An End To Blinkered Thinking?

The relationship between racehorses and equipment - especially headgear - is often misunderstood. That can be mainly credited to two factors. Firstly, some equipment, such as different tack (sheepskin or cross nosebands, for example), is not required to be declared. And secondly, the general public have nothing more than a notation in the racecard relating to the wearing of other equipment.

Some publications do make a point of highlighting that a horse is, for instance, blinkered first time. But is the initial fitting of blinkers a positive? If not, what about subsequent runs?

The Tongue Tie: A Precursor to the Blinkers Study

I did some research a while back on the fitting of tongue ties and, specifically, whether a first time tongue tie was advantageous. The findings - reproduced below - were surprising, to me at least.

 

 

The first application of a tongue tie, shown here in the top row, A) 0, has the lowest win strike rate of all. It also has the lowest place strike rate, and the highest negative ROI.

What was really interesting was the linear progression in performance thereafter. The second application of tongue tie - B) 1 - was better than the first, but not as good as the third; the third was better than the second but not as good as the fourth; and so on.

These observations were made on what, in racing terms, is a vast sample size of 115,000 runners, and it is clear that, in general, it takes time for a horse to learn to breathe with its tongue tied down. There is also a point - somewhere between the 10th and 25th application - beyond which the tongue tie's effectiveness wanes.

**

"Wait, I can't see much! What's that behind me?!"

So why put blinkers on a horse anyway? I was re-reading Nick Mordin's brilliant Betting For A Living for the nth time recently, and stumbled upon his version of the origination of racehorse blinkers. It goes thus:

Reproduced from Nick Mordin's Betting For A Living

Reproduced from Nick Mordin's Betting For A Living

 

More recently, US turf writer, Marcus Hersh penned this 'qualitative' article, quoting many top American trainers as well as John Gosden, who began his training career in California.

Both Mordin and Hersh make the point that blinkers replicate the 'fight or flight' (in this case, 'flight' being the hoped for response) instinct in a domesticated herd animal.

Before a horse sports blinkers, it is almost always assumed not to need them and is given a chance to run unencumbered by headgear. For many, though, this is a temporary hiatus before some sort of garb is reached for, generally in search of hitherto undiscovered improvement.

Variations on the blinker theme exist in the form of eye shields, visors, and cheek pieces, as well as actual differences in the 'cup' on a pair of blinkers. The research below focuses solely on blinkers in all their forms (British racing does not require trainers to declare, for example, half cup blinkers, or full cup blinkers), and uses 'no headgear' as a control.

'No headgear' means the horse did not wear blinkers, or an eye shield, or a visor, or cheek pieces, or indeed anything. By extension, then, the study does not include all runners: those sporting equipment other than blinkers, or in combination with blinkers are excluded from this research.

I have included short form comparisons with equivalent data from Irish racing.

UK Blinkers vs No Headgear, 2014 to 2016

The first dataset I looked at was 'all UK racing' for the three year period 2014 to 2016.

During that period, the average win strike rate for a blinkered horse was 10.43% compared with an average for horses wearing no headgear of 11.76%. It should surprise nobody that horses wearing blinkers typically under-performed in strike rate terms compared to horses without headgear of any sort.

But I wanted to dig a little deeper. My experience with the tongue tie dataset had led me to wonder whether there were any similar patterns for blinkered horses: did familiarity with the equipment improve performance?

As you can see from the table below, there was an improvement in the win strike rate for blinkered horses through their early runs in blinkers.

 

 

Horses having their first run in blinkers - i.e. with zero previous blinkered runs (Prev Blkrd Runs) - won just 9.53% of the time, compared to 11.76% for horses not wearing headgear.

But, on their second run in blinkers (i.e. one previous blinkered run), that improved to 10.96% winners. Whilst still lagging some way behind the control this represented a significant step forward, and can reasonably be put down to the benefit of experience.

Now here's the interesting thing: horses having their third start with blinkers won at a rate of 12.74%, markedly better than runners without headgear. Moreover, their place strike rate was higher and their ROI was much closer to zero than other group. The second best ROI group was 'first time blinkers'.

For some reason, the sixth run in blinkers was also very close to the 'no headgear' group. That was odd and my first reaction was that it is happenstance.

 

Blinkers vs No Headgear, 2012 to 2013

That initial look piqued my interest for a deeper dive, so I went back another couple of years to see if the fairly strong patterns were replicated on a separate sample.

Here, the control - i.e. no headgear - group won at a rate of 11.15%. Again the blinkered first time subset performed poorly, scoring just 9.56% of the time albeit for a very respectable -5.6% ROI at starting price.

Second time blinkers showed an improvement on the initial effort in win and place strike rate terms, and third time blinkers was another step forward. Indeed, as with the first dataset, third time blinkered outgunned the 'no headgear' group by a statistically significant margin: 11.83% vs 11.15% is an improvement of more than 6%.

Note as well that both fifth and sixth time blinkered also fared better than the control sample, especially sixth time... again. Weird.

 

 

There seemed to be quite strong similarities between the two UK datasets, so I extended my analysis to Irish racing for the same period. The results were again similar, though this time no blinkered sub-group beat the control:

 

The group that came closest to matching the control was - you guessed it - the third time blinkered set. Quite why the 'no headgear' set had the edge this time, I'm not sure, though I suspect it may be something to do with the might of the Mullins and O'Brien yards which dominate Irish racing and rarely use blinkers (just 300 blinkered runs collectively versus 5122 'no headgear' runs, or 5.5%).

 

Blinkers vs No Headgear, Handicaps (All Codes), 2012-2016

Having looked at the five year supersets for all races, I then honed in on handicaps. Below is the UK story for number of blinkered runs in handicaps under all codes (i.e. flat, all weather and National Hunt).

Here again, perhaps unsurprisingly given the related contingency nature of the subset, that third run in blinkers shows a marked improvement over 'no headgear'. This time it is 12.29% vs 11.18%, a 10% advantage; and both place strike rate and ROI are notably better than those without facial accoutrements.

The best performing group in ROI terms was 'first time blinkers', again.

And again we see that odd sixth time in blinkers group outperforming the rest. Weeeirrd...

 

 

The Irish comparison was interesting. When looking at all races above, I surmised that a possible reason for the 'third time in blinkers' group performing less well than the 'no headgear' group was the preeminence of the Mullins and O'Brien stables. If there was any truth to my conjecture, we might expect to see the handicapping third time blinkers group fare better: after all, most of those super power batallions race in open company (80% in the survey period vs 20% handicaps).

Your first 30 days for just £1

Sure enough, we do see the third time blinkered Irish handicappers outperform the 'no headgear' handicappers, and by a sizeable margin: the same 10% as the UK handicappers.

 

Blinkers vs No Headgear, Handicaps (Flat Turf & All Weather), 2012-2016

Finally, I wanted to split out flat (turf and all weather) handicaps from National Hunt handicaps. You'll be familiar with the view by now...

 

 

The same themes emerge yet again: third time in blinkers significantly outpoint UK flat handicappers without headgear - by 9.34% this time (11.94% vs 10.92%). The best ROI performance comes from those blinkered first time once more. And those sixth time blinkered UK 'cappers almost match the 'no headgear' mob. Quirky...

 

Bringing in the Irish scene, we see more of the same:

 

Third time blinkers had a nigh on 10% strike rate, compared with 8.74% for those wearing no headgear. That 13.84% uplift is the largest of all the disparities in my research, but I can't logically apportion it to anything other than variance. Nevertheless, the core principle is alive and kicking: third time lucky for many in blinkers.

 

Blinkers vs No Headgear, Handicaps (National Hunt), 2012-2016

The last subset under scrutiny was National Hunt handicaps. And it was the group for which blinkers seemed to be the most consistently advantageous, at least when compared to not wearing headgear of any description.

Here we see that, after that initial blinkered start - perennially an under-performing run in strike rate terms, regardless of dataset - those wearing 'the blinds' outran those unfurnished for almost all of their next five runs (excepting fifth time blinkered - go figure).

Top of the pile, on place strike rate at least, is the third run in blinkers brigade. They were actually marginally usurped in win strike terms by the sixth run crew. Again, I have no idea why that might be.

From an ROI perspective blinkered first timers went closest to breaking even, as they have done throughout my analysis.

 

 

The Irish National Hunt handicappers perspective has a familiar look to it:

 

Third, and especially sixth, time blinkered horses were in front of those without headgear. Again.

 

Conclusions: Winners? Or Profit?

This was a fascinating, and to some degree, surprising study. When I set out, I expected to see some improvement from run to run as per the tongue tie research I'd previously performed. What I didn't really expect was such a strong correlation in the data, regardless of race type, time frame or racing jurisdiction.

That slicing and dicing of the dataset adds a layer of credibility which is quite rare in horse racing research. Normally, samples are too small or conclusions too obvious. But here we can clearly see something which I've not seen expressed anywhere before:

Horses wearing blinkers for the third time, on average, win significantly more often than horses wearing no headgear.

Horses wearing blinkers for the third time, on average, win significantly more often than horses wearing no headgear.

Whilst that is worth knowing, and can be the starting point for making a bet, or at least marking up the chance of a horse with a borderline value case, it is not profitable in or of itself.

Indeed, it is interesting that, while first time blinkered horses perform comparatively very poorly in strike rate terms, their starting price return on investment is closer to zero than those without headgear across almost all samples above. That implies they are sent off at bigger odds than ought to be the case.

Remarkably, backing all UK blinkered first time horses from 2012 to 2016 would have yielded a positive ROI of 10.66% at Betfair SP.

Remarkably, backing all UK blinkered first time horses from 2012 to 2016 would have yielded a positive ROI of 10.66% at Betfair SP.

That's 675 points on 6338 bets. Granted, the results are skewed by Morning Post's 100/1 win in 2013, which returned 470 on Betfair!

Looking only at handicaps - Morning Post ran in a conditions race - still gives an ROI at BSP of 9.28%. That's a profit of 440 points from 4743 bets where no winner returned bigger than 75 on the exchange.

Irish handicaps also returned a profit at BSP, though it was as a result of three triple-digit payoffs.

At the end of the day, then, it comes down to the age old punters' dilemma: do you want winners, or profit? First time blinkers will give you less winners than any other group in the sample. But... they are your best chance of turning a profit.

Given the size of these datasets and that there are brain dead angles contained within, the scope to develop profitable micro-systems is vast.

 

What Next?

This research is merely the tip of the iceberg. I haven't looked at trainers who fare especially well (try Gary Moore and Tim Easterby first time blinkered in handicaps); or compared with last time out headgear (blinkers off to blinkers on, switch from another headgear type); or looked at how exposed horses were prior to the fitting of blinkers; or looked at the donning of blinkers in conjunction with other equipment; or indeed looked at the general effect of other types of headgear.

In other words, there are many routes to explore from here. Each may be a profit-sucking cul de sac or an expansive value boulevard. Go forth, and let this be an end to your blinkered thinking... or perhaps just the beginning!

To help your exploration, we've now added numerical suffixes to horses wearing equipment, as follows:

Geegeez racecards now have a count on the number of times equipment combinations have been deployed on a horse

Geegeez racecards now have a count on the number of times equipment combinations have been deployed on a horse

In this example, we can see that Piazon is wearing the combination of hood, tongue tie and cheekpieces for the first time (h,t,cp1). Meanwhile Artscape and Taajub (on his 77th career start!) are wearing cheekpieces and blinkers respectively for the first time, Coastal Cyclone has a second run in blinkers, while Mad Endeavour has run at least five times in blinkers.

We know from the above research that first time blinkers have a low strike rate but a profitable bottom line, such is their 'marmite' impact. And we know that 'bl3' may be something to look out for from a win strike rate perspective, notwithstanding that value judgements will have to be made about the animals deploying such kit.

To assist with such judgements, clicking the 'TODAY'S HEADGEAR' check box on the Full Form tab will shine a light. Here's Mad Endeavour's record with blinkers:

Select TODAY'S HEADGEAR to see a horse's record with the same equipment

Select TODAY'S HEADGEAR to see a horse's record with the same equipment

Matt

FOOTNOTE: I was reminded that there is an inherent 'sampling bias' in the progressive number of blinkered runs. That is, a horse which has a positive experience in blinkers is more likely to retain them for a subsequent run/runs; conversely, a horse with a negative experience is more likely not to run in them again.

Whilst that is undoubtedly true, it does nothing to invalidate the findings which are, as stated, intended as a) food for thought and b) a starting point for the curious among you to undertake further research. Moreover, the only element of 'directly profitable takeaway' concerns first time blinkers, for which there is no such sampling bias.

Horse's racing careers are necessarily bound by sampling biases in that their trainers and owners will want to replicate successful outcomes and avoid repeating failed ones.

Your first 30 days for just £1
6 replies
  1. Blokeshead says:

    “Happenstance” Matt? I thought you were a Londoner these days – when did you become a Sherman? ;o)

    More seriously, thanks for a fine article. Must have taken you a good while to pen that little lot – much appreciated.

    Reply
    • Matt Bisogno says:

      Haha, thanks Stuart. It did take a while. I’m limbering up for the book! (Actually, this is a small section of the ‘equipment’ chapter…)

      Hope all’s well in Sverige,
      Matt

      Reply
  2. paul mason says:

    Hi a great report well worth the read, another trend that i was interested in was horses that have been gelded how they do on there 1st or 2nd runs. From initial viewing it looks ok.

    Reply
  3. Richard Noy says:

    So how about Geegeez racecards showing “b1”, “b2” and “b3” for the first THREE times that blinkers are used? I’m guessing that the answer is that it’ll slow down the software which produces the data too much because of the need to interrogate past performance not once but twice.

    Reply
    • Matt Bisogno says:

      No, Richard, the answer is that we’ll have it online in a fortnight or so, with a following wind. Not just for blinkers but all headgear; and the notation will be b1 to b4 and then b5+

      Glad you asked!

      Best,
      Matt

      Reply

Leave a Reply

Want to join the discussion?
Feel free to contribute!

Leave a Reply

Your email address will not be published. Required fields are marked *