Betfair Official Blog

'Trends, schmends,' says Jack Houghton - A lot of this Epsom 'analysis' is pure bunkum

Click here to view market

Our racing columnist is starting fights with the stats mob as he assesses the value of trends when betting on the horses

Derby imminence provides the opportunity to regurgitate a well-worn lecture, one I've given regularly, about the usefulness (or more usually uselessness) of trend analysis in identifying big-race winners.

It also provides the opportunity to start a fight between two of betting.betfair's venerable contributors. Because it seems every time I read something by Simon Rowlands, he is carping on about what dreadful people trend analysts are: describing them, and their methods, as "crude", "over-valued", "sub-human" and "fascist morons."

Okay, I made the last two up. But read between the lines my friends and it's obvious he considers our dedicated followers of trends with sneering contempt.

Meanwhile, the work of Wayne Bailey, the trendiest of trendsters, gleefully parades itself right next to Rowlands' admonishments.

So who's right?

Well, who cares? My job here is to simply rile them to the point when they agree to settle matters with a street fight. What an event that would be! Betfair could live-stream it; offering a market on the outcome.

Did you know Simon is a vegetarian? Hardly the lifestyle choice of a seasoned brawler is it? But then I read that Wayne is a qualified librarian. A veggie versus a bookworm; it's starting to sound like this match-up has about as much pugilistic fascination as watching an amoeba fight a piece of Blu-Tack. Still, schedule it in the downtime between British and US Racing and I'd bet on it.

However, if they're unwilling to trade physical blows to resolve the trend analysis argument, let them at least appear on betting.betfair with a metaphorical jab or two on the subject. To help them on their way, I'll start the discussion...

Using trend analysis as a method of identifying big-race winners is futile. And those who peddle the analysis in the guise of thoughtful investigation are perpetrators of heinous crimes against rationality. (You see Veggie, I'm on your side).

Examining historical data to predict future events is most commonly used in financial markets; broadly categorised as Technical Analysis. This financial trading technique uses masses of data on how the price of a commodity has fluctuated in the past and draws conclusions therein as to how it will behave in the future.

The technique is widely criticised as having no scientific basis. Peter Lynch, a famous Wall Street investor, once cuttingly summed up the activity: "Charts are great for predicting the past." And if trend analysis is viewed circumspectly in this environment - where it is predicated by oodles of accurate data - how viable can its use be when examining horse races based on such meagre inputs?

The established practice of horseracing trendsters is to look at the previous ten runnings of a race and, based on this minimal input, draw conclusions as to the most likely future winner. Take the Derby. In a national newspaper before last year's renewal, one trendster drew a number of conclusions as to the characteristics the winner would possess. They included such gems as: he would have won last time out; he would have made his debut at a Group One track; he would be in the top six in the market; and, he would have won over at least seven furlongs as a juvenile.

You have to ask, what did that analysis tell us other than last year's Derby would be won by a horse with strong form and a proven capacity for stamina? Did we really need the bogus machinations of trend analysis to tell us this?

In the run up to this year's Derby, I've already heard one pundit inform viewers of: "the strong trend of Dante winners going on to Derby success." Ignoring momentarily the fact the Dante has only provided three of the last ten Derby winners (the Derrinstown Stud Derby Trial has proved an equally good indicator), did the pundit really think he was presenting a useful trend? Because all he really said was that Derby winners tend to have proved themselves in recognised trial races. Hardly groundbreaking analysis, is it?

But perhaps these attacks aren't relevant to the practice of trend analysis itself, but rather to those who do it badly. So come on Wayne and Simon, move the debate on. We're waiting to hear what you think.

In the meantime, consider this. 18 per cent of Derby winners have had names beginning with the letter S. Sindajan isn't looking like a [1000.0] shot now is he?