Racing systems: Using systems to create shortlists

In my first article I gave readers an introduction to racing systems, writes Dave Renham. In this follow up I am going to look at a slightly different way of using systems, namely for shortlisting. and we'll use the Cheltenham Festival as an example.

Most people prefer the clear cut system method: a set of rules from which, when a horse qualifies, you simply back it; the theory is that this system has proved profitable in the past, and should continue to do so in the future.

However, for those system punters out there, not all systems remain profitable despite past results. I discussed in the previous article that many systems, even very good ones, are likely to have a limited shelf-life in terms of profitability. Meanwhile some other past ‘winning’ systems are basically flawed due to poor construction and have only proved profitable essentially because of luck or back-fitting.

But systematic approaches can also be used in a broader sense, for instance to create shortlists for specific races. Let’s review how that could work.

Creating a Cheltenham Festival system

With the Cheltenham Festival around the corner I thought I'd offer a shortlisting system idea that gives us a chance to make a profit at the meeting. This system will produce a shortlist of runners that is about as easy as you could get. The system rules are these:

1. Non-handicap races at the Cheltenham Festival
2. Top four in the betting

That’s it. The focus is just on the top four horses in the betting in these races. You may be thinking how is this system potentially going to make me a profit? Well the strange thing is that if you had bet all runners from the top four of the betting in every non-handicap festival race since 2009 you would have made a blind profit to Betfair SP. If you had placed £10 on every selection you would have made a profit of £661.50; a return of just over 7 pence in the £. The graph below how each year would have panned out from a return on investment perspective:

Nine of the 13 years produced a profit; and losses in the other four years were far from damaging. Strike rates were pretty consistent, varying from 14% in the ‘worst’ performing year to 20% in the best.

One issue with systems that use market factors is that it is not always possible to be 100% sure which horses will finish where in terms of position in the betting market. Of course, the later you place the bet or bets, the easier it is to be confident the horse will end up in the top 'x' of the betting. I appreciate, though, for some people this is not an option; but the good news is this ‘shortlisting system’ would have worked equally well if you had used the top four of the betting forecast in the Racing Post.

Before examining what we can do next with our shortlist of four runners, let me quickly discuss why I think this idea might have worked in the past. The one thing we know about the Cheltenham Festival is that trainers and their horses will be giving 100 per cent. In non-handicap races we are going to get fewer shocks and more races running closer to form. Yes, the races will be competitive, but even so wins for outsiders will be rare. Indeed, horses priced greater than 25/1 have won just 11 times in the 13 years from 1197 runners. Backing all of them would have secured a loss of 20p in the £ to BSP.

In terms of longevity, this top four in the betting idea should not have a shelf-life per se, as the prices of these runners are going to be very similar in years to come. So, as a shortlisting method, focusing on the top four in the betting in these races looks a highly logical starting point.

From here it is a question of personal preference in terms of what to do next and how much time you wish to spend on each race. For those who have no real time to devote to form study there are two possible options. The first option is that you could just back all four runners each time. A second option is to look for one further rule to incorporate into the system. Once found, assuming one can be found, the selection process is short and swift.

Looking for an extra system rule brings into focus the trap of back-fitting so one needs to be careful. I think before testing any further ideas out, we ought to draw up a list of possible rules that make sense. Here are some ideas to that end:

1. Look for horses that have won that season
2. Look for horses that were sensibly priced last time out
3. Look for horses that have not been off the track for too long
4. Look for horses with sound sire stats

This is not an exhaustive list, but for the purposes of this article it makes sense not to examine too many. Let us consider them one by one.

Horses that won earlier in the season

It makes sense to me to focus on horses that have won at least once earlier in the season. I am someone who generally looks for horses that have won before and at this level I think it is crucial. The problem of course, as you may have guessed, is that it will not trim the shortlist much! For horses to be in the top four of the betting at the Cheltenham festival it is likely that they have shown a decent level of form that year. This extra system rule knocks out just over 100 qualifiers and overall profits improve by £3.56!! This equates to an improvement in terms of returns on investment by around 1p in the £. All in all, not a system rule I would be too fussed to add.

Horses priced 'sensibly' last time out

Looking for horses that were not huge prices last time is something I do if I am focusing my attention on the front end of the betting market in non-handicap races. My reasoning is that horses that were big prices last time out were clearly not fancied. If they ran well and outran their odds I think there is sometimes an overreaction next time and their odds in the current race may end up being shorter than they perhaps should be, thereby offering little or no value.

The difficulty for any system punter once they come up with a rule that involves price in some form or other, is to choose the right price. Not only that, this is where back-fitting can easily raise its head. It is so tempting to look at the stats, see the price bracket that gave the best profit and choose that. But, as we know, this is a doomed approach. We should choose a price that seems to make sense and stick with it, regardless of what the data spits out. In non-handicap races I would rarely back horses priced 12/1 or bigger, so the price limit I would use here is LTO price 11/1 or shorter.

Looking at the results of adding this system rule we get a similar pattern to the basic system:

Overall returns improve a little to 9p in the £, but ultimately only 52 horses were ‘knocked out’ from the original group of just over 900. So again, for me, it would make little sense to use this rule.

Days since last run

When dealing with the fitness of horses, I am always a little sceptical about runners that have been off the track for too long, especially if it is later in the season like the Cheltenham Festival is. Again the problem is choosing the right time frame, and sticking to it. With Cheltenham taking place annually in mid-March I would not choose a specific days since last run figure, but simply use January 1^st as the cut off. So what happens if we stick to horses that have run at least once in the calendar year? Here are the stats:

We have trimmed the number of selections a bit more here. This extra rule would mean backing three horses in each race on average. However, the figures have not really improved. A slightly better ROI%, but due to losing 200+ selections actual profits have dropped by around £110 if betting £10 level stakes. Overall this might be an option for some people but, once more, it's not one for me.

Sire Stats

Using a sire rule for a system is a rarity for most, but actually sire stats are quite detached from the usual system rules punters use. They also don’t really clash with other rules/variables, that is there are not normally any 'related contingencies'. For that reason, I often like to investigate sire angles. Moreover, sire data is not as accessible for many punters as the more bog standard racing data.

So what parameters to use for the sire system rule? I will keep it relatively simple by sticking to sires that have had a strike rate of at least 10% in the last year, assuming they provided over 25 winners during that period. I could have chosen a near infinite number of sire stat combinations, but this is relatively simple to check so that helped in making my decision. The results for the top four of the betting combined with this sire rule present quite a rosy picture, at least on the face of it:

We now have just over two bets per race on average and profits have soared. To £10 level stakes you would have been in profit to the tune of £1645.80. This gives returns of nearly 34p in the £. The yearly returns are broken down below:

The pattern remains similar to our original graph but this one is more volatile, as might be expected: 2014, for example, saw a strike rate success of just 5.1% (two winners from 39) which is the reason for the low ROI% figure for that year, whereas 2015 saw a strike rate of 34.5% (10 winners from 29). The more you drill down into systems the more likely they are to show volatility. It is down to personal preference whether that is tolerable or not.

The question now is, ‘should we incorporate these additional sire parameters into the original system?’.

Well, there is no easy answer. I must admit I have doubts this adapted system would perform as well over the next 13 Festivals securing profits of over 30p in the £, but I certainly do believe it has a good chance of showing some sort of profit long term.

With systems, racing, and betting, there is no "right way", no single correct answer. However, there are plenty of wrong ways and wrong answers and, as punters, we need to avoid them as often as possible. System based betting continues to divide opinion, but using the sound logic of a basic system as a starting point does make for some solid options.

As mentioned earlier you don’t have to go down the route of additional rules that I have looked at here. Instead, you may simply focus on the quartet that head the betting and treat it as a four runner race, using your usual race reading techniques to land on a selection or selections. This more flexible approach will appeal to some of you, I hope.

Before finishing, I want to remind you of two NH systems I mentioned at the end of my first article.

System 1: Won at Cheltenham LTO

Just one rule. This system had a bumper 2021 making huge BSP profits. Before that, from 2009 to 2020, it made a small BSP profit of £100 to £10 level stakes. However, anything as basic as this which has broken even or better over a long period of time is a really good starting point from which to build.

Cheltenham winners in general are pretty smart which is probably why results have been good over the long term – also these last time out Cheltenham winners go on to win 25% of their next races, which is greater than 4% above the average figure for last day winners.

System 2: Quick returning winners

1. Won last time out
2. Last ran 5 days ago or less

Quick returners have been the basis of many systems in the past and this is one that has continued to prove profitable in recent years. Specifically, it has achieved BSP Profits of £924.70 to £10 level stakes since 2009.

Both of these systems, with profit and volume of runners, offer good starting points from which to investigate further. Good luck!

- DR

Other Recent Posts by This Author:

• ‘SR’ Ratings on the Flat (Turf)
• From a Place, to a Different Place
• Comparing Starting Price with Betfair SP
• Favourites on the Flat in April
• Early Flat Season Trainer Form