The market as a guide

SYSTEMS: 'When a horse's price shortens from its forecast odds it is generally not a good bet. All the value has gone'

I HAVE frequently criticised both the quantity and quality of information that is offered to British punters. But one area where I have no complaints is the betting forecast.

It is easy to take for granted that the betting forecast in your racing daily is a very good predictor of what the betting will actually be. But this is not the case elsewhere.

In all other racing countries that I have knowledge of, the betting forecast, where it is provided, is hopelessly inaccurate.

I think the staff on our racing dailies do an extraordinary job of getting inside the heads of punters and reading how they will bet on the various runners.

They have a tremendous gut feeling about how most bettors will respond to a particular type of horse in a particular type of race.

This being so, I have long wanted to carry out a survey to see what happens when a horse's starting price differs from its forecast price. The sheer volume of research involved has prevented me previously.

But now that the Racing System Builder software has been made available (see Weekender Systems page, June 24), I can finally tackle the task.

The Racing System Builder software records the betting forecast and starting price for every Flat race run in Britain over the last nine years. So I set it to work on the last two seasons results to see what it could turn up.

To start with I looked at the relationship between the forecast price and the starting price:

Win % Loss %

F/c less than 40% of SP 0.6 77.4

F/c 40% to 59% of SP 2.8 58.7

F/c 60% to 79% of SP 5.7 40.4

F/c 80% to 99% of SP 8.4 29.9

F/c 100% to 119% of SP 11.4 19.6

F/c 120% to 139% of SP 15.2 17.8

F/c 140% to 179% of SP 18.6 18.8

F/c 180% to 259% of SP 20.9 25.0

F/c 260% to 419% of SP 24.7 22.5

F/c 420% + of SP 33.6 3.4

What these figures show is basically that the shorter a horse's starting price is in relation to its forecast price, the more likely it is to win.

If on the other hand, it drifts out from its forecast odds, it is less and less likely to win.

At the same time, the more the horse is bet, the less money you lose, and the more it drifts, the more money you lose.

This sounds like great news at first. It seems like the way to win (or at least lose less) is to simply follow the money.

Alas, life is not that simple. In reality, the preceding figures have a built-in deflationary factor that is produced by what the researchers call the "fav- ourite-longshot bias."

The favourite-long-shot bias is a well catalogued phenomenon that affects all betting markets.

Essentially, it means that punters tend to bet too much on longshots and too little on favourites.

The result of this is that you lose a lot less money if you bet favourites rather than long-shots. Basically, the bigger the price, the more, on average, you will lose.

It follows that since the various categories in my survey contain horses whose average starting price is progressively lower, they will show a smaller and smaller loss.

In other words, the results are pretty much what you would expect by picking groups of horses randomly at progressively shorter prices.

To prove this, I looked at horses in a narrow starting price range (less than 2-1) and divided them into two groups - those whose forecast price was longer, and those whose forecast price was shorter.

It turns out that the short-priced horses whose SPs were less than their forecast prices actually lost punters more than those who drifted up from their forecast odds (7.1% versus 2.2%)

This makes sense to me. After all, if the betting forecast expert thinks a horse should be 5-4, it probably represents value at 7-4. But it may well be a poor value bet if it is cut to odds on.

I wasn't content to leave the situation there, however. Even if overall a reduction in betting odds means little, I reasoned that there must be some circumstances in which it is significant.

I tried various angles to uncover such situations and the nearest I got initially was when I divided the results into the various official classes of racecourses.

It turns out that the higher the class of track, the less significant market moves appear to be.

On Grade One tracks such as Ascot, York and Sandown, horses whose forecast prices were at least 20% longer than their starting prices won only 15.9% of the time and lost punters 22.8% of the money they bet.

Such strongly bet horses won 17.1% of the time on Grade Two tracks, 17.6% of the time on Grade Three courses and 20.4% of the time on Grade Four tracks, where they cost only punters 6.5%.

I think the cause of this trend is almost certainly our old friends the Big Three off-course bookmakers.

They are by far the biggest influence on the betting market, and are unquestionably more active on the higher class courses.

They tend to send so-called "office money" through to the track, when the meeting is a big one, especially if it is televised. They are less active at the smaller meetings, because they do not have so much turnover at stake.

Office money is obviously not inspired gambling on the part of the horse's connections. It is simply an attempt to shorten up the price of one of the fancied horses in order to increase off-course profits.

It follows that on tracks where office money represents a larger slice of turnover, market moves will be less significant.

Similarly, market moves appear to be less meaningful in handicap races. It is in these races of course, where off-course bookmakers try to stimulate betting action most (presumably because their profit margin is bigger).

In any event, you lose 22.8% of your betting money if you follow horses whose forecast price is at least 20% longer than the SP in handicap races. In other races, you lose less than 20%.

The academics, who produce learned papers on betting markets, would probably categorise office money as a type of "silly money". This is the term they use for betting money that has little significance.

Their quest is for what they call "smart money", that is money bet by people who have a better insight than most (presumably the horse's owner and trainer, and perhaps professional gamblers).

Smart money, according to the experts, is invariably wagered late. The shrewdies apparently don't want the horse to shorten up too much in price, so they hold off betting until the last possible minute.

I'm not sure this is the case in Britain. If you've ever visited your local betting shop, you can't fail to have noticed the general stampede towards the counter that occurs when the announcement is made that they're under orders.

From my personal observations, I would say that while smart money may be bet late, so too is silly money and plain daft, desperate and even drunk money as well.

It is true that one of the big bookies did produce a paper a few years ago which seemed to indicate that big punters who held back their bets till the last couple of minutes actually made money.

But a close analysis of their research reveals that their sample size was so small that one decent priced winner would have been enough to skew the results completely.

My own personal quest for smart money revolved around situations where the stable seemed likely to be trying especially hard. In such situations, I reasoned, any betting action might well be significant.

My first stab in this area was to look at the type of races where stable gambles are thought to be landed most frequently - namely sellers. Sadly, it turns out that horses who shorten up considerably from their forecast price (by 25% or more) still lose you over 15p in the pound.

Gambled on horses in maiden races do only marginally better, costing 13p in the pound.

After researching umpteen angles, I only came up with a couple that made a very small profit.

Firstly, when a horse is the only mount at the meeting of a jockey who rides at least 12% winners, it appears to be significant if they shorten up in price from their betting forecast (i.e. forecast price is 20% longer than SP). They earn you a 7.9% profit.

Secondly, horses that travel at least 220 miles to the meeting are worth a close look if they get bet down in a similar way. They earn you 5.4%.

Obviously, in both these cases, there are good reasons for thinking that the betting action is significant. Good jockeys don't go to a meeting for a single ride, and trainers don't send a horse a long way unless they mean business.

ALL in all though, after researching scores of angles, I am forced to the conclusion that when a horse's price shortens from its forecast odds, it is generally not a good bet. The value has gone from the price.

In fact, it seems easier to make money by betting horses whose odds drift upwards. For example, it is pretty obvious that when a horse from Britain's biggest and most expensive training centre, Newmarket, is sent a long way (220 miles plus) to contest a mere maiden race, it must have a great chance of winning.

But you can't make money betting on the ones that start at their forecast odds or shorter. The profits only kick in when such maidens drift out from their forecast odds. They have returned a 45% pre-tax profit over the last two seasons.

I am sure that this will prove the same in virtually all cases. And it is only commonsense.

After all, if a particular type of horse wins at a particular rate, it's daft to bet it if its odds go below the point that your statistics indicate offers value for money.

Newmarket maidens, for example, win 25-33% of the time when they travel 220 miles or more to a meeting. You are therefore just bound to lose money if you only bet them when they start at less than 2-1.

The lesson here I believe is that you should be a great deal more demanding of the prices you take from bookmakers. Before you make any bet, you should assign a price that you believe offers value for money.

If you cannot get that price, you simply shouldn't bet. If the odds drift upwards from your minimum price, however, you should step in and bet again.

I know it's hard to do from a psychological and emotional standpoint. But the statistics and the mathematics indicate that in the long run it is foolish to follow the market. The way to make money is by opposing it.