Roman Koz provided us with a fascinating look at stats about favourites on pages 18-19 in the November PPM. It really is a good way to start thinking about a structured manner of attacking your betting and is something I have enjoyed fiddling around with all my racing life.

However, he expressed some surprise about a couple of the findings and so I thought that this month I would share some stats of my own to show that there is some agreement here and the findings he revealed really do stand up.

Firstly, my sample will be a little different. He looked at all favourites in the price range from 1/1 to 5/1 running every day of the week. I shall look at a system that works only on Saturdays and public holidays but I feel it allows some conclusions to be mutually reached. This system is one I have referred to before and was developed by myself and a keen racing mind that also loves figures and the recording of them.

My partner in this research is John Williams - no, not the guitarist, or the composer. John's music is the sound of the TAB responding to the payment of a winning ticket - a sound neither of us hears as often as we would like.

Nevertheless, John has kept records of our method (no rules to protect his slim profits!) for over 12 years and some 3631 bets. A fair sample you will agree, and it shows a profit of 3.7 per cent at this time. Now we have asked what can we learn if we break these results down and look at subgroups of results and one such area of analysis was days since last start.

We found early in the piece that results for 14 days were so bad that we stopped recording them. As Roman found, favourites that started 14 days ago seemed to do worse than most - this group suffered a 17.99 per cent loss on turnover.

But without going through all the groupings, we found results that would mirror and support Roman's results. Working on horses that started recently - here meaning 1 to 5 days - we dropped down to 169 bets but profit rose to a staggering 68 per cent!

Even pulling out a 60/1 winner in that group on the grounds that it could be an out-layer still gives 32 per cent return. Clearly the message is to look closely at any horse, but particularly a favourite or in-form horse that is backing up smartly.

We also found that results for horses starting again in 7 days were poor and this may reflect the tendency to overbet these horses. Some people will religiously bet on a last-start winner in seven days even though long term these are poor bets because of price. But it should be noted that even in Roman's sample the result was no worse than average - it just was not as good as we thought it might be.

Now one question we asked, that last month was not touched on, was what if you further break down the 8 to 13 days group, which was noted as losing only 4.85 per cent and so showing "some promise"?

Well, Roman, me old shiner, go back to the computer and look at days 8 and 9. We found from our sample that the number of bets dropped down to 697 but profit rose to 12.7 per cent. Perhaps favourites in this group may show an improved result as well.

Ten days, which for us was Wednesday to Saturday, was a terrible loser. So if we just focused on days 1 to 5 and days 8 and 9, we had about one-third the bets and profit rose from 3.7 per cent to 18.07 per cent. Who said we should ignore stats in racing?!

Incidentally, we found form from Sunday racing (six days) was dreadful in our sample.

Before moving on, I will throw in some other thoughts from our work as it may help other system buffs or analytical thinkers. We know that females against males are not popular with punters, and even some trainers, unless you are talking of a Sunline type.

Well, worry not. We found that females did not have a worse winning percentage but paid better, perhaps reflecting the bias against them. Profit rose to 11.9 per cent by betting on those races where a lady was our choice against the guys.

Barriers I have mentioned before in other articles but to briefly repeat for new readers - barriers 1 to 8 do win significantly more but pay less, so more profits await you if you gamble on the draw and survive the longer losing runs.

As with Roman's findings, we did not see distance as variable - our results were similar across all journeys. And another heretical view, claiming jockeys won as often but gave better dividends. Lower weighted competitors also had much better winnings.

And do not be too alarmed if your choice comes back in distance. A small sample (63 runners) back in distance by a whopping 300 metres or more, showed a very healthy account.

So you can see that stats and record-keeping plus regular analysis can help in the battle. Hopefully, you have received some thought provocation from some of the findings above - and keep those figures rolling, Roman.

Until next month - have fun punting!

By Clive Allcock

PRACTICAL PUNTING - DECEMBER 2002