- DRF Bets
- Handicapping & PPsThoroughbred Past Performances
ReportsPremium NewsDigital PapersHorsemen's Products
- DRF Classic PDF PPs
- DRF Formulator PPs
- DRF EasyForm PPs
- Daily Racing Program PPs
- Equibase PPs
- TrackMaster PPs
- NewsCategoriesTrack Notes
- DRF TV
- StorePast Performances
- Compare all DRF PPs
- DRF Formulator PPs
- DRF Classic PPs
- DRF EasyForm PPs
- Daily Racing Program PPs
- Expanded Closer Looks
- Equibase & Trackmaster PPs - Thoroughbred
Byron King: Two common handicapping angles can be misleading
Ask a group of horseplayers to name two factors deemed least important in handicapping a race, and there would be a wide variety of answers – everything ranging from connections to class, speed figures, and pace.
As for me, I’d say track bias and ground loss. I find them about as significant as noting wind speed and direction, which is to say, I don’t find them significant at all.
First, track bias. I do believe track biases can exist, but believe a study of track bias should be made over a wide sample of races, not based on an individual day, or even worse, a few races.
Anything can happen over a short sample. You can flip a coin five times and get heads every time, but that doesn’t prove the coin is rigged. It’s just chance.
Yet walk through the grandstand and listen to bettors after five races (or fewer), and you’ll hear “speed is dying,” or on polar-opposite days, “all you have to do is make the lead to win.”
In actuality, the winners of these races might simply have been best, or maybe, that the results were influenced by luck, just like the coin flip. Perhaps front-runners did well because the pace was to their benefit, or because they were favored to win.
Or if closers won, perhaps they did so because there were heated battles up front, or because the leaders had a history of puking in the stretch.
Ask any statistician: The smaller the sample size, the greater the margin for error. It’s true of political polls, and it’s true of track-bias analysis.
Ed Feng, a Stanford Ph.D. who runs the number-crunching sports website www.thepowerrank.com, detailed these sentiments in a column last year, titled “One thing everyone ought to know about sports analytics.” In this Google-worthy piece, Feng illustrated with computer-generated numbers how an “average” coach with hypothetical 50 percent chance of winning his NFL games, could in one sample random computer-generated trial, end up winning 31 of 50 games, or 62 percent.
But by continuing this trial over 200 games, this average coach’s win total dropped to 103 victories, a 53 percent win mark far more in line with what should have been expected.
Because horses, even favorites, have far less than a 50 percent chance of winning, there are even wilder swings to be expected in horse racing.
All that noted, I do believe in long-term track bias – those that are backed up by weeks of data, most notably one that has been long proven over time – that front-runners at nearly every dirt track in America are at an advantage.
But try to analyze track bias every day or every few races, and you are going to get unpredictable swings by chance, just as the stock market might go up or down from day to day depending on news that one often can’t predict.
As for ground loss in evaluating a horse’s trip, I believe by strictly looking at that number, whether it is noting a horse was three wide, or lost 20 feet via Trakus, is an oversimplication of a horse’s trip and too often ignores all the other things that happen in a horse race.
I didn’t flunk geometry. I know that if you have two equal runners racing a lap on a high school track, and one races in lane one, and the other in lane two, that the inside runner is going to win.
But horse racing is not that simple. It’s full of trade-offs. A horse tucked on the inside, but racing behind the pace, may have his move delayed by what unfolds with the horses in front of him, not to mention the fact that he may get discouraged from dirt being kicked back in his face.
Meanwhile, that horse stalking in the three path could enjoy a sustained clear run that the ground-saving horse did not, offsetting the ground loss he experienced.
Yet if a horseplayer looks at data from Trakus, or from speed-figures companies that factor in ground loss, the rail-riding horse won’t look as good on numbers as that clear-running horse that raced in the three path.
My opinion is simply this. There are good and bad trips with inside-running horses, just as there are with those that race on the outside. And the key is to watch these races individually and form a conclusion, and not merely rely on the number.
One caveat: I do believe ground loss to be of much more importance on turf because of the contrast of how these races are run relative to dirt races. Inside-rallying turf horses don’t have to deal with dirt kickback on the grass, obviously, and beyond that, turf races, more so than dirt races, tend to often come down to late tests of acceleration. So if a horse can save ground and quicken over a short distance, an inside trip under those conditions is beneficial.
mike oaktag I believe that a lot of handicappers forget about race dynamic's ( the way a race shapes up) but you can't tell me their's no such thing as a bias aided racetrack all you have to do is remember joe king from nyra those balcony tracks at belmont were fun!!
do u also take into consideration that the first 5 races may have had a closer or closers as the favorite, possibly even odds on in each race. a horse winning at 20-1 on front end 5 races in a row, over an odds on fav in each race,could very well be a speed bias. EVERYTHING HAS TOBE TAKEN INTO ACCOUNT
The idea that 200 games somehow clarifies a coach's ability to win a game is absurd. Teams play 16 regular season games a year in the NFL and up to an additional 4 games in the playoffs. Obviously over a stretch of 10-12 seasons, the make up of a team is going to vary widely. Even if a coach stays with the same franchise for that length of time, the personnel will have changed. Certainly there will have been more than three featured running backs, there either will have been significant turnover at the important quarterback position, or at least the franchise quarterback will have risen and fallen in his ability at the job. A smaller sample size will tell you how that team and its coach are doing now better than some distorting long view.
sometimes jockeys think thier is a speed bias and put thier horses on the lead....and then win. However often the horse wins simply because he was the best horse and would have won comming from 7 lengths back. This results in a false confirmation of a speed bias. Royal Delta in the BC Distaff is a good example.
I think thier is a BIG difference between having good speed and dueling 4 wide... then rallying 4 wide on the turn.
I've seen far too many days with a complete bias that this article make me scratch my head that it was actually published. Trying to bring statistical sample size into the argument to make it a legitimate argument seems to be reaching to make the point. Sure the best horse might win every race that day and those horses might all have the same running style, and maybe there are zero other horses in the race that had a chance to win with a different running style....sure....but I think your statistical sample size argument argues against that being a likely possibility. Ok maybe on some days you can't jump to a conclusion because the 1st 4 races were just stacked....but give the average handicapper a little credit to recognize thats the case. I'm not a big ground loss person at all and agree with this point. But your bias argument is poor at best. Tracks change frequently...weather, maintanence, etc all come into play on a daily basis. For some reason I think the article is simply trying to show that speed rules here in NA..and to prove that take a look at long term track bias...and ask any statistics professor to back up the claim. Instead try looking at breeding and the typical distance of races here and you'll get a better answer on why front runners are at an advantage. It has little to do with statistical track bias. Bias just helps further the stat that speed rules.
I agree with this article up to a point: the greater the sample size from which to derive information, the more accurate the information will be as regards indicating a 'trend,' in this case track bias. Take Philadelphia Park (Parx) as an example: the inside has been dead there for a long time. Jockeys seem to avoid it like the plague. If your horse is stuck on the inside, good-bye money: a horse will rally from the outside and beat you almost every time. Here's where I disagree with the article: Jockey's get hot; trainers get hot. For whatever reason, they're 'in the zone' for some period of time. Look at Rosario this spring. His win percentage at the Keeneland meet will not be reflective of his percentage for the entire year, but if you were betting Keeneland at that time you should have been cognizant of the fact he was on fire. Ed Feng's ideas are not wrong but I think they're useless from a handicapping standpoint. And we're all aware of the trainer who 'gets hot' (or gives hot shots): look at the recent Aqueduct numbers of the two NY trainers Jacobson and Rodriguez: whatever they claimed over the winter and ran right back won at an alarming rate. This statistic is not reflective of their over all training stats, but as a bettor betting at THAT TIME you did well if you were aware of this. Can anyone see how what I am saying contradicts the basic thesis of King's article? I took a lot of time to articulately compose my thoughts and I would appreciate comments. Thank you.
jockey is last thing to consider when handicapping
I thank you very much for ignoring track bias. Someone has to bet the wrong horse!
ok Preakness weekend ....turf course, any closers win? last Saturday notacatbutallama sure looked good on a fair course... tomorrow lets see how embarr does at colonial.