Archive for November, 2011

If it ain’t broke….

November 6, 2011 Leave a comment

First of all I would like to apologise for the lack of recent updates here. I thought being back at work would focus my efforts more during my spare time and as a result I would be updating this blog on a more regular basis but that’s not really been the case for a variety of reasons.

Going back to work has been quite a shock – in a good way. I’m surprised at how easily everything has come back to me. Getting up in the mornings isn’t a problem. The commute to work takes a bit longer than is ideal but isn’t any real hassle and I can do it half asleep. Plus quite a lot of the commute involves walking which gives me plenty of time to think about things and get my head in order, which is very welcome. The work itself is enjoyable and despite the fact I haven’t really used my software testing skills over the last couple of years I have been able to get straight back into it and am making a real contribution. The domain I am working in (oil industry) is one I have worked in before, in fact I have worked with these very same developers before which has probably helped me settle in quickly. I’m only supposed to be part-time at the minute but there is a lot of work on as we’re approaching a software release so this last week I ended up working the full week, and I don’t feel any worse for it which is an especially pleasing thing. But it does mean I haven’t had a lot of spare time of late.

What spare time I have had to spend on gambling activities has been spent making a number of improvements to my betting records. I mentioned recently that I had taken a spanking on one of my lay systems (Lay’em) and as a result I had started to record more details for each horse racing bet, namely the racing code and whether or not the race was a handicap. I wrote some VBA to go back all my previous bets and update them with the same information but it took a bit of time to get that code working 100% and to verify the output was correct. I also had to clean my data a bit as this code highlighted a few mistakes in my earlier records, errors I have now addressed. Once I had done all of that I needed to decide on how best to use this information. The idea behind recording these additional details was to allow me to break my systems down and look at stats based on individual racing codes.

The example I gave at the time was to look at whether the Lay’em performance was better on the turf or AW but I soon found that the way I displayed stats for each system/tipster was too constrained and wouldn’t easily extend to cover different racing codes so I had to redesign the way I handle all my stats. Initially I didn’t really know what the best way of doing this was and I began to write a number of user-defined functions that would calculate the stats for each system for each racing code on the fly. Unfortunately it soon became obvious that that way madness lie as my whole spreadsheet regularly ground to a halt while I recalculated stats over more than 70000 horse racing bets. Unacceptable so I had a rethink and went back down the normal VBA path with routines that are run as and when I tell them too rather than every time anything changes.

Throughout this I wasn’t really sure of the best way to present my data so I ended up playing with a number of different output formats. I knew I needed more stats but wasn’t sure of the best way to present them in a way that would mean I could get the best out of them. This is the wrong way round for me – I normally start with a much clearer idea of how to present the data but then I have to work backwards to generate the data in the first place. And of course every time I changed my mind about how best to display this data I had to tweak the VBA output code to conform to the new style. I eventually worked out I wanted separate sets of stats for my current portfolio and for all systems/tipsters regardless of whether or not they are part of my current betting plan. I wanted a breakdown of each system/tipster by racing code (or sport or whatever) but also wanted the figures broken down by month so I could view my progress throughout the year and it was this that caused me a few issues initially. But I got there even if it did take rather long than I had hoped for. The glory of it now is that I have had all this in place for a week or more now and have yet to find reason to change it which suggests I have ready access to all the stats I think I need. At bloody last!

So returning once more to the Lay’em system, which was after all the driver behind all these changes, I can now compare the turf and AW figures. It seems the system is profitable on both, netting 187.66pts on the AW and 117.56pts on the turf. Looking a bit deeper though the AW ROI is 1.68% compared to 0.26% on the turf. Concentrating on results from just this year (which I can do just by setting a combo box – yay!) the system has seen a poor time of it on turf with a loss of 17.53pts but it’s another story on the AW with a profit of 56.81pts so I am pleased the turf season is now over and we’re into the AW racing. Mind you, last year the turf lays made a profit of 72.01pts compared to 23.79pts on the all-weather so it goes to show you never can tell how things well go – past performance being no guarantee of future success and all that.

The long and short of it though is I wanted to look at the Lay’em system in more detail as I had taken some beatings on this one recently and had thought about deconstructing it and reviewing the rules via a system builder. That was around mid-October and with a remarkable change in form the system went on to record its most profitable month ever and I have been running it for around five and a half years now! The second half of the month saw a profit of more than 50pts which certainly helped turn things around. Had I not had such a month of two halves I may not have thought to add in the additional data I now record and I would still be in the statistical dark ages as it were so the poor run at the start of October has certainly put me in a better place overall.

Now I have these stats what can I do with them? The main way I plan to use them is to help me press my edges better by tailoring my staking for each system/tipster according to their strengths and weaknesses. For each racing system in my portfolio I now have a full breakdown by racing code and for the sports tipsters I have a breakdown by individual sport. I considered breaking down the football bets my division or betting market but I couldn’t really decide what I would do with such info at this stage so I have held off that. I can always revisit that decision in the future. Anyway, the idea now is that I can examine the figures for a tipster such as ProBandit’s A List and note that the overall ROI of 1.18% doesn’t tell the whole tale by any means as the AW bets have profited at a rate of 6.43% while the flat turf bets show a losing ROI figure of -4.11%. Take Sports Bet USA as another example. This service advises bets on all four major US sports (American football, baseball, basketball and ice hockey) but the performance across those sports is anything but equal. The NFL bets are running at a ROI of just over 12% but the corresponding figure for the baseball bets is -42.27%, albeit over a smaller sample of bets. It goes to show though that not all tips are equal and I should be able to improve my bottom line by staking more where the tipster has a good record and less where his performance has been poor.

Before I make any more staking changes I need to sit down and properly review the figures. For example, had I looked at last year’s Lay’em figures and changed my staking as a result I would have been staking more on the turf lays than the AW lays this year and as a result may not have made a profit at all. I also need to make sure I have an adequate amount of data on which to base my analysis. The Sports Bet USA baseball bets may show pretty hefty losses but those figures are based on just 31 bets. I want to be sure that I have a significant number of bets before deciding to deviate from my usual staking for that tipster, although it’s hard to put a definite figure on that minimum number of bets. I have a figure of 100(ish) in mind but bear in mind that is 100 bets for each given sport or racing code, not just 100 bets in total. I have recorded over 500 bets from the Insider service and my figures show the tennis tips show the greatest rate of return but as I only have 79 tennis bets from that service I don’t feel I have enough evidence to support an increase in stakes relative to other sports, especially as the bulk of the profits came from a single bet a couple of years back.

I plan to review my whole portfolio with a view to adjusting stakes based on the figures for individual racing codes/sports so that I press my edges better. I will update this blog with all the review details as writing things down helps clarify things greatly in my mind. It’s as if by explaining it to someone else I have to first get it all straight in my own head. But for now the work I have put into my betting records over the last couple of weeks means I won’t be running the Lay’em rules through a system builder – the system is clearly profitable as it stands and if it ain’t broke, don’t fix it.

Categories: General