Nick Mordin in The Weekender

Share on Twitter

Nick Mordin has written a number of books on horse racing (e.g. Betting for a Living; Mordin on Time) and has a regular column in The Weekender, which I buy every week. Nick is held in very high regard on the racing circuit, any book and his thinking on horse racing, speed figures, tracks and punting are a must for any racing enthusiast.

Reading Nick Mordin’s column in The Weekender, he makes some very valid points about hard facts being in short supply in horse racing – particularly in Britain and Ireland. Facts that are taken for granted in other countries such as the USA, include sectional times, horse’s body weights and medical records, full details of the the equipment they carry and official gallop reports. Even some of the so-called facts that are available to UK punters can be wildly inaccurate. Official race distances can be out by 50 yards or more.

Gathering real hard facts takes some effort but can lead to profit. In this week’s column Nick’s observations about horses that sweat in the their early starts would appear to be profitable – returning £28.31 to a level £1 stake.

Free bets?

Share on Twitter

Quite a number of the online bookmakers have offers of free bets. However, when you look at the small print, I’m not quite sure how ‘free’ these are. Often they are a matching deposit, up to a certain limit, based on the amount of money you put into the betting account when you open it. But, you can’t just open an account, deposit some money, collect the matched deposit from the betting company – and then withdraw it all. They usually want you to turnover the initial deposit by two or three times. After that, you might still be lucky to have some money left! I guess you could always bet on near certain bets – and most of the time you’ll get away with it. Such bets are more likely on football or tennis than horseracing – so it may still be possible to make money from these offers.

It is worthwhile shopping around – but be sure to read the small print with all of these free betting offers. You want to make sure that the conditions are such that you’ll get your money back – after all, the point of this is to make money and not line the pockets of the bookmakers!

Online training…

Share on Twitter

Talking about online training resources, as I did in my last post, when I mentioned videolectures.net – I have recently come across KhanAcademy.Org. Again a website with a large number of video based lectures on a wide range of subjects. I have found some of the videos in the probability and the statistics sections to be very useful and well taught. I think this is a resource that will get a fair bit of use from me.

While rummaging through mathematical books that relate to horse racing on the Internet – I came across Precision by CX Wong. If anyone has read this book and could give their thoughts on it I’d be very greatful. I might want to get a copy but don’t want to spend £30 without a bit more info on the book’s content and level. Unfortunately the Amazon preview of the book doesn’t show a table of contents or index, so it’s hard to tell.

Online learning resource – Bayes and MCMC

Share on Twitter

A really useful website that I have come across recently is videolectures.net. As the name suggests, this website has a wide variety of videos of lectures on a wide range of subjects. I came across a particularly useful lecture from the Machine Learning Summer School (Cambridge 2009) by Iain Murray on Monte Carlo Markov Chain (MCMC) methods, that helped clarify my thinking on the subject and how I might start thinking about applying Bayesian inference to horse races.

Talking of Bayes’ theorem – I am currently reading and enjoying Sharon Bertsch McGrayne’s book “the theory that would not die” (Yale University Press 2011). There is not much maths in the book – but a lot of background on the people that have been involved with Bayes’ theorem over the years and some of its practical applications. A riveting read!

Further investigations into horse racing

Share on Twitter

Over the last few weeks I have been reading all I can about horse racing handicapping and form. The works of Andrew Beyer, Mark Coton, Tom Ainslie and others alongside more quantitative approaches from the likes of Bausch and Ziemba – have been the mainstay of this background reading. Resulting from this I am starting to reach some questions worthy of further research and can then start gathering the necessary data and build statistical models to help determine which factors look to be the most profitable.

Some primary data will be gathered from the annual form books (Raceform Flat and Jumps books for the years 2009-2010 and 2010-2011) as well as the Racing Post Weekender. It may be that I need data that these publications cannot provide. The US Daily Racing Form (DRF) seems to publish more detailed timings information that UK punters are unable to enjoy.

And a possible fly in the ointment is the recent ‘guidelines’ from BHA about use of the whip, which are likely to have an effect on the way some races are run. Top jockeys may not ride where the risk of a 5 day ban puts a big pay-day at risk – and in some jumps races there may be no finishers. Let’s hope that sense prevails and sometime soon – otherwise I may have to change tack and look at another betting market completely.

More python wrangling

Share on Twitter

I have decided that I prefer using Python over Perl, and so I have started getting some basic Betfair functionality working to learn the basics of what I need.

I now have a script that uses the Betfair SOAP API to login, get a list of win and place markets for GB horse races, get additional information for each market and then store this information into a MySQL database. I have had to get to grips with a few Python packages for all this and it hasn’t been too painful. Now I just need to make it production quality and make sure the functions are suitably defined and packaged.

Scanning HTML web pages looks like it might be a bit of a challenge!

Python or Perl

Share on Twitter

I have been reading through Colin Magee’s book on automated exchange betting which makes use of the Betfair API using a number of perl scripts. I have been having some success with these, but perl somehow seems rather unsatisfactory and the scripts are not obviously structured, so I have started to recreate the functionality using python instead of perl.

My Ubuntu laptop has python 2.6.5 installed, and the only extras I have needed are SOAP functionality (so far). I was going to go by the recommendations of Dive into Python but have ended up installing SUDS instead. SUDS comes with decent documentation and I have managed to get the Betfair login and logout functions working without any real issues.

My next piece to deal with will be extraction of race and horse data from a suitable web-site. This functionality may need to remain in perl as it is quite strong with this type of activity (html parsing). I’ll research further what python libraries might help.

Plummeting and rocketing share prices

Share on Twitter

The recent turmoil in the stock markets round the world shows little sign of ending any time soon. I’m not sure how automatic trading systems can cope with such high volatility. Whipsawing in fairly normal trading can be an issue – now it’s a whole new ball game. Probably best to be in cash and stand on the side lines and hope sanity returns.

The ongoing Euro crisis needs to be resolved soon. Either the Germans need to throw loads of money at Greece – or Greece needs to leave the Euro and sort itself out. Ongoing death by a thousand cuts is not going to resolve the problem and allow any sort of stability to emerge.

Interesting times!

The Joys of Linux

Share on Twitter

I spent a lot of the week-end getting to the point where I could try out and further develop the perl scripts from Automatic Exchange Betting having decided to set up Linux on an old laptop.

Attempts to get Ubuntu 11 on the machine were foiled initially by a bad CD and then a hardware incompatibility. Trying a different Linux distro, Puppy 5.2.8, showed that I could get Linux on the machine and running – but problems resulting from missing libraries after installing additional software such as perl and mySQL made the machine unusable.

Finally I tried Ubuntu 10.04, and apart from one minor glitch with the screen going blank (which was resolved with some additional boot parameters) it has all been plain sailing. The old laptop now starts in seconds rather than the several minutes that my Windows XP laptop takes. Perl and MySQL are up and running. Now I can get down to the more serious action of developing the perl scripts!

Sports Betting

Share on Twitter

I have started to take a look at the investment opportunities presented by sports betting on websites such as Betfair and Betdaq. These two sites have an API which allows for easier connection to a computer based trading solution.

I am currently in the prices of acquiring data to perform some further detailed statistical analysis and determine what, if any, trading opportunities are likely to be profitable. However, if I do find some I won’t be able to share them here until I have made a decent profit first!