Monday, 18 June 2012

European Systems Download

Similar to what I have done with the system reviews from last season, I have pulled together a summary of the performance of all the Euro systems during backtesting.  You can download the spreadsheet using the following link:

The spreadsheet is over 10MB, so may take a while to download depending on your connection speed.  I suggest you download it and save it somewhere you can access when it comes to reading the previews of the Euro systems in the next week or so.

Here are a few caveats when looking at this data:

·         The systems have no live results. Historical results may prove to be a good indication of future performance but sometimes, systems don’t perform anywhere like they did during backtesting when they go live. I explained why in the last post and won’t repeat it here.

·         Be careful when comparing results by season. Seasons 2006/7, 2007/8 and 2010/11 are skewed somewhat by some of the data being used in the backfitting process. Hence, if you want an unbiased view of historical profitability, ignore these 3 seasons completely.
·         Analysing backtested results may help you spot potential weaknesses in the rating algorithms or systems but be VERY careful about increasing or decreasing stakes based on a certain criteria unless it is based on a large sample of live data.  IMO, a large sample needs to be hundreds of bets, possibly thousands of bets. If you start interfering with the system bets and trying to cherry pick a certain type of selection, you are likely to decrease your chances of winning long-term.

·         Don’t buy into trends too much as it is only a trend, not a guaranteed change. Hence, a league which may be the best season one season may be the worst next season.  Likewise, a league which is losing one season may be winning next season. Don’t fall into the trap of backfitting your own portfolio to follow!

·         Be careful about writing off the Asian Handicap returns as clearly, when you see the graphs compared to the outright returns, they are nowhere nearly as good. However, for a subsection of bets, it may be the case that AH can smooth your returns.  In addition, I believe the historical backtested returns for AH returns may be understated a little for the Euro systems, so be careful about dismissing the returns completely.

·         There may be very minor data issues apparent within the data.  With any large samples of data, there can be small discrepancies thrown up within the data.  As far as I am aware, if any issues are found, they won’t be material or even close to being material.  Of course, this data is only a few weeks old, so as we progress, we may spot something that requires investigation but to my knowledge, the data is reliable and trustworthy.

OK, the next step is for me to walk us through the systems one by one in blog posts and hopefully highlight anything to look out for. The conclusions should help us populate the summary performance table for the 11 Euro systemns as shown here.  


No comments:

Post a Comment