Posted By : S.K.SAGAR surendraksagar@hotmail.com 01/07/2011 IMPORTANCE OF TRACK CHARACTERESTICS IN EVALUATING TIME FORM RATINGS     This article is the fourth in a series of articles on the technical aspects of racing and may be read in conjunction with the first three articles available in Guest writer`s column of India race viz `Time Form Rating`, Impact of Penetrometer reading and False rails on Track Index`, and `Unanswered questions’ posted on 13-06-2010, 26-07-2010, and 17-08-2010 respectively. The basic requirement for a system to yield `absolute` values of the Time form rating of horses is that either the track conditions must be perfect, or they must be perfectly understood. For comparative values it might be enough to know the Short Term Track Index `Ts` which is a measure of the Penetrometer reading values that determine the day to day variations in the track conditions caused by rain and storms. But for `Absolute` values it is imperative to know the long term Track Index `Ti`(Defined as the distance travelled by a horse rated 150 on the TF scale carrying 120lbs weight in a time of one minute on that track when that track is at its fastest with least Penetrometer reading value)) of the course for all distances, and for this purpose all the track characteristics such as the geometric layout indicating the profile in plan with radii of all curves, the undulations and gradients, and the ground characteristics (nature of soil, specially the top layers), etc etc are required to be known and understood. In my third article entitled `Unanswered Questions` I had asked several questions on these aspects, to the various turf authorities. No answers were forthcoming. Here, I would like to say that I recently started analyzing the 2006 performance of ‘Deep Impact` in a race called `Teno Sho` at the Kyoto race course (Japan) where he ran 3200m in 3m 13.4sec, an unbelievably fast time for that distance and in my quest to determine the TF rating of DI, someone from the JRA (Japanese racing authority) actually helped me in this connection by giving valuable inputs (Refer `Deep Impact – Tribute` on YouTube). I will discuss Deep Impact`s case later in the article. Coming back to my `UQ`, While our Turf Clubs were busy not answering my questions, I started my own research by studying the timings for various distances for various classes in particular the record timings in order to arrive at the long term track index. I did this for the Bangalore race track, and observed that the timings for shorter distances (1100m) are relatively unimpressive by about 2.5 to 3% when compared to those of middle distances and those of longer distances (2400m) relatively unimpressive by about 1 to 1.5 % when compared to middle distances. What does this mean: Obviously that there is a certain downward gradient from some point near the middle distance zone for about half the track perimeter and a corresponding rise in the remaining half. To some extent the sharp curves may also contribute to this. These assumptions were more or less found to be accurate when after extensive search on Google I came across the following para pertaining to the Bangalore race track: "The Bangalore Race Track is an oval-shaped, right-handed course measuring approximately 1950 meters with four sharp curves and pronounced gradients. The downhill backstretch drops 13.10 (43 feet) meters from 1800 meters to 800 meters and climbs 11.53 meters (38 feet) from that point to the winning post with a further rise of 1.5 meters (5 feet) from the winning post to the 1800 meters. This demanding and testing race track, with its gradients, bends and a distinct short straight, places a premium both on the speed and endurance of the horse and the skill and experience of the jockey; a win on this race track is therefore a significant achievement" When I correlated this info with my earlier findings and assumptions I was able to work out the Long Term Track Index (let’s call it `Ti` ) for various distances for the Bangalore race track, which I give below: 1100M : 1048m, 1200m :1055m,1400m : 1066m, 1600m : 1077m, 1800m :1082m, 2000m: 1077m, 2400m: 1072m, 2800m : 1066. Please do not ask how the above values have been worked out from the given gradients of the course. There are enormous complexities involved. The methodology used is purely empirical (and not based on a scientific study of motion along inclined surfaces).In broad terms I have considered a gain in time @0.45secs per 100m from 1800m to 800m, a loss in time @ 0.55 secs per 100m from 800m to wp, and @0.3secs per 100m from wp to 1800m. In addition a loss in time of 0.75 secs for each semi circle curve is also considered. It has also to do to a considerable extent with the correlation between standard timings and ratings for various grades/classes of horses, it goes without saying that some intuition was also necessary in arriving at the final figures. In a similar way in a separate write up I had worked out the `Ti` value for the Epsom Derby course for the mile and a half as 1056, which gives fairly accurate results. With these values of the Long term Track indices, it is now possible to calculate the Absolute value of the Time form rating of a horse based on a specific performance at the Bangalore race track. This is best explained by an example: Consider Siachen`s win over 1200m on Sunday 5th June 2011 carrying 60 Kgs weight and clocking 73.75 secs, Penetrometer reading on the day = 4.20cm Please note the following assumptions made while evaluating the TFR: a) False rails if provided are taken care of by reducing the distance travelled by adjusting the location of the starting stalls. Alternatively if false rails parameters (width) etc are given and the precise location of starting stalls is given, then it is possible to calculate the exact distance travelled by geometric calculations. b) It is assumed that the horse while running along the curves maintains a steady and average distance of about 2m from the rails. If you are a smart observer and know the precise route (trajectory) followed by the horse, then you are in a position to determine the precise distance travelled by the horse. c) The rating does not take into account: 1) The reserve potential of the horse, ie how handy it was near the winning post or 2) The pace of the race – the fractional timings etc - ie whether it was a slow run race or a fast run, or 3) The fitness of the horse. If you are an expert in this category and have a good reliance on your judgment, you might consider all these aspects and conclude that the horse may have cut a faster time by such and such an extent. Factors `a`, `b`, `c`, referred above, including many other factors correspond to the uncertainty element of the race. It is here that the caliber or the expertise of an observer can make a difference while giving a figure by which the timing of the horse must be adjusted, to take into account all the uncertainties. The most important aspect to be remembered is that even if you are the greatest of experts in the line and have made an absolutely perfect evaluation of the Time form rating of a horse even to the nearest 0.5 point, there is no guarantee that the next time around the horse will run true to its rating. At best you can only fix probabilities of success. Coming back to the example in question: First the formula: TFR = 2.22*W + 30 – 22222*( T – T1)/D W = Weight carried in kgs which in this case = 60 D = distance travelled in metres which in this case = 1200 T = the time taken to traverse the distance which is to be adjusted to take into account the long term track index (due to permanent track characteristics) as well as the short term track index (due to the conditions prevailing on the day, determined by the PR Reading). A track index of 1090 corresponds to a perfect track in this regard. For the Bangalore race track the short term track index ( say Ts )values may be arrived at by reducing from 1090 in the following way ( refer 2nd article in the series, however the values have been revised slightly due to additional inputs on the track characteristics received since then ): a) For PR reading (say `p`) less than 2.5cm: No reduction required. b) For `p` between 2.5 and 3.5 reduce Track Index by (p – 2.5)*5 c) For `p` between 3.5 and 4.0 reduce Track Index by 5 + (p – 3.5)*10 d) For `p` between 4.0 and 4.5 reduce Ti by 10 + (p – 4)*25 e) For `p` between 4.5 and 5.0 reduce Ti by 22.5 + (p – 4.5)*35 f) For `p` exceeding 5.0 reduce Ti by 40 + (p – 5.0)*50 In the current example with PR =4.20 cms, the short term reduction in the track index value = 10 + (4.2 – 4.0)25 = 15 which means the short term index (Ts) = 1090 -15 = 1075. Long term track index (Ti) for 1200m for the Bangalore race track = 1055. Hence adjusted time T = 73.75*(Ts/1090)*(Ti/1090) = 73.75*(1075/1090)*1055/1090) =70.40 secs. Values of T1 corresponding to a TFR of 150 for various distances is given below (see also 1st article in the series, some values revised slightly) : Distance: 1000m Time T1: 54.7 secs Distance: 1200m Time T1: 66.4secs Distance: 1400m Time T1: 78.4 secs Distance: 1600m Time T1: 90.7 secs Distance: 1800m Time T1: 103.2secs Distance: 2000m Time T1: 115.9 secs Distance: 2200 m Time T1: 128.8 secs Distance: 2400m Time T1: 142.0 secs Distance: 2800m Time T1: 169.0 secs Distance: 3200m Time T1: 196.7 secs Note: T1 for a particular distance corresponds to the time taken to cover that distance by a horse rated 150 on the TF scale, while carrying 120lbs (54.5kgs) weight on a perfectly fast track having a track index Ti of 1090m. T1 for the current example = 66.4 secs Hence Time form rating of Siachen is worked out as: 2.22*60 + 30 – 22222 *(70.4 – 66.4)/1200 = 89.2 points Another example: Immense`s win on 11th June 2011 carrying 62.5kgs, clocking 85.34secs over 1400m, with p =3.8cms: W = 62.5, Ts = 1082,Ti = 1066, D =1400 Adjusted time = 85.34*(1082/1090)*(1066/1090) = 82.85secs TFR = 62.5*2.22 + 30 – 22222*(82.85 – 78.4)/1400 = 98.2 The reason for selecting these two examples of `SIACHEN` and `IMMENSE` is that in each case the element of uncertainty is minimal as both runners gave their career best performance and ran true to their merit. I would say they may have scope for further improvement. As these two horses ran after a long layoff, I am inclined to give them a slightly higher rating, placing Siachen at 92 points and Immense at 100 points. On the Indian rating system which gives a 26 point higher rating (reasons not known) this would correspond to Siachen at 118 and Immense at 126. As in most cases the uncertainty element is generally quite high, it is advisable to determine the absolute value of TFR for only a select few horses only where this element of uncertainty is minimal in the eyes of the observer. For all other horses the TFR values may be determined by comparative studies only and for this purpose closely fought races only must be considered. Coming back to Deep Impact`s record breaking Performance in the Teno Sho (in 2006) where he clocked 3min 13.4 secs while carrying 58 kgs impost, over the 3200 m course at the Kyoto Race track. I have been trying to evaluate his Time form Rating based on that showing. The Japanese public rates this performance as better than Secretariat’s showing in the Belmont, but the UK experts did not rate him higher than 134 which is curious. In a conversation with some guys on YOU tube in the posting `Deep Impact – Tribute`I had commented that such a time was impossible to believe unless a portion of the track which is covered twice during the 3200m run is having a net downward inclination of substantial quantum in that portion. To my surprise, someone from the JRA itself responded promptly to my comment. His response given below: `@SKSAGAR100 From the JRA - "features a flat home straight of just over 400 meters. The course also is known for its hill laid out over the third and fourth turns, where the track rises from the 1,200-meter mark to the 800 mark, then dips right to the mouth of the stretch. The trip for the Tenno Sho lasts one-and-a-half laps around the oval." The outer oval is 1894m. Go to the JRA site, it has charts with course details (overview, undulation, etc.). How does it effect your figures?` I accessed the site and found the layout and the geometrics. In some portion of the track which is covered twice in the 3200m run there is a rise of 1 in 90 for 360m followed by an extremely sharp fall of 1 in 40 for 160m.This last part could rejuvenate a tiring horse to a good extent. Further interaction with a certain `711ATOM`presumably from JRA on `YouTube` revealed there are three more occasions when a sub 3m 15sec time has been registered over 3200m at the Kyoto Race track. Compare this with the World Record time of 3m 16.4secs clocked by `Kingston Rule`in the Melbourne cup, and the time of 3m 18secs over 2 miles (3218M) by `Petrone` at Hollywood Park in 1969. With all these inputs and again with a great deal of `intuition`- as these inputs are far from sufficient (I have asked for the course record for 1600m over the outer and inner oval tracks which info, if it comes will be quite useful) - I am inclined to give a long term track index of 1122(about 3 % higher than a perfectly fast track) to the Kyoto Race track over 3200m. Accordingly the adjusted time of Deep Impact is worked out as 193.4*1122/1090 = 199.07 secs and the Time form rating of Deep impact is worked out as: 58*2.22 + 30 – 22222(199.07 – 196.7)/3200 = 142.3 This puts Deep Impact in the Elite group of race horses rated 140 or higher. Though he may not figure in the all time top ten list, he still has to be the highest rated horse of the 21st century so far, at least 3 points better than Sea the Stars or Harbinger. But the UK Experts could not rate him higher than 134 which is curious. No doubt he lost in the French Arc, finishing 3rd. But I have no doubt that on his home turf in Japan he would have beaten the best in the world. It appears that the UK Time form rating experts might be a bit biased in favor of European horses. Even some top class American horses are rated lower than their actual worth. What I am surprised at is: Why is there no scientific system in place - acceptable to all countries and turf authorities - that takes care of all the technical parameters, and arrives at the accurate values of the ratings. It’s true there are complexities, but it’s not Quantum Physics. All that is required is an accurate evaluation of the Long term track index ( defined earlier) of the track for various distances. This can easily be done either by experimental observations over a period of time or by a detailed statistical study and analysis of the past performances on the track.