Question for the database people

[Valuist]

Curious if someone could run numbers on:

LONE 3 year old in a race, non-claiming races only.

[sjk]

I get a win pct of 17% but a return of -29%.

[gm10]

I get a win pct of 17% but a return of -29%.

Is there any month-to-month evolution?

[sjk]

month win pct AvgOfreturn CountOfDATE
1 0.00% -100.00% 23
2 21.95% 0.37% 41
3 12.56% -40.05% 199
4 17.58% -26.57% 364
5 15.64% -44.17% 697
6 17.16% -36.23% 711
7 16.69% -28.99% 665
8 20.00% -21.77% 590
9 20.29% -7.77% 478
10 14.71% -40.75% 435
11 20.37% -8.18% 324
12 13.49% -33.65% 215

Pardon the formatting. Seems to be consistently bad. Few 3YO and up races in Jan-Feb so the sample size is small.

[Valuist]

Can we break it into dirt and turf, sprints and routes? The win percentage is solid; apparently they are always heavily bet.

[sjk]

SURFACE Expr2 win pct AvgOfreturn CountOfDATE
DIRT R 18.25% -26.84% 778
DIRT S 18.62% -27.42% 2976
TURF R 13.27% -35.08% 324
TURF S 11.30% -39.26% 664

[Valuist]

Thanks.

No angle should be bet blindly since most aren't going to produce long term profits. That 18% win on dirt is impressive, even though it isn't profitable. But in multirace wagers, the lone 3YO on dirt should always at least be considered.

[Valuist]

SJK-

One more question: does "dirt" include synthetic?

[castaway01]

Thanks.

No angle should be bet blindly since most aren't going to produce long term profits. That 18% win on dirt is impressive, even though it isn't profitable. But in multirace wagers, the lone 3YO on dirt should always at least be considered.

If bets on these horses not only lose, but lose over 50 percent more than the average win takeout, why should they always be considered? The results were strongly negative. I'd think they should NEVER be considered rather than ALWAYS be considered.

[cj]

If bets on these horses not only lose, but lose over 50 percent more than the average win takeout, why should they always be considered? The results were strongly negative. I'd think they should NEVER be considered rather than ALWAYS be considered.

This is often a mistake people make. The "average win takeout" doesn't mean much. Run a query making a bet on every horse and see what you get for an ROI. It will be considerably worse than the average takeout.

I'm in no way saying this is a good angle, but it isn't as bad as that would lead some to believe. It is probably just a hair worse than random, which is still bad.

[Valuist]

Put it this way: if a trainer had similar numbers (18% win percentage and a similar ROI), would you not ever bet him? When playing bets like P4s, you have to sometimes include horses that might be a bit overbet but that have good chances to win.

[sjk]

SJK-

One more question: does "dirt" include synthetic?

It does.

[therussmeister]

This is often a mistake people make. The "average win takeout" doesn't mean much. Run a query making a bet on every horse and see what you get for an ROI. It will be considerably worse than the average takeout.

I'm in no way saying this is a good angle, but it isn't as bad as that would lead some to believe. It is probably just a hair worse than random, which is still bad.

I believe it will be somewhere in the neighborhood of 0.80

[cj]

I believe it will be somewhere in the neighborhood of 0.80

It is worse.

[castaway01]

This is often a mistake people make. The "average win takeout" doesn't mean much. Run a query making a bet on every horse and see what you get for an ROI. It will be considerably worse than the average takeout.

I'm in no way saying this is a good angle, but it isn't as bad as that would lead some to believe. It is probably just a hair worse than random, which is still bad.

I know I was oversimplifying, but just trying to clarify why the OP thought these horses ALWAYS had to be considered. It's sort of hard to sum up the results of an angle that makes no sense vs. the negative results a database that I don't have, but the point that the horses aren't exactly making much money for bettors was pretty clear.

[Valuist]

I know I was oversimplifying, but just trying to clarify why the OP thought these horses ALWAYS had to be considered. It's sort of hard to sum up the results of an angle that makes no sense vs. the negative results a database that I don't have, but the point that the horses aren't exactly making much money for bettors was pretty clear.

How many angles can produce 18% winners? My guess is that at one time, it was probably solidly profitable. As time goes on, these horses have been getting heavily bet.

Since the dirt numbers include synthetic, and since synthetic plays more like grass, have to wonder if the exclusion of synthetic races would be closer to 20% winners.

[lansdale]

Thanks.

No angle should be bet blindly since most aren't going to produce long term profits. That 18% win on dirt is impressive, even though it isn't profitable. But in multirace wagers, the lone 3YO on dirt should always at least be considered.

Hi Valuist,

This has been an interesting thread. But it reminded (as if I needed it) how tough this game can be, and how effective the precision of a database can be in correcting even seemingly sound assumptions. Over the years, I've heard many players mention similar angles, which seem logical, yet these guys never made money long-term.

If you look at the example you cited, the angle has a hit rate of .18 or 9/2. But the public bets it at an average of roughly 4-1. That's really not much of an error, 1/2 an odds tick, but it's big enough to produce a hefty loss.

Another aspect of the likely unprofitability of this angle - it's very simple, very easy for the public to pick up on. IMHO, in this game, there is no longer any low-hanging fruit.

Cheers,

lansdale

[Valuist]

I want to point out, I have never advocated blindly betting this or any other angle. That is extremely lazy handicapping. Angles should be used to support an opinion. Like I said in an earlier post, if you liked a horse that was a second time starter and you see the trainers numbers are 18% but a negative ROI, would you automatically not use it just because THAT ONE angle said it was a negative ROI? Of course not. If I liked a horse that was the lone 3 year old in a dirt (real dirt) race would I back off because this one angle said it had a negative ROI? Of course not. What if the horse figured to have a pace edge on the field? What if the trainer was in the midst of a hot streak? How about if the horse ran a decent race last out despite being on a dead rail?

Granted before I saw the numbers I would've expected a better ROI from this angle. But I also would've expected a lower win percentage than 18% (dirt).

[Valuist]

Another point is that the differences in win pct between the dirt runner sample and the turf runners is significant. I would've anticipated win % of around 12% for each, meaning the turf 3 YOs have no advantange at all but the dirt 3yos are winning about 50% more than they could've been expected to.

I would have to think it is because one advantage 3 year olds have is less wear and tear on the tires. But grass runners tend to stay sound so the older grassers can keep their form longer so the 3 yo turfers do not win at a higher than expected rate. Curious to see if synthetic numbers would mirror the grass......while not a fan of synthetic it could provide some (anecdotal) evidence that synthetic racing can keep horses sound longer.

[windoor]

Two more to go.

Add in a factor (with a value # if required) and start over.

There is gold in them there hills.

Regards,

Windoor