Complicated/sophisticated ratings

[highnote]

Speed and pace figures are fairly difficult to create. So I was wondering if anyone uses a single factor that is as complicated to make and as useful as speed/pace figures?

I can't think of any off the top of my head, but I'm sure there must be some -- maybe a power rating?

How would you go about making a power rating or even some other complicated rating?

[aaron]

Barry Meadow uses his Master Win ratings for California tracks.

[blind squirrel]

BRIS "PRIME POWER"

[Zaf]

CJ's Performance Figure & HTR's K Rating :cool:

Proprietary.

Z

[orlando]

FAST FRED PCL RATINGS.

[ryesteve]

Proprietary.
Z
Ditto JRating... but Jeff might wander by and give a general answer to the "How would you go about it?" question.

[highnote]

Ditto JRating... but Jeff might wander by and give a general answer to the "How would you go about it?" question.

Thanks for those ideas. I wonder how the horse performs when all the services point to the same top ranked horse?

What does go into those rankings? Inquiring minds want to know!

[kitts]

All-In-One V6 has a Power rating (factors comprising it decribed in detail in the User Guide) and they also show a Double Advantage horse designation on horses that fit that criteria

[Bruddah]

software, JCapper, has an excellent power number rating. However, the software does so many excellent things.

[traynor]

The most important consideration in combined ratings (all the way back to John Meyer's adjustments of 0-Fin fps by class) is whether on not the combined values outperform the component values. False consensus can be unprofitably deceptive--an entry ranked best in a dozen different ratings may be beaten by a longshot that was first or second in another rating.

One way to determine results is to model each rating, and each combination of ratings, over time. In a lot of cases, it will be clear that the balance of combined ratings to mutuel payoffs negates any advantage gained in win percentage; everyone can see the standout entry, and bets it down.

The dichotomy is between win percentage and available mutuel prices--as one goes up, the other goes down (usually, but not always). The bottom line is that single, all-in-one ratings more often illustrate false consensus than predictive ability. If you think convenience is more important than profit, they are great.

Lastly, combined ratings (or "energy ratings" as the Sartinistas call them) are only useful to the extent they outperform the component ratings. If you can qualify minor track contenders with something simple like EPS or APV, then extract the winner from those contenders with Early Pace, it does not follow that combining EPS or APV with Early Pace will put you on more winners.

[highnote]

Thanks, Traynor.

Anyone got anymore compound ratings. I'll check all of these ratings out at some point. They are on my to-do list. I'd like to see how well they perform in the real world.

I wouldn't expect any of them to be profitable by themselves - except on certain populations of races. I don't think any single factor can outperform all the races.

Aside from the final odds, is there a single factor that can win more than 33% of all races?

[Jeff P]

I've spent a lot of my time this summer (when not immersed in live play) doing tons of research while developing and tweaking several different comprehensive Power Ratings. The algorithms I've been working on each weigh several hundred separate factors before spitting out a Power Rating. My current (Q3 2006) database contains approximately 5,000 races and 40,000+ starters. Don't be put off by the small sample size(s) I'm about to post. Databases from other time periods all show results that are nearly identical.

Here are results for one of the raw Power Ratings, for all horses, under all race conditions, in my Q3 2006 database broken out by rank:

By: Pwr Rating Rank (Raw)

Rank Gain Bet Roi Wins Plays Pct Impact
1 -671.70 10272.00 0.9346 1581 5136 .3078 2.4583
2 -1159.70 10276.00 0.8871 1026 5138 .1997 1.5947
3 -1631.90 10276.00 0.8412 748 5138 .1456 1.1626
4 -2393.10 10260.00 0.7668 555 5130 .1082 0.8640
5 -2158.60 10104.00 0.7864 443 5052 .0877 0.7003
6 -3347.90 9412.00 0.6443 324 4706 .0688 0.5498
7 -2057.30 7734.00 0.7340 224 3867 .0579 0.4626
8 -2074.30 5716.00 0.6371 126 2858 .0441 0.3521
9 -1742.30 3818.00 0.5437 66 1909 .0346 0.2761
10 -826.20 2382.00 0.6531 45 1191 .0378 0.3017
11 -570.30 1278.00 0.5538 19 639 .0297 0.2375
12 -319.40 698.00 0.5424 9 349 .0258 0.2059
13 -173.80 178.00 0.0236 1 89 .0112 0.0897
14 -112.00 112.00 0.0000 0 56 .0000 0.0000
15 -4.00 4.00 0.0000 0 2 .0000 0.0000
16 -8.00 8.00 0.0000 0 4 .0000 0.0000
17 0.00 0.00 0.0000 0 0 .0000 0.0000
18 0.00 0.00 0.0000 0 0 .0000 0.0000
19 0.00 0.00 0.0000 0 0 .0000 0.0000


One thing I've been experimenting with is the concept of using the collective intelligence of both the morning line odds maker and the betting public for the purpose of improving the Power Rating itself. I've discovered that taking the odds into consideration can improve the win rate. However, I've also discovered that doing this involves a trade off. Improvements in win rate almost always result in a decrease in roi.

Here are results for the Power Rating when Morning Line Odds are taken into consideration:

By: Pwr Rating Rank (Allowing for ML - Before Odds are Known)

Rank Gain Bet Roi Wins Plays Pct Impact
1 -1109.80 10272.00 0.8920 1626 5136 .3166 2.5283
2 -1243.10 10280.00 0.8791 1069 5140 .2080 1.6609
3 -1331.40 10274.00 0.8704 803 5137 .1563 1.2484
4 -2699.00 10260.00 0.7369 551 5130 .1074 0.8578
5 -2386.70 10114.00 0.7640 451 5057 .0892 0.7122
6 -2696.40 9460.00 0.7150 290 4730 .0613 0.4896
7 -2406.80 7734.00 0.6888 176 3867 .0455 0.3635
8 -2097.00 5758.00 0.6358 93 2879 .0323 0.2580
9 -1347.10 3792.00 0.6448 58 1896 .0306 0.2443
10 -631.40 2396.00 0.7365 35 1198 .0292 0.2333
11 -827.20 1398.00 0.4083 10 699 .0143 0.1142
12 -440.60 680.00 0.3521 4 340 .0118 0.0940
13 -10.00 86.00 0.8837 1 43 .0233 0.1857
14 -24.00 24.00 0.0000 0 12 .0000 0.0000
15 0.00 0.00 0.0000 0 0 .0000 0.0000
16 0.00 0.00 0.0000 0 0 .0000 0.0000
17 0.00 0.00 0.0000 0 0 .0000 0.0000
18 0.00 0.00 0.0000 0 0 .0000 0.0000
19 0.00 0.00 0.0000 0 0 .0000 0.0000
And just for comparison purposes, here are results for Morning Line Odds by rank:

By: Morning Line Rank

Rank Gain Bet Roi Wins Plays Pct Impact
1 -2195.00 10540.00 0.7917 1665 5270 .3159 2.5231
2 -1863.70 10558.00 0.8235 1072 5279 .2031 1.6217
3 -2522.50 11464.00 0.7800 829 5732 .1446 1.1550
4 -2790.80 11582.00 0.7590 607 5791 .1048 0.8371
5 -2085.30 10536.00 0.8021 440 5268 .0835 0.6670
6 -1638.70 9088.00 0.8197 272 4544 .0599 0.4780
7 -2248.40 7298.00 0.6919 139 3649 .0381 0.3042
8 -1973.80 5136.00 0.6157 75 2568 .0292 0.2332
9 -414.30 3278.00 0.8736 42 1639 .0256 0.2046
10 -764.90 1692.00 0.5479 17 846 .0201 0.1605
11 -423.30 880.00 0.5190 7 440 .0159 0.1271
12 -201.80 348.00 0.4201 2 174 .0115 0.0918
13 -94.00 94.00 0.0000 0 47 .0000 0.0000
14 -34.00 34.00 0.0000 0 17 .0000 0.0000
15 0.00 0.00 0.0000 0 0 .0000 0.0000
16 0.00 0.00 0.0000 0 0 .0000 0.0000
17 0.00 0.00 0.0000 0 0 .0000 0.0000
18 0.00 0.00 0.0000 0 0 .0000 0.0000
19 0.00 0.00 0.0000 0 0 .0000 0.0000

Also, just for comparison purposes, here are results for the Bris Prime Power Number covering the same set of races:

By: Prime Power Rank

Rank Gain Bet Roi Wins Plays Pct Impact
1 -1244.10 10384.00 0.8802 1658 5192 .3193 2.5502
2 -1533.80 10658.00 0.8561 1076 5329 .2019 1.6125
3 -1968.60 10784.00 0.8175 791 5392 .1467 1.1715
4 -2478.60 10818.00 0.7709 568 5409 .1050 0.8386
5 -2455.40 10454.00 0.7651 422 5227 .0807 0.6448
6 -2809.10 9494.00 0.7041 274 4747 .0577 0.4610
7 -2583.30 7422.00 0.6519 175 3711 .0472 0.3766
8 -1712.00 5350.00 0.6800 100 2675 .0374 0.2985
9 -1234.70 3398.00 0.6366 53 1699 .0312 0.2491
10 -472.50 2044.00 0.7688 32 1022 .0313 0.2501
11 -301.80 980.00 0.6920 13 490 .0265 0.2119
12 -395.00 494.00 0.2004 2 247 .0081 0.0647
13 4.60 170.00 1.0271 2 85 .0235 0.1879
14 -66.20 78.00 0.1513 1 39 .0256 0.2048
15 0.00 0.00 0.0000 0 0 .0000 0.0000
16 0.00 0.00 0.0000 0 0 .0000 0.0000
17 0.00 0.00 0.0000 0 0 .0000 0.0000
18 0.00 0.00 0.0000 0 0 .0000 0.0000
19 0.00 0.00 0.0000 0 0 .0000 0.0000

It should be clear from the roi that a good Power Rating can outperform the Morning Line.

Here are results for the Power Rating when post time odds are taken into consideration:

By: Pwr Rating Rank (Allowing for the Odds - After the Odds are Known)

Rank Gain Bet Roi Wins Plays Pct Impact
1 -1333.80 10350.00 0.8711 1784 5175 .3447 2.7531
2 -1431.20 10268.00 0.8606 1156 5134 .2252 1.7982
3 -2075.70 10292.00 0.7983 796 5146 .1547 1.2353
4 -2467.80 10224.00 0.7586 535 5112 .1047 0.8358
5 -3072.50 10176.00 0.6981 352 5088 .0692 0.5525
6 -2277.30 9458.00 0.7592 245 4729 .0518 0.4137
7 -2613.60 7676.00 0.6595 125 3838 .0326 0.2601
8 -1670.30 5692.00 0.7066 87 2846 .0306 0.2441
9 -465.40 3804.00 0.8777 52 1902 .0273 0.2183
10 -909.90 2412.00 0.6228 22 1206 .0182 0.1457
11 -457.70 1240.00 0.6309 8 620 .0129 0.1030
12 -357.30 742.00 0.5185 4 371 .0108 0.0861
13 -58.00 134.00 0.5672 1 67 .0149 0.1192
14 -60.00 60.00 0.0000 0 30 .0000 0.0000
15 0.00 0.00 0.0000 0 0 .0000 0.0000
16 0.00 0.00 0.0000 0 0 .0000 0.0000
17 0.00 0.00 0.0000 0 0 .0000 0.0000
18 0.00 0.00 0.0000 0 0 .0000 0.0000
19 0.00 0.00 0.0000 0 0 .0000 0.0000

And just for comparison purposes, here are results for Post Time Odds by rank:

By: Odds Rank

Rank Gain Bet Roi Wins Plays Pct Impact
1 -1702.80 10744.00 0.8415 1921 5372 .3576 2.8558
2 -1918.60 10182.00 0.8116 1070 5091 .2102 1.6785
3 -1872.90 10188.00 0.8162 785 5094 .1541 1.2307
4 -2632.90 10192.00 0.7417 511 5096 .1003 0.8008
5 -2410.30 10032.00 0.7597 363 5016 .0724 0.5779
6 -2649.10 9404.00 0.7183 224 4702 .0476 0.3805
7 -2614.60 7704.00 0.6606 131 3852 .0340 0.2716
8 -1184.70 5700.00 0.7922 78 2850 .0274 0.2186
9 -213.80 3798.00 0.9437 55 1899 .0290 0.2313
10 -1236.20 2382.00 0.4810 15 1191 .0126 0.1006
11 -275.50 1262.00 0.7817 10 631 .0158 0.1266
12 -351.10 676.00 0.4806 3 338 .0089 0.0709
13 -96.00 172.00 0.4419 1 86 .0116 0.0929
14 -92.00 92.00 0.0000 0 46 .0000 0.0000
15 0.00 0.00 0.0000 0 0 .0000 0.0000
16 0.00 0.00 0.0000 0 0 .0000 0.0000
17 0.00 0.00 0.0000 0 0 .0000 0.0000
18 0.00 0.00 0.0000 0 0 .0000 0.0000
19 0.00 0.00 0.0000 0 0 .0000 0.0000

A good Power Rating can also outperform the odds set by the betting public - at least from a flat bet win roi perspective. Although I am working on it, I have yet to produce a Power Rating where the top ranked Power Rating horse in the race wins significantly more often than the race favorite.

-jp

.

[highnote]

All-In-One V6 has a Power rating (factors comprising it decribed in detail in the User Guide) and they also show a Double Advantage horse designation on horses that fit that criteria


Can the user guide be downloaded from the website or do you have to buy the software. I couldn't find a link to the user guide.

[kitts]

Swetyejohn-

The user guide used to be for sale. You might check their website:
http://www.cynthiapublishing.com

There are 17 factors that make up the power ratings. You can modify the impact of any factor. I, for instance, zero the impact for Jockey, Trainer and Breeding.

If you get over to their website check out their "WMF" rating which can be a very potent factor in race shape.

[highnote]

Thanks, Kitts. I did check their website lastnight, but didn't see it. I'll check again. It sounds good.

I used to use All-In-One about 15 years ago! I think Dick or Steve actually helped me set it up. Time flies.

[kitts]

Steve and I started with Cynthia Publishing the same week back in the day. I am reasonably certain they answer their sales office phone and would be glad to sell you a User Guide. They would even offer you a deal on upgrading your old software, maybe. :o)

[highnote]

Thanks for the suggestion!

[thoroughbred]

[QUOTE=swetyejohn]Thanks, Traynor.

Anyone got anymore compound ratings. I'll check all of these ratings out at some point. They are on my to-do list. I'd like to see how well they perform in the real world.

I wouldn't expect any of them to be profitable by themselves - except on certain populations of races. I don't think any single factor can outperform all the races.

Swetyejohn,

Check out the "Boxer Number." You can view the derivation of this overall rating at out website: www.revelationprofits.com. On the home page click on "Documentation" and then on "Addendum to Engineering Analysis of Thoroughbred Racing." It's generally NOT to be used as a single factor.

[PriceAnProbability]

Speed and pace figures are fairly difficult to create. So I was wondering if anyone uses a single factor that is as complicated to make and as useful as speed/pace figures?

I can't think of any off the top of my head, but I'm sure there must be some -- maybe a power rating?

How would you go about making a power rating or even some other complicated rating?

You're talking about a power rating.

Generally, if your power rating method is good, the top rating will show a flat-bet profit, and the "value" horses (longshots near the top) will make for great singletons in races with no clear top rating.

[Dave Schwartz]

Generally, if your power rating method is good, the top rating will show a flat-bet profit, and the "value" horses (longshots near the top) will make for great singletons in races with no clear top rating.


You make this sound very easy. Have you found that to be the case?


Regards,
Dave Schwartz

[PriceAnProbability]

You make this sound very easy. Have you found that to be the case?


Regards,
Dave Schwartz

Found what? That the top power rating should be profitable? That's a given for any method.

Is making power ratings easy? It gets easier over time as you can use your experience to fine tune your method. I'm just saying that if your method is good, your top rating should show a flat-bet win profit.

It's never easy no matter what because even when you win you never really know if it was luck or skill.

[sjk]

Knowing whether its luck or skill is really not an issue for those that are winning.

I would agree with Dave that making a single power rating that provides winning bets is quite a challenge.

[classhandicapper]

A good Power Rating can also outperform the odds set by the betting public - at least from a flat bet win roi perspective. Although I am working on it, I have yet to produce a Power Rating where the top ranked Power Rating horse in the race wins significantly more often than the race favorite.

-jp

.

That's because the public is damn good. IMO, the public's mistakes are generally isolated to very specific circumstances. Plus, we as horseplayers also all make mistakes because we are working with incomplete knowledge and sometimes faulty information and/or understanding. The more of the public's typical errors you identify, the more you will be able to isolate the top ranked horses within your sample that are creating all the ROI value relative to the take you are seeing - because most of them actually don't have much value. Once you isolate them, you have a winning method.

[highnote]

I've never seen a power rating that makes a flat bet profit.

Maybe it could be profitable in certain, limited circumstances, but not in every race.

[Tom]

What he said. ;)

[JimG]

One thing for sure, if the power rating is published, it will not be profitable for long. If you have a rating where the top horse turns a profit, you best keep it to yourself.

[Dave Schwartz]

Truthfully, I subscribe to theory that everyone eventually boils it down to a single number for each horse.

For some people it is a "Aristotlian truth number." That is, it is a zero or a one, where ones are plays and zeroes are not.

For others it is a power number that refeclts a strength of horse.

For me it is a probability that sums to 100% and is used to make a line.

Whatever type of number you make/use/buy it is not the number itself that makes the profit. Rather, it is how you learn to use that number.


In the Basics of Winning
Handicapping
Exploitation (or wagering)
class I try to stress that winning demands a two-phase approach:

In the handicapping phase you assign your "numbers" to the horses.

In the wagering phase you determine the best way to take advantage of how the public has wagered relative to your handicapping. That is, you exploit their errors.

These two phases must be totally separate! One problem is that most handicappers try to mix the two.


Regards,
Dave Schwartz

[PriceAnProbability]

Knowing whether its luck or skill is really not an issue for those that are winning. I would agree with Dave that making a single power rating that provides winning bets is quite a challenge.

It's not an issue ***while they are winning***. It becomes an issue the seconed the "luck" ends, however.

It may be a challenge to produce a single power rating that is profitable, but today's game virtualloy demands it of those who are going to make money, and unless you want to not play doubles, p3, p4, p6, p9, and p-whatever-they-come-up-with-next, you also need to know how to handle every type of race, because they don't conveniently group the rolling p3 by class or surface or age.

Seven years after I began attempting this, my ratings are profitable enough that I could demonstrate them in public and not get eggs thrown at me, well not on most days. One thing about ratings is that because you are taking a stand on a horse's ability in that way, you're going to be wrong rather spectacularly, rather often. The other problem is that even one hole in your formula can be fatal (like the guy who wrote about how not understanding trainer patterns messes up his value line).

[PriceAnProbability]

That's because the public is damn good. IMO, the public's mistakes are generally isolated to very specific circumstances. Plus, we as horseplayers also all make mistakes because we are working with incomplete knowledge and sometimes faulty information and/or understanding. The more of the public's typical errors you identify, the more you will be able to isolate the top ranked horses within your sample that are creating all the ROI value relative to the take you are seeing - because most of them actually don't have much value. Once you isolate them, you have a winning method.

The public is "damn good" but the pros are damn better at keeping their mouths shut (or they don't stay pros).

Public mistakes are plentiful, but usually we are the "public" that makes the mistakes, like when Joe Bombdropper brings his one paydirt winner to Saratoga from a minor-league track and no one catches on for five or six years, or doesn't care even when they do.

Most of the public mistakes are what we call conventional wisdom.

[PriceAnProbability]

I've never seen a power rating that makes a flat bet profit.

Maybe it could be profitable in certain, limited circumstances, but not in every race.

Actually, people who use power ratings profitably prefer to bet every race, because that reduces the likelihood of bad racing luck or a bad outcome in one race hurting us.

A power rating system that cannot rate every type of race is incomplete, because horses leave those types of races and go into other types of races, so if you don't know how to make ratings for something, it'll come back to bite you even in your areas of expertise.

[PriceAnProbability]

One thing for sure, if the power rating is published, it will not be profitable for long. If you have a rating where the top horse turns a profit, you best keep it to yourself.

Beyer did okay by giving up the speed figure formula. The trick is to milk a book deal or a media gig out of publication, so you wind up making money on NBC before the Derby etc. Most players don't have that, so exploiting it at the windows is the primary option.

My concern about publishing the method I use isn't that the public would catch on fast, but that a few syndicate guys would maul it into neutrality/unprofitability in no time flat. At the end, I chose not to publish it, a decision that was reinforced by this very board a few weeks ago when I used an angle I found here to bump up a horse's rating (I had no opinion beyond the winner anyway), threw it into my key, and watched it get second at 30-1, enabling me to hit a $782 trifecta with a 7-5 on top and a 4-1 in third.

You also forgot the other problem that comes with publication, which is being flamed off of internet fora for being an arrogant, lying know-it-all who just wants to sell something.

[Dave Schwartz]

Price,

I am not trying to be difficult here but to clarify something:

You aren't saying that your top horse should be profitable in each race, are you?


Regards,
Dave Schwartz

[PriceAnProbability]

Price,

I am not trying to be difficult here but to clarify something:

You aren't saying that your top horse should be profitable in each race, are you?


Regards,
Dave Schwartz

I'm saying BOTH the top-rated horse and the "value" horse (if you know how to convert your ratings into an odds line), should show a flat-bet profit. Often they will be the same horse, but not always.

One theory I've never explored much but know has a value is that the top rating should be able to exert better control over the pace, and that may be worth a few points in and of itself (i.e., a "top rating bump"). Like if your top rating is 75, and nothing else is 68, the 75 might actually be a 77 or even an 80.

[Dave Schwartz]

Price,

Yes, that makes sense.


My personal style of play is to play several value horses in a race and add one hedge horse, then dutch the whole bunch. Had a race recently (12 horses) where there were 9 strong value plays plus the hedge horses.

Conversely, on many occasions we have a single value play and a single hedge.

Rarely we get a race where there is no hedge horse available. (This happens when the top 2-3 probability horses are all "value" plays.)


BTW, our purpose in hedging is to extract the profit earned by the value horses (usually longer-priced animals) without waiting for the price horse to actually hit.


Regards,
Dave Schwartz
Our purpose

[PriceAnProbability]

BTW, our purpose in hedging is to extract the profit earned by the value horses (usually longer-priced animals) without waiting for the price horse to actually hit.


I assume by "hedge" you mean horses who are contenders but underlays. I throw them in if I have extreme value in the other slots, but even then sometimes I don't play it safe, since that gets expensive if my main opinion is wrong.

Usually I put the value horses (those where my ML is lower than the program's) into the A group, and the remaining top ratings in the B group, though if a horse has a much higher rating than the filed, I might make it an A in the multirace wagers.

With so many races a day, however, it's not a big deal to sit through a "long" losing streak, and if you improve your ROI by cutting out the "hedge" horses (or top-figure underlays), you get your bankroll out of the danger zone even more efficiently. The general advice "don't try to buy the race" is my guide. If a horse is 4-5 with a 106 power rating and a few 97s are "value" I'm more likely to box all three rather than stick to my guns, but if that same 4-5 shot is a 99, he goes off my ticket completely and I'm more likely to toss in some bombs in search of a big score rather than try to ensure that I cash a ticket.

[Dave Schwartz]

What I heard you say (paraphrase) is that you have multiple strategies for how to exploit a race depending upon the race's individual scenario.

What you have described is a excellent and everyone should be aware of this approach.

Perhaps you might consider posting some specific examples (from the past) of these different scenarios and their applicable strategies?

(Hopefully I have understood you correctly.)



Regards,
Dave Schwartz

[PriceAnProbability]

What I heard you say (paraphrase) is that you have multiple strategies for how to exploit a race depending upon the race's individual scenario.

What you have described is a excellent and everyone should be aware of this approach.

Perhaps you might consider posting some specific examples (from the past) of these different scenarios and their applicable strategies?

(Hopefully I have understood you correctly.)



Regards,
Dave Schwartz

Every race, I am assigning a rating (probability) to every horse, and the rating is then converted into a ML value (price). Unless the program or the public matches me on every horse (has actually happened a few times), I will have a value play, and if I don't have a value play, then I just default to ranking them by power rating. Usually in turf-to-dirt races where there is no ML to guide the public, I go by rating and don't try to compare my odds line to the public's.

Was there a specific scenario you were talking about? Generally what I do is give each horse a power rating, convert the ratings into a morning line, and then bet the best value horses according to my line against the publics, while throwing in the top ratings underneath or in multirace tickets if they are different than the top values or lay over the field (like a 2-5 shot who "should" be 3-5 and is an underlay).

As for how to make the ratings, well if I published that, it wouldn't be profitable for long even assuming it's profitable now.

[highnote]

What's the most important factor that goes into making a power rating?

[PriceAnProbability]

What's the most important factor that goes into making a power rating?

Age.

As people already know, older horses are far different from younger ones, who tend to improve rapidly and must be viewed differently. The exception is when an older horse appears to be returning to its career form.

[Murph]

I've never seen a power rating that makes a flat bet profit.

Maybe it could be profitable in certain, limited circumstances, but not in every race.Here is the recent record for some thorostats top power selections. I cherry picked a few tracks that are showing a profit over the past month.

Races Winners Win% Avg $Won Won Cost Profit ROI
Belmont
199 64 32% $2.09 $416 $398 $18 4%

Beulah
45 19 42% $3.35 $151 $90 $61 66%

Finger Lakes - last month
189 60 31% $1.91 $362 $378 -$16 -4%
Finger Lakes - last week
45 21 46% $2.63 $118 $90 $28 31%

Hawthorne
199 59 29% $2.14 $425 $398 $27 6%

Keeneland
67 19 28% $2.36 $158 $134 $24 17%

Remmington Park
153 42 27% $2.39 $366 $306 $60 19%

Thistledown
148 53 35% $2.39 $353 $296 $57 19%

Zia Park
80 26 32% $2.65 $212 $160 $52 31%

ALL tracks, All races last week
1006 293 29% $1.89 $1,899 $2,012 -$113 -5%

Race sets larger than 200 races (see FL above) will rarely show a flat bet profit for our top power selection. For the past 12 months many tracks show win percentges in the 25%-30% range but cannot outperform the track take. The figures are similar to the above line for those race sets.

This seems to support swetyejohn about being profitble over every race. A power number that performs within the track take supports it's consistency in selecting contenders though.

Breaking down the races by conditions will quickly point to conditions where the power rating will outperform the take. This helps me decide how to proceed in isolating win contenders. Some races you like the top power runner and some you don't.

Murph

[Tom]

What's the most important factor that goes into making a power rating?

Marketing! :rolleyes:

[Murph]

Marketing! :rolleyes:EVEN when positive ROI expectations on a single win percentage are presented every day for FREE??

[Murph]

I would call that a percentage advantage!
Murph

[sjk]

Age.

As people already know, older horses are far different from younger ones, who tend to improve rapidly and must be viewed differently. The exception is when an older horse appears to be returning to its career form.

I have never differentiated age after Nov. 1 of the 3YO year. Do you find it profitable to do so?

Are you using actual birthdays (and if so can you point to a convenient source for this data) or the usual Jan 1?

[PriceAnProbability]

I have never differentiated age after Nov. 1 of the 3YO year. Do you find it profitable to do so?

Are you using actual birthdays (and if so can you point to a convenient source for this data) or the usual Jan 1?

Age is *always* an important factor.

Juveniles develop rapidly, and one must adjust their ratings accordingly. Be especially wary of a horse who wins one or two starts at two for a top barn and is laid up until a Triple Crown undercard where it comes back with a top jock.

Three year-olds also develop, but usually in one or two spurts.

Four year-olds also develop, but not as much as three year-olds, except on the grass. They often improve steadily and become more consistent.

Five year-olds can either start wearing from early competition (like Funny Cide), or they can be late-bloomers who become incredibly powerful (like John Henry did at six).

When making a horse's power rating, the very first thing I look at is the horse's age, so it is the most important factor.