Aliando's Football Ratings 2007/8

[aliando]

I've spent many a long night by candle-light working on this system so hopefully it'll pay off. It's an expansion of my old "computer says yes" thread with an upgraded prediction spreadsheet. The short story is that I have used formulas to give ratings to every game each team plays. Similar formulas are used to rate each teams recent form and these ratings combine to give an overall match value. Due to the recent form element of the formulas, I've chosen not to use the predictions until there have been 12 rounds played in the division - the rationale here is that teams will have had time to settle and the overall match value should be more accurate. Technically, the results should show that the higher the rating, the stronger the probability for a home win. I will be using the English Premiership through to the old Conference divsion. Initially, the intention was to get a system that predicted winners but early results appear to show that laying the away team may be the most profitable way to go. I've not had even nearly enough time to backtest this but have looked at the Conference since Round 12. The next post shows the results returned and a small analysis of the results in order to show where my thinking comes from. I'll put up similar stats every now and then when time allows. Its early days, I know, but once the system is up and firing fully, if any of the staistically minded folks can analyse the info and demonstrate the best way of maximising profits then I'm happy to listen (but please use simple language for me :$). I'll be using 3 staking systems initially: Staking Plan A - Boxed Singles Bets laying the away team will be placed one at a time and profits mounting up until the box has completed 5 successful lays. An unsuccessful lay means a losing box and a loss of the original liability. I will use a starting bank of 40pts and each box will have a starting liability of 2pts. Bets will initially be placed on matches that have rating of 200+ as this appears to be the cut off line where results are more predictable (see next post + posts 14 & 15) although I may need to up this level if the early figures are not continued. Staking Plan B - Top 8 Lays Accy Simply I'll use the top 8 lays to use in an accy. This will only be applied to weekend bets where at least 4 of the five divisions have as good as a full programme. I will either use betfair multiples or use a bookmaker through double chance bets. I will use a starting bank of 40pts and each bet will be 5% of the bank. Staking Plan C - Double Up I will use the top rating matches in order to get enough odds to double my stake (ie. evens/2.0) in an accy. This will only be applied to weekend bets where at least 4 of the five divisions have as good as a full programme. I will either use betfair multiples or use a bookmaker through double chance bets. I will use a starting bank of 40pts and each bet will be 5% of the bank. For all staking plans, I will have an intial target of 100pts where I will then take back my original bank (40pts) and continue from there. The post below will not be used for after- eventing, just to show how the early conference results have panned out. All banks start from this point. I will say now that with baby pending and work/family commitments that I may not be able to post all the time but will do so as much as possible. Although much of the workings are updated automatically, theres still a bit of manual inputting needed. I will update things here as we go if needed - please feel free to ask any questions through the thread as i invariably miss stuff out in the opening posts :eyes

[aliando]