Tuesday, March 11, 2014

The 411 on Bias

This article originally appeared in December 2013's Horseplayer Monthly magazine.  To read that issue, or any of the previous issues, chalk filled with handicapping articles and stats, please click here. It is 100% free.  

In November 2013 a reader of the Horseplayer Monthly asked this question, "I would love for someone to explain the track bias and how to identify it. How could there be an inside bias and the seven and eight horses are winning?" 

We asked several horseplayers to share their thoughts on this question, and their answers are as follows. 

Ed DeRosa - I think many handicappers put too much emphasis on track bias. It's too often seen as an oracle to explain why a certain result happened rather than just as another piece of the puzzle. 

Is it a speed bias when a horse goes gate-to-wire after setting a pace with fractions below par? I don't think so, but let's take an extreme example. Let's say there's a Sprinter showcase day at a track featuring six graded stakes races all going six furlongs. In every race the horse with the lead coasts through an opening quarter in :25 and a half in :49. Is it a speed bias, then, when these horses are all able to win after such soft fractions? Of course not, but if you showed those charts to a group of handicappers, I guarantee most would conclude that the track "favors front runners." 

That's not to say biases don't exist, but most people use that word as a catch all to explain away anomalies when really it's just physics at play. To take another extreme example, think of a half-mile harness track. Is there an inside bias because posts seven and eight are so bad? No. It's just common sense that starting that far out at full speed with the first turn rapidly approaching isn't ideal. Without question certain tracks favor certain dynamics, both in terms of pace and position on the racetrack, but strange weather aside, those dynamics are more typically long-term than a short-term bias.

Melissa Nolan - Track bias occurs when either certain areas of a track are winning more than expected, or certain running styles are winning more than their fair share of races. For instance, perhaps the most well know track bias occurred at Keeneland on the main before the Polytrack was installed in fall 2006. Deemed the "Golden Rail," it was well known that speed horses who got to the rail had a tendency to keep going when their past performances indicated otherwise. Essentially horses could "ride the bias" to victory. A famous example of the “Golden Rail” was Sinister Minister, who wired the field at 8-1 for Bob Baffert and Garrett Gomez in the 2006 Bluegrass Stakes. 

Another example was the track at Santa Anita on the first day of the 2013 Breeders' Cup where it was also well-perceived as speed biased though surface maintenance that evening resulted in a much fairer surface on Breeders' Cup Saturday.

As opposed to the old Keeneland surface where it was advantageous to not only be on the lead but on the rail, the main track at Parx Racing is notorious for having a deep and slow "dead rail."  At that track, it doesn't so much matter if you horses are speeds or closers as long as your jockey knows to get a few paths off the rail towards the middle of the track.  I've heard Jerry Brown of ThoroGraph mention that he gives more "dead rail" denotations at Parx than any other track in the United States.

Being able to identify a bias and which horses were helped or hindered by a track bias/profile can be very profitable if you can identify what's occurring before other handicappers.  If horses are running on when you think they should be stopping, or stopping when on the rail at comfortable fractions, you may be encountering a bias.

Pay attention and you may be laughing all the way to the IRS windows when others are still wondering why no horses can close. 

Jeff Platt – HANA President - As a horseplayer, once you recognize a path bias, part of your job is predicting which of the horses are likely to take advantage of it.

Take the case of a bias where the footing along the rail (or inside) is better than the footing on the outside. Such a bias does not necessarily mean that horses with inside posts will get the best of it. Quite often races unfold in such a way that the horse with the best early speed, even if that horse drew an outside post, gets over to the rail - and on days when the inside is best - that's the horse most likely to take advantage of an inside path bias.

The Cangamble Blog - The naysayers have gone extinct or changed their tune, track bias exists.   There are two types of track bias that could affect the results on a given day, inside/outside and running style (speed/closer). 

Sometimes the rail is better than average, sometimes the outside has an advantage.   Sometimes speed horses have an advantage, and of course there are days when closers do exceptionally well.  And many times, there will be two biases that seem to be prevalent, for example, speed-rail, or outside-closer. 

The problem with bias is that it can be very subjective and sometimes a correct assessment is hard to figure out.  Having an objective number for track bias can be very useful in explaining past races, but also, if calculated after three or four races on a live card, can be very advantageous to the horseplayer.

In the handicapping book, Power of Early Speed, Steve Klein came up with a simple way to objectively come up with a running style bias.  Here is a variation on it, as well as a way to tackle the rail/outside bias as well, and shouldn't take more than a couple of minutes to calculate the bias for a card. 

A couple of things, turf and dirt are separate when doing bias numbers, and sometimes due to wind or weather, a bias can change halfway through a card, but this will become apparent when looking at your work, and the result will be two sets of bias numbers for that specific day. 

For the running style bias, add one point if the winner was positioned within the top half of the field, first call (that shows lengths beaten).  In a field with an odd number of horses, a horse positioned in the exact middle is considered to be in the top half.  Add one point if the horse who finished second was positioned in the top half of the field first call.  Add another point if the winner was on the lead first call.  And finally, if the race favorite started in the back half of the field and didn't finish in the top three, add another point.  Repeat this for every race on the same surface for the card.  Divide the total number by the amount of races used.  If you are OK looking at a number that has decimals, round off to two decimals, if you don't like decimals, multiply by 100. 

Some tracks have a general speed bias and some do not, by comparing the number daily with numbers from the same track, you'll have a good idea very quickly how much of a bias existed.

When it comes to inside/outside bias, the best way to go is make notes, or watch replays, but if you want a decent general way to figure it out within a minute or two, using the following works well:  You may want to differentiate between one and two turn races, and perhaps only rate one turn races.  Add one point if the winner had a post on the inside (again, if it is a nine horse field, for example, the 1-5 posts are considered inside).  Add one point if the second horse had an inside post.  If the favorite didn't finish in the top three, look at the comments, if they give you the impression the horse was three-wide or greater, add another point.  Finally, look at the comments for the winner.  If the impression from the comments and the horse's post are that the horse was three-wide or more, subtract a point.  Once more, take the total number from all the races calculated for that day, and divide by the number of races used. 

Finally, for calculating live biases, you'll need to take notes if you are at the track and/or try to watch the replay as well.  This is a much easier task if betting from home, but if you have internet access at the track, you can also use what is available at Equibase, for example, as the charts come out pretty quickly these days.

No comments: