Post a new comment on this article
Post Your Discussion Comment
Please follow our guidelines for comments and review our competing interests policy. Comments that do not conform to our guidelines will be promptly removed and the user account disabled. The following must be avoided:
- Remarks that could be interpreted as allegations of misconduct
- Unsupported assertions or statements
- Inflammatory or insulting language
Why should this posting be reviewed?
See also Guidelines for Comments and Corrections.
Thank you for taking the time to flag this posting; we review flagged postings on a regular basis.close
Questions and answers
Posted by 02 Feb 2011 at 04:38 GMTon
The authors of the study offer this further commentary:
A critic of our study has approached us recently with a series of concerns about the merits of our work. His comments and those of others appear in uppercase below. We thought it would be helpful to share our responses to these criticisms.
QUOTING FROM THE PAPER: “Whipping does not increase a horse’s chances of finishing first second or third.” THIS MAY OR MAY NOT BE TRUE. HOWEVER, THE STUDY CANNOT CONCLUDE THIS FROM THE LIMITED FINDINGS.
Our response: A crossover study with and without whip use in conditions that replicated racing would be desirable. However, such an experiment was not possible to organise. We therefore started with retrospective analyses of what is actually happening during a sample of races regulated by the Australian Rules of Racing.
QUOTING FROM THE PAPER: “Horses run their fastest when they aren’t being whipped.” THIS CANNOT BE EXTRAPOLATED FROM THE FINDINGS AT ALL. WHAT THE STUDY FOUND IN THIS LIMITED SAMPLE WAS THAT THE HORSES RAN THE FASTEST SECTION BETWEEN THE 600M AND 400M MARK. THAT’S ALL. NOW IF I COULD SHOW YOU RACES (WHICH I CAN) WHERE THE FASTEST SECTIONS OF RACES WERE AT THE END (WHEN HORSES ARE BEING WHIPPED), WOULD I BE ABLE TO CONCLUDE THAT WHIP USE HAS AIDED THESE HORSES? OF COURSE NOT.
Our response: In the five races studied, with velocity measured as time for 200 metres, the horses ran fastest in the 600 to 400m, not the 400 to 200m, or 200m to finish. There was only one section when whipping was not present, and it coincided with the fastest of the last three 200-metre sections. There is no cause and effect argued, and the limitations of measuring velocity as time for 200 metres are presented in the Discussion. Likewise, the Discussion notes that different race tactics might produce different results.
QUOTING FROM THE PAPER: “The strongest predictor of racing success is how a horse runs in the section of the race where he isn’t being whipped.” THIS IS JUST LAUGHABLE – THERE IS NO SINGLE PREDICTOR OF RACING SUCCESS IN THE TERMS MENTIONED HERE. LET ME LIST JUST A FEW OF THE VARIABLES TO BE TAKEN INTO ACCOUNT:
**RACE PACE – THIS RELATES TO THE GENERAL PACE – COULD BE SLOWLY RUN OR A FAST EVEN PACE, ETC, ETC, ETC.
**LENGTH OF THE RACE. YOU PICKED A PARTICULAR TRACK AND RACE DISTANCE.
**RACING PATTERNS – DIFFERENT HORSES HAVE DIFFERENT RACING PATTERNS – THEY MAY BE LEADERS – THEY MAY RACE ON THE PACE – OR THEY MAY BE SWOOPERS.
**WEATHER – YOU ISOLATED ONE PARTICULAR WEATHER PATTERN WITHOUT ANY THOUGHT OF WHAT EFFECT THIS MIGHT HAVE ON THE STUDY.
** TRACK VARIATIONS (e.g., SIZE – CIRCUMFERENCE OF BENDS).
Our response: All of the variables listed above might influence results, and that is also noted in the Discussion. In a recent radio interview, a prominent jockey argued that performance was most reliant on how the horse performed as it galloped to the 400 and 200m position. The results do not conflict with that view.
IF WE ARE TRYING TO ASCERTAIN THE EFFECT OF WHIP USE, THEN WE NEED TO RUN TWO PARALLEL RACES (ALMOST IMPOSSIBLE AS I FOUND OUT) – ONE WITH NO WHIP USE AT ALL AND ONE WITH WHIP USE – YOU WOULD THEN NEED TO MEASURE THE RELATIVE SPEED IN BOTH EXAMPLES TO SEE WHETHER ONE WAS FASTER THAN THE OTHER.
Our response: We agree, but find little value in arguing for a controlled study that is impossible to conduct. It is almost impossible to conduct an experiment on this issue, especially if we attempt to replicate all the conditions found in a normal race, with Stewards observing jockey behaviour, and betting being conducted. What this study offers instead is an analysis of actual races and therefore a reflection of real conditions. It would be difficult to have a blind study because jockeys would know what was being measured and could alter their behaviour. If more research were conducted, it would be unlikely to be in the form of an experiment. Further research would involve observing more races and more horses, and including more factors.
FROM THE WAY THE STUDY HAS BEEN DONE, IT APPEARS THAT NEITHER RESEARCHER HAS ANY EXPERIENCE IN STUDYING RACES. IF YOU ARE GOING TO STUDY SOMETHING AS COMPLEX AS RACING, THEN SURELY YOU WOULD HAVE NEEDED A SIGNIFICANT AMOUNT OF ADVICE FROM PEOPLE WHO UNDERSTAND HOW RACES ARE RUN. JUST TO SHOW YOU THE VARIATIONS IN RACE SPEEDS, I’LL MENTION SOME GENERAL TRENDS IN HARNESS RACING. IN THE US, THESE RACES ARE NEARLY ALL RUN AT MAXIMUM SPEEDS (ALL OVER THE MILE DISTANCE) AND THEY TEND TO BE SIMILAR TO THE FIVE RACES YOU MENTIONED IN THAT THE LAST QUARTER (400M) CAN OFTEN BE THE SLOWEST. VERY OFTEN THE FIRST 400M CAN BE THE FASTEST. IN AUSTRALIA, VERY OFTEN THE LAST SECTION WILL BE THE FASTEST BECAUSE AUSTRALIAN RACE PATTERNS ARE DIFFERENT FROM THOSE IN THE US. YOU MAY NOW SEE THE RELEVANCE OF MY EARLIER COMMENT RE WHIPS – THERE ARE SOME RACES WHERE WHIPS WILL BE USED DURING THE FASTEST SECTION OF THE RACE. THE ONLY COMMON DENOMINATOR IS THAT, IN ALL COUNTRIES, THE WHIPS ARE USED DURING THE LAST 400M (APPROXIMATELY). IF YOU HAD SPOKEN TO RACING ANALYSTS REGARDING RACING PECULIARITIES, THEN YOU WOULD HAVE HAD SOME KNOWLEDGE ABOUT THE ACTUAL VARIABLES, AS WELL AS THE KNOWLEDGE THAT RACES CAN VARY SIGNIFICANTLY DEPENDING ON THOSE VARIABLES. THIS WOULD HAVE ALERTED YOU TO THE FOLLY OF USING FIVE RACES OVER A SET DISTANCE ON ONE PARTICULAR TRACK, WHEN RACES ARE RUN FROM 800M–3200M IN THOROUGHBREDS (MAINLY 1600M–3200M IN HARNESS RACING). THE RESTRICTION IN THE SAMPLE SIZE AND THE FACT THAT THE VARIATIONS IN DISTANCE WERE NOT TAKEN INTO ACCOUNT WOULD INDICATE THAT IT IS IMPOSSIBLE TO MAKE ANY WORTHWHILE FINDINGS AT ALL.
Our response: The limitations inherent in extrapolating to other distances, racetracks, ground conditions, etc, are acknowledged in the Discussion.
QUOTING FROM THE PAPER: “We were also interested to determine whether performance in the latter stages of a race was associated with performance in the earlier sections of the race.”
YOU INDICATE HERE THAT YOU CAN DO THIS BY MEASURING WHIP STRIKES AND THEN TRY TO ASSOCIATE VELOCITY WITH WHIPPING AND NON-WHIPPING. THIS IS JUST RIDICULOUS. THE RACES YOU MEASURED ARE SPRINT RACES AND IT IS TRUE IN THESE RACES THAT POSITIONING IS IMPORTANT SO IF A HORSE NEEDS TO MOVE TO A MORE FORWARD POSITION THEN IT WILL ACCELERATE. IN SPRINT RACES THE GENERAL TREND IS THAT THE HORSE NEEDS TO BE ON THE FRONT END – THERE ARE STATISTICS AVAILABLE THAT INDICATE THE NUMBER OF LEADING HORSES THAT WIN AND THE RELATIONSHIP BETWEEN RACE TEMPO AND HORSE POSITION DURING THE RACE. THE VELOCITY (SPEED) OF HORSES IN THE LAST SECTION OF THE RACE IS RELATED TO EARLIER SECTIONS OF THE RACE BUT THAT HAS VERY LITTLE TO DO WITH WHIP USE DIRECTLY. THE SPEED OF THE LAST SECTION OF THE RACE IS DETERMINED BY THE HORSE’S EXPENDED ENERGY DURING THE RACE. FOR EXAMPLE, IF IT IS A FAST RUN RACE (IF EARLY SECTIONS ARE QUICK), THEN IT IS LIKELY THAT THE LAST SECTION WILL BE RELATIVELY SLOW. IF THE EARLY SECTIONS ARE SLOW THEN IT IS LIKELY THAT THE FINAL 400M WILL BE VERY QUICK. TRACKTYPE – LENGTH OF STRAIGHT – LENGTH OF THE RACE – FIELD SIZE – RACE TEMPO – WEATHER – FIELD POSITION WILL ALL INFLUENCE THE SPEED OF THE LAST SECTION (WHETHER YOU ARE MEASURING IN 200M OR 400M SECTIONS).
Our response: We have properly addressed the limitations of the study. As for race tactics, it seems that in the five races studied the average tactic was to get the horses to the 400 and 200m positions as quickly as possible, then strike them 6-7 times, on average, during fatigue. That might not happen in all horses, or in all 1200¬1250m races. Again we concede the limits to the general applicability of these findings in the Discussion.
THERE IS GREAT DETAIL AND USE OF STATISTICAL ANALYSIS, WHICH YOU INDICATE LEADS YOU TO CERTAIN CONCLUSIONS. HOWEVER, THE WHOLE PREMISE OF THE STUDY IS WRONG. YOU HAVE NOT TAKEN ANY ACCOUNT OF OTHER “RACING VARIABLES” MENTIONED ABOVE SO THE STATISTICAL ANALYSIS BECOMES IRRELEVANT. TO BE FAIR, SOME LIMITATIONS OF THE STUDY ARE DETAILED IN THE RESEARCH PAPER, AND IT ACCEPTS THAT THE STUDY DOES NOT FIND A CAUSAL LINK, AND THE STUDY DOES ACKNOWLEDGE THAT FURTHER STUDIES ARE NEEDED.
Our response: We have noted in the Discussion the limitations in the study, with particular reference to other circumstances that might influence the results.
THE FINAL PARAGRAPH, HOWEVER, PROVIDES A WONDERFULLY SUBJECTIVE STATEMENT THAT BEARS NO RELATIONSHIP TO THE STUDY, BUT PROVIDES A WONDERFUL QUOTE FOR THE FUNDER, THE RSPCA, TO QUOTE THE STUDY TO JUSTIFY ITS IDEOLOGICAL POSITION. IMHO THAT IS DISGRACEFUL AND I HAVE NOTIFIED THEM OF MY CONCERNS. IN THE VIDEO “INTERVIEW” THE IDEOLOGICAL POSITION IS PUT AND THE DISCUSSION PROVIDES WONDERFUL “SPIN” FOR THE RSPCA. IT DOES RESPECTABLE RESEARCH NO FAVOURS. IDEOLOGICAL POSITIONS ARE QUITE OK IN THE RIGHT PLACE. HOWEVER, THIS IS A RESEARCH PAPER NOT A POLITICAL PAPER. IF THE ETHICAL POSITION IS THAT THERE IS NO PLACE FOR WHIPS IN ALL HORSE SPORTS, THEN THAT CAN BE STATED. IF THAT IS THE CASE THEN THERE IS NO NEED FOR THE RESEARCH – IT WOULD BE SUPERFLUOUS. WE CAN JUST ALL HOLD OUR RELATIVE IDEOLOGICAL POSITIONS, BUT THAT IS NOT A GOOD SCENARIO.
Our response: Ideology was all that was available to "inform" debate before the paper was published. There is now evidence to add to the debate, even if that evidence has its limitations because it is not a controlled trial, and it does not cover all possibilities. We look forward to more research into aspects of whip use and racing outcomes. Science is a process of development and retesting of hypotheses.
HAVING TAKEN ON THIS RESEARCH, AND BEEN AWARE OF THE IDEOLOGICAL POSITION OF THE FUNDER, THE RSPCA, I WOULD HAVE THOUGHT THAT THE RESEARCHERS WOULD HAVE BEEN VERY CAREFUL TO ENSURE THE INTEGRITY OF THE RESEARCH, AND THAT IN ANY INTERVIEW THE FINDINGS AND LIMITATIONS WOULD BE OUTLINED. IMHO THAT IS SADLY LACKING. THAT MAY BE THE FAULT OF THE RSPCA – I’M NOT SURE – BUT THE VIDEO INTERVIEW HAS PUT THE RESEARCHERS IN AN UNFAVOURABLE LIGHT.
Our response: The RSPCA did not design the study, and did not choose which races or racetrack to study. The authors likewise did not choose the racetrack or races, or do the whip counts. Employees of Racing NSW conducted the above elements of the study. The statistics were not assembled by the authors. They were independently worked by Dr Michael O'Neill, formerly Associate Professor, Biometry, University of Sydney. We refer the critic to his website: www.stats.net.au.
I HAVE READ ALL OF THE COMMENTS MADE IN THE VARIOUS NEWSPAPER ARTICLES AND IT CONCERNS ME. EVEN PEOPLE WITH A NON-RESEARCH BACKGROUND HAVE PICKED THE “LIMITATIONS” TO THE STUDY. THIS IS MOST UNFORTUNATE BECAUSE I CAN SEE SEVERAL NEGATIVE RESULTS.
Our response: The limitations are outlined in the paper. We note that some of the comments in the online responses to media reports also point this out to those critics who have got themselves into a lather about the limitations of the study.
THERE WERE CORRELATIONS BETWEEN WHIP USE AND RACE PLACINGS. HORSES THAT WERE WHIPPED MORE HAD SUPERIOR PLACINGS, SO SURELY WHIP USE AIDED PERFORMANCE?
Our response: This relationship was expected because horses better placed at the final 400m and 200m marks were whipped more frequently from that point onwards. Even without whipping, these horses had an increased chance of going on to achieve a place.
DO YOU MEAN TO SAY THAT HOW A HORSE RUNS OVER THE LAST 400 AND 200M HAS NO EFFECT ON ITS RACE PERFORMANCE?
Our response: On average in this study, variation in those speeds was not a significant predictor of superior race placing, expressed as finishing 1st, 2nd or 3rd. That must mean that variation between horse speeds in the first 850–1050m of the races was a more important predictor. Those variations in the first 850–1050m were achieved without any use of whipping (no horse was whipped before the final 400m).
HOW DO WE KNOW THAT WHIPPING ISN’T STOPPING HORSES FROM SLOWING DOWN? HOW DO WE KNOW THAT WHIPPING ISN’T ENSURING THAT HORSES USE EVERY LAST BIT OF FUEL IN THE TANK?
Our response: Horses were slowing down. The point is that how a horse performs in that last 400m (when jockeys start whipping it) is NOT the critical factor in whether or not it’s going to achieve a place. The horse’s performance up until that point is what counts (and in this study no horse was whipped up to that point).
PERHAPS SPEED VARIED OVER THE LAST 100M SO AS TO AFFECT THE LIKELIHOOD OF BEING PLACED 1ST, 2ND OR 3RD?
Our response: Yes, and more frequent measurements of speed would be needed to measure that effect. However, the lack of an effect of whipping on variation of 200m sectional times seems sufficient argument to ban whipping. Many more races would need to be studied to investigate whether whipping has an effect on the probability of finishing in first place.
WHAT DO THESE RESULTS MEAN FOR THE ONGOING USE OF THE WHIP IN RACING?
Our response: The results of this study do not offer any support for retaining whipping in horse racing. This is the first study to confirm that whipping does not increase the chances of a horse finishing 1st, 2nd or 3rd. Horses reach their highest speeds in a race when they aren’t being whipped.
IF ALL JOCKEYS IN THE STUDY WERE MEETING THE WHIP RULES, HOW IS IT THAT 98% OF HORSES WERE WHIPPED? SURELY THEY COULDN’T ALL HAVE BEEN “IN CONTENTION”?
Our response: Whether or not a horse is “in contention” is subjective and not easily defined. This, more than anything else, highlights the difficulty that stewards face in policing the rules of racing.
WHY WOULD JOCKEYS WHIP HORSES MORE IF IT HAS NO EFFECT?
Our response: For any number of reasons.
Punters want to be sure that the horse they’ve backed is “ridden out” and whipping is certainly a powerful visual to support that.
A whipped horse may change its stride giving the perception that it’s going faster.
There may be transitory effects on speed or locomotion (e.g., over 30–40m) that are perceived by jockeys as contributing to improved performance.
SOME INDIVIDUAL HORSES MIGHT HAVE HAD A RESPONSE TO THE WHIP THAT INFLUENCED ITS PERFORMANCE.
Our response: Yes, and that effect might have been positive or negative. It is not appropriate to retain whipping in all horses just in case an individual horse might improve its performance. Science and evidence is about averages and statistical significance.
If there is a perceptible increase in performance with whipping, it is surprising that the horses in inferior placings at the 400m and 200m were then whipped less than the leaders from that point onwards. Surely one would expect horses in inferior placings to be whipped more if it does cause meaningful acceleration?
NO STUDY HAS EVER PROVED CONCLUSIVELY THAT WHIPS CAUSE PAIN.
Our response: If whips didn’t cause some pain or discomfort, then they wouldn’t be used. If whipping isn’t associated with racing outcome, then isn’t it best to give horses the benefit of the doubt in terms of pain and not use them?
WHAT WAS THE AIM OF THE STUDY?
Our response: To investigate whether whip use improved racehorse performance. The study looked at how often horses were whipped and what influence, if any, whipping had on race outcome, expressed as finishing 1st, 2nd or 3rd.
WHAT DID YOU EXPECT TO FIND?
Our response: We hypothesised that whippings would be associated with superior performance and that those superior performances would be explained by an effect of whipping on horse velocities in the final 400m of the race. What we found was that velocity in the final 400m was not a critical factor in racing success.
WHAT DID YOU MEASURE AND HOW?
Our response: The authors analysed race data provided by Racing NSW. They looked at sectional times for the last three 200m sections: 600m–400m, 400m–200m, and 200m–finish. Total whip counts were made for each eligible horse in each section. Placings of each horse were also described at 400m, 200m and at the finish. They also looked at where horses were positioned at the end of each section and at the finish.
Times and rankings for these sections were derived from an electronic timing system that sits underneath a horse’s saddle cloth and communicates with transmitters in the ground. Race outcomes were described as final placings and whether or not a horse finished in the first three placings (that is, placed 1st, 2nd or 3rd, in any order).
Potential predictors of a horse’s final placing were considered, including: the number of starters; race distance; horse placing at the 400m and 200m marks; sectional times and total number of whip strikes between 400m–200m and 200m–finish.
The study found that in the last 600m of the race, horses on average ran fastest in the section that they weren’t whipped (600m–400m). Not only did horses decrease their speed from the time of whipping in the 400m–200m section, but their speed continued to decrease as whipping increased in the 200m to finish section.
Horses in more advanced placings at the 400m and 200m positions were whipped more frequently from that point onwards than the lower-ranked horses at those positions.
The strongest predictor of racing success was where a horse was placed at the 400m and 200m marks. Neither whip counts in the final 400m or 200m, nor velocity in the final 200m, significantly explained the probability of finishing 1st, 2nd or 3rd.
WHO FUNDED THE STUDY?
Our response: The study was funded by RSPCA Australia. Racing NSW provided in-kind support.
THE STUDY ONLY LOOKED AT ONE RACECOURSE AND FIVE RACES – HOW DO YOU KNOW THE RESULTS ARE RELEVANT TO OTHER RACETRACKS AND RACES?
Our response: The sample size is statistically valid and it is reasonable to expect to find the same outcome in other races of the same distance across the country.
THIS PAPER DOES NOT DESCRIBE AN EXPERIMENT. IT IS JUST AN ANALYSIS OF OBSERVATIONS. WHY ARE THE RESULTS OF ANY INTEREST?
Our response: It is almost impossible to conduct an experiment on this issue. What this study offers instead is an analysis of actual races and therefore a reflection of real conditions. It would be difficult to have a blind study because jockeys would know what was being measured and could alter their behaviour. If more research were conducted it would be unlikely to be in the form of an experiment. Further research would involve observing more races and more horses, and including more factors.
WHY LIMIT THE ANALYSES TO FIVE RACES? ONE RACETRACK? THOSE DISTANCES? A NARROW RANGE OF TRACK RATINGS?
Our response: To reduce "background noise" in the experiment – that is, variance due to those factors that could reduce the likelihood of demonstrating the effects of whipping if they did exist.
HOW ACCURATE ARE THOSE TIMES FOR 200M?
Our response: That was not checked. They are a standard measurement in the industry. Even if they were slightly inaccurate, that inaccuracy would affect all horses equally.
MAYBE THE STEWARDS HAD A VESTED INTEREST IN THE RESULTS, AND WERE BIASED IN THEIR COUNTS OF WHIP STRIKES?
Our response: Two stewards collaborated in the counts. The stewards did not know how the results would be analysed. Stewards count whips routinely during regular administration of the rules, and were asked to use the same approach in this study.
MAYBE THE OMISSION OF "SLAPS" COMPROMISED THE STUDY?
Our response: Perhaps, but slaps are not regulated by ARB Rules.
A PROMINENT TRAINER HAS ARGUED THAT IT’S MAINLY THE BEATEN HORSES THAT ARE WHIPPED MOST.
Our response: Incorrect. Horses that were at a superior position 400m from the finish without being whipped, were whipped more in the ensuing 400m.
IS THIS PAPER THE LAST WORD ON THE ISSUE, OR IS MORE RESEARCH NEEDED?
Our response: The results support the banning of whip use as described in this study. More research could look at different tracks, distances, prize money and other factors, and could investigate locomotory responses to whip use over short distances (e.g., 30–40m) to investigate the influence of whip strikes on the probability of finishing first.
However, it is clear that under current ARB rules it is highly likely that there will be many horses whipped every day in Australia while fatigued and that this will have no measurable impact on their racing success. Whipping is NOT required to produce peak speeds during races.