Betfair | Betting systems examined...

Betting systems examined...

Wayne Bailey weighs up the pros and cons

There's one approach to betting that's loved and loathed in equal measure: systems.

The reason people love them is because once the initial groundwork is done, you simple follow the rules and find your selections within minutes. Others however, would baulk at the idea of backing a horse because a set of rules says so.

For those who aren't familiar with systems, they simply involve following a number of rules in order to come up with the selections to bet on. Nearly every punter has some sort of systematic way of betting whether they realise it or not. The guy who backs every second favourite is essentially following a system (albeit a losing one).

The rules of a typical system may read something like:

• Handicap hurdles only
• Horse was 4th or better last time out
• Horse is trained by Joe Bloggs

Every time a horse matches the criteria, a bet is placed.

How are systems created?
Systems can be created in two ways. The first involves trial and error either on paper or with real money, recording the results as you go along. But by far the quickest way is to use data mining software. An idea is tested out, and if it proves profitable, it can be followed.

Do systems work?
Like all types of gambling, systems involve risk. Sometimes systems go bust for unexplained reasons, yet others will continue to churn a profit year after year. This is why you shouldn't rely on one system alone. No smart investor would buy shares in just one company, and it's the same with systems. If one of your systems is showing a loss, hopefully your other systems will compensate.

Why do systems fail?
Systems fail for a number of reasons:

Poor staking
This is the most common reason. Punters let their own opinion come into play and start putting extra money on horses that they feel is a 'cert'.
When using systems, very strict discipline is required. The staking plan that has proved successful historically must be adhered to, and long-term thinking is required.

Too many followers
Systems that are out in the open for a long time lose their value as more people get involved, and the market readjusts. Very often however, and old system can be tweaked slightly and it becomes profitable again.

Data is 'fitted'
If you analyse any set of data for long enough, you will find a trend eventually. In the lotto, one of the balls has come out the most number of times historically. This does not mean that this is a 'lucky' ball and should be followed. The fact of the matter is that one of the balls MUST be drawn the most.

Psychological factors
When using systems, you must change your whole mindset regarding punting. You must see the system as an investment, and not deviate from the rules and staking. The downside of this is that it often takes away from the enjoyment of following racing. Your system may tell you to back the outsider where you clearly feel it is a no-hoper. You must remind yourself that the system is based on proven historical data, and NOT your opinion, otherwise it's a game of luck. If you want to play a game of luck, go do the lotto.

People trying to 'get rich quick'
If you're thinking of doing a few systems, I'd hold off telling the boss where to shove his job for now. The reality of gambling is that over 90% of people come out with a loss. If your system doubles its bank in a year, consider yourself very lucky. If you feel disappointed that your starting bank of £100 is only £150 by the end of the year, then systems are not for you. Consider though, that a 50% increase in your investment is a lot better than your high street bank will offer. Certain punters are not good at handling losing runs, especially on systems with a low strike rate. After a few losers, many punters will drop a system before it has a chance to get off the ground.

If you're interested in seeing some examples of systems, post a comment and we can provide the rules of one or two for discussion. Similarly, if you have a system of your own, post it up for people to try out!

20 December 2007 / About Wayne Bailey

Tags: Betting Systems / Digg It / Del.icio.us / StumbleUpon / reddit

Comments (9)

Wayne

Very interesting article; as you say it is easy to make the data fit and find a trend. Recently, I have been researching results and considering formulating systems using well known software, the difficulty seems to be finding something that offers a consistent LSP year in year out. Do you think that a system should, in the first instant, provide a LSP? Also, form performance based systems seem to provide little or no value, as punters appear to have latched on to the selections, as you might expect. I would be interested to research other elements in a system, course, going etc, what do you think?

Bill | 28 December 2007

Thanks for the feedback Bill.

For me, the strike rate will be the first thing I check, followed by the LSP.

(For those readers that are unfamiliar with the term LSP, it stands for Level Stakes Profit.)

Some sort of consistency in the strike-rate is usually a good sign that your idea is on the right track and if it shows a good LSP, then all the better. LSP is obviously very important but can be misleading on occasions. A single high priced winner could alter the LSP on any given year and skew the results somewhat, so consistency in the strike rate is just as important. Personally, I don't like systems with strike-rates of less than 15%, as the losing runs can be difficult to handle. It all depends on your temperament I suppose. I normally look for at least 20% strike-rate, or higher if the prices are likely to be short.

Regarding form-based systems, it's true to say that they are difficult find value in. I like to try to spot angels that others wont bother with. If a horse has been pulled up or fell last time out for example, punters will be immediately put off and this may be a good place to start. Certain jockey's are successful with horses that fell last time out and this angle can be exploited (see this article for one particular system based on that:

http://betting.betfair.com/horse-racing/betting-strategy/when-is-a-bet-on-mccoy-value.html )

That's just one example of where it's good to bet against the crowd and the conventional idea that you shouldn't back fallers again.

Course based systems can prove quite profitable as long as you keep up to date with any changes to the track. (For example, Southwell's track was re-laid after the weather damage and this may change things a bit). I've a system based on favourites, as backing the favourite in certain types of race at certain tracks shows a good strike rate and LSP year on year.

Hope some of this helps! If you've any of your own, feel free to post them up.

Take care,

Wayne

Wayne Bailey | 03 January 2008

Wayne,

Very interesting article. Do you know of any companies that provide historic horse racing data in an electronic format that would be suitable to load into a data mining application? I have been looking around for a supplier but haven't come across anything suitable.

Thanks James

James Harris | 07 January 2008

Hi James,

There are three principal companies that provide horse racing data that can be mined:

RSB (Racing Systems Builer) http://www.rsbweb.com/IE/IE_index.htm
RSB give a free trial CD, but it only contains data from 1995/1996 just to give a taste of what it can do. It must be said that it's pretty powerful and can run all sorts of queries. There are hundreds of variables and the option of getting your system qualifiers sent to you daily (at an extra cost).

Raceform Interactive
http://www.raceform.co.uk/
RI has been around for a while now and has a good reputation. Again, almost any query under the sun can be asked, and the results are provided. The best part about it is that official formbook comments are available, and can be mined. Unlike RSB, they don't have a trial version, so don't buy unless you really feel you will get some use out of it.

Timeform-i
http://www.timeform-i.com/
This is a pretty exciting exciting development. It combines the timeform database(including access to their ratings) with a powerful systems builder. The whole thing interacts with the betfair interface too. One of the best things about this is program is that you can set the parameters for your system, and it will automatically inform you when you have a qualifying horse. The downside is that it's not cheap. The full version for 12 months cost over £1,300 so you'd want to be serious about your racing if you are considering buying. They don't offer a trial, but you can buy 3 months for the flat or NH (or both).

Hope this helps,

Wayne

Wayne Bailey | 08 January 2008

Thats excellent Wayne, thanks for the help.

James

James Harris | 08 January 2008

Hi Wayne,

I think I have a kernel of an idea for a system. I have noticed in particular small field chases that runners with a win LTO often go off at biggish prices (i.e. 5-1+). This may be because they won a lesser-grade race or that they are up against a hype horse.

Do you think there is any lileage in this?

Cheers,

Stu

Stu Bailey | 17 January 2008

Some interesting thoughts Stu.

I've had a look through some data and found a couple of interesting trends relating to your theory.

Firstly, I looked at horses running in small field (seven or less) chases, who won last time out, and are priced 5/1 or more.

Interestingly, 2007/8 has shown a profit so far, but previous years show heavy losses. It did not make much difference to the results if the horse was stepping up in class or not.

Distance wise, it was the longer trips (3m+) that showed by far the best return, and the most successful type of race was novice chase.

This left me with a number of revised rules to research:

· Novice chases only
· Distance is 3m or more
· Race has seven runners or less
· Horse won last time out
· Price is 5/1 or more

The above criteria had 18 winners from 118 bets (15.3%) in the past eight years and provided a profit of £560 to £10 stakes for the backer using official SP ( £1001 to betfair prices). This is a very decent return indeed, but it comes with a few warning bells...
The qualifiers are sporadic, so you may be waiting weeks, or even months for a bet, which can be frustrating.
Also, not every year has shown a profit so you'd definitely have to be in for the long term, and not miss any qualifiers in case it's a winning one.

Historically, the longest sequence of losers in a row was 14, and the longest sequence of winners was two.

It's certainly a new angle for last time out winners, and some good food for thought. Thanks for the input Stu!

Wayne Bailey | 17 January 2008

Cheers for churning that out Wayne - sterling work as ever.

The only other sporadic system that I use is dosage index for early season 2YO races. For the past two years I have had two big-priced winners in the same race at Bath (Carson's Spirit and Vhujon) that have ensured it has been profitable for me.

The rule is back the two horses in each race that have the highest Dosage Index.

Stu Bailey | 23 January 2008

Thanks for that Stu.

I've never used the Dosage index but it's something I want to delve into further. I've shied away from 2YO races in the past, but breeding could be the key. Might try it out on the 2008 flat season.

Thanks for the info,

Wayne.

Wayne Bailey | 23 January 2008