-7 bets a week that is.

fffastt

[Valuist]

There is no one correct way. The best player I know plays only turf races on multiple circuits and then any kind of race on his main circuit. Another pro I know plays only New York and Chicago racing. These are not programmers. That's not to say that software can't work; just that its definitely not the only way.

[skate]

this subject is very elusive in its attempt to become clear.


all depends, experience, bankroll, luck and experience and bankroll.

i've been at my best while hitting at 12% and i'm expecting for my win % to drop to around 10% to 11% and my profit to go up.

i'm not "fun" at the track. while playijng 3 or 4 races on a weekend.

skate

[thoroughbred]

Studies have shown that under the following conditions:

1.- If you do only a win bet on a single horse in each race, and,
2.- If you bet every race,

Then,

The win rate maxes out at about 40%.

I've thought about this a lot and have come to understand why. It's analagous to the reason no baseball player has had a seasonal batting average above 400, i.e. 40%, since Ted Williams did it in the early 1950's.
That's over half a century and no one else has done it.

[the_fat_man]

Studies have shown that under the following conditions:

1.- If you do only a win bet on a single horse in each race, and,
2.- If you bet every race,

Then,

The win rate maxes out at about 40%.

I've thought about this a lot and have come to understand why. It's analagous to the reason no baseball player has had a seasonal batting average above 400, i.e. 40%, since Ted Williams did it in the early 1950's.
That's over half a century and no one else has done it.

I wonder if you wouldn't mind expanding on this a bit.

More about the studies (empirical or strictly statistical)

and the analogy to the 400 club in baseball?

For example, why did so many (relatively) hit 400 pre 1940's and none after?

[thoroughbred]

I wonder if you wouldn't mind expanding on this a bit.

More about the studies (empirical or strictly statistical)

and the analogy to the 400 club in baseball?

For example, why did so many (relatively) hit 400 pre 1940's and none after?

Hi the_fat-man,

A number of years ago, I wrote a simulation program to learn what kind of profit could be expected, on the average, based on an individual's handicapping capability, i.e., what his win rate is.

On the old Prodigy Bulletin Board, I asked for, and received statistical data from thousands of races, breaking down the percentage of times horses with different final odds won. I set this up for random occurrences, and also had a random generator that would win at a rate consistent with a choosable win rate. In this simulation, every race was bet on only one horse to win, and every race was bet.

Results went something like this. If the win percentage was less than about 29-32% or so, there was an expected loss. At higher win rates, of course, proftits were generated. And, the maxing out at 40% was also a result.

Interestingly enough, after I got these results, I was told, that some others had already done similar research and came up with the same 40% maximum.

It would take quite a bit of discussion to do full justice to the baseball analogy.

Here's the gist though.

To get into the big leagues, as a hitter, you need to have at least something like a 290 average. And, something like over 300 is good.

Now to be good baseball hitter, in the major leagues, you need good attributes for eye-to-hand coordination, strength, agility, and many more. So if you look at the general population, and, graph the standard normal curves for such attributes, those that make it into the big leagues, are very likely to be in the upper percentiles for each attribute. When such batters face the typical pitchers, the end result of the interplay between batter and pitcher leads to the typical 300 or so batting average.

But, Ted Williams was an exception. He had something extra. Every so often, because I look for it, I see a newspaper reference to this fact about Williams. He entered the Navy, in Florida, during World War II, and it is reported that no one since has ever equalled his eye sight capability. So, he had this one extra superior attribute that let him break through the 400 barrier.

In some similar way, the interaction of handicappers, faced with all the variables of horse racing, if they bet every race, and ONLY bet to win, will, on the average, not exceed the 40% win rate.

I need to emphasize that these are very specialized conditions, e.g., betting every race. If the simulation were designed to accommodate other betting strategies, the results would surely be different. But, at least, this provides some idea of possible win rates.

As an aside, I'm not sure that there were many 400 hitters before Williams.

[Jeff P]

For example, why did so many (relatively) hit 400 pre 1940's and none after? Two reasons. First thing (just my opinion) is equipment. Ever seen or better yet worn any of those old style gloves? There isn't a whole lot to them. Imagine yourself putting on a plain leather gardening glove and then trying to spear a hot smash while diving to your left. Far easier to do with a Wilson A2000 and some elbow and knee pads. I think a lot of balls got through the infield 75 years ago that go on the scorecard as ground outs today. Second (again just my opinion) pitching. Today's pitchers are taught very early on to throw stuff (split finger/cutters, etc) that didn't exist 60 years ago. I think they change speeds and disguise pitches better than the pitchers of yesteryear. Also, in the old days pitchers started and finshed their games and even pitched again the next day. When they got tired they just continued to throw the ball. Ever tried to throw a curve ball after your arm has turned to jello? It's called a hanger. In today's game a hitter who plays a full game is likely to face 2-3 different hurlers. When these guys come in from the bullpen they're fresh - and quite often far tougher than a tired starter.


I tend to agree about 40 percent acting as a pratical upper limit for win percent under the conditions that you have to limit yourself to using past performance data only and single one horse in each race and play every race. The reason it isn't higher, aside from things like trips and biases, is that no matter what the past performance lines say, not all horses are physically ready to run on race day. If you are good at reading the way horses look and act and use that to make play or pass decisions on your otherwise 35-40 percent past performance selections, then I think 50-60 percent winners is very realistic. Likewise, I have seen and used a handful of highly selective computer generated models where you pass 90-100 races for each one that you play. These models are mechanical in nature and are quite capable of selecting between 65-70 percent winners strictly from past performance data alone. Combine that with the ability to read a horse's body language and it's pretty realistic to get 80-85 percent winners (at very low win mutuels) from a very tiny and select handful of plays.

-jp

.

[toetoe]

Fats,

25% @ $9.00 I would consider fantastic. 12.5% profit before rebates.

When does DS do video conferences? I'm interested.

[Fastracehorse]

Two reasons. First thing (just my opinion) is equipment. Ever seen or better yet worn any of those old style gloves? There isn't a whole lot to them. Imagine yourself putting on a plain leather gardening glove and then trying to spear a hot smash while diving to your left. Far easier to do with a Wilson A2000 and some elbow and knee pads. I think a lot of balls got through the infield 75 years ago that go on the scorecard as ground outs today. Second (again just my opinion) pitching. Today's pitchers are taught very early on to throw stuff (split finger/cutters, etc) that didn't exist 60 years ago. I think they change speeds and disguise pitches better than the pitchers of yesteryear. Also, in the old days pitchers started and finshed their games and even pitched again the next day. When they got tired they just continued to throw the ball. Ever tried to throw a curve ball after your arm has turned to jello? It's called a hanger. In today's game a hitter who plays a full game is likely to face 2-3 different hurlers. When these guys come in from the bullpen they're fresh - and quite often far tougher than a tired starter.


I tend to agree about 40 percent acting as a pratical upper limit for win percent under the conditions that you have to limit yourself to using past performance data only and single one horse in each race and play every race. The reason it isn't higher, aside from things like trips and biases, is that no matter what the past performance lines say, not all horses are physically ready to run on race day. If you are good at reading the way horses look and act and use that to make play or pass decisions on your otherwise 35-40 percent past performance selections, then I think 50-60 percent winners is very realistic. Likewise, I have seen and used a handful of highly selective computer generated models where you pass 90-100 races for each one that you play. These models are mechanical in nature and are quite capable of selecting between 65-70 percent winners strictly from past performance data alone. Combine that with the ability to read a horse's body language and it's pretty realistic to get 80-85 percent winners (at very low win mutuels) from a very tiny and select handful of plays.

-jp

.

Corking the bats that is :)