- This topic has 78 replies, 10 voices, and was last updated 20 years, 5 months ago by
.
- Posts
The aggregate goal supremacy for last week’s games(77), comparing the actual results with those predicted by the formula was -24. This is very accurate but it is only a relatively small sample, so I’ll be keeping a check on it. Ideally, the figure should be positive in the long term, since this would be evidence that the formula is a useful predictive tool.
I looked at the RFO goal supremacy predictions for the same games and their aggregate figure was -1153. The formula Alex Deacon uses gives more weight to the home advantage than the one I am using. I stress again that 77 matches is not a significant sample, but it could be worthwhile monitoring the RFO’s goal supremacy performance comparing actual and predicted results. I’m a little bit surprised that Alex doesn’t use the goal supremacies to monitor the performance of the index.
I’ve worked out the ratings for the weekend’s games and can only find two matches where there is a significant variance between ratings odds and actual odds.<br>                ratings odds      actual odds<br>Bolton(+31)         5/4            11/5(Coral)
Berwick(+22)        11/8            2/1(Lad.)
The recommended bets, therefore are:
Bolton to beat Charlton 25 pts@ 15/8(Bet 365, Stan James: Coral have trimmed price to 8/5)<br>Bolton won 2-1       +47pts<br>LAY Arbroath to lose 50pts on the exchanges @ 2.25 or less.<br>Arbroath won 2-0     -50pts
The main reason I started using the ratings was as an indicator of likely draws(with the pools in mind). In this respect, I would say they have been helpful bearing in mind the difficulty involved in finding score draws. I think they have found a few of the less likely draws this season but not sufficient for me to win any dividends using a small perm. In the absence of other betting opportunities, it could be worth supporting the top-rated draws this weekend. I know a few people like to do multiple bets using draws, but I prefer to stick with single bets.
Recommended: the following games to DRAW<br>20pts on each. The odds vary with different bookmakers, but Bet 365 just about has the edge overall and it will be worth checking out exchange prices. nothing should be less than 9/4.
QPR(+1) v Leeds 1-1<br>Torquay(+1) v Port vale1-0<br>Stenhousemuir(+1) v Montrose 1-0<br>Coventry v Wolves(+3) 2-2<br>Kidderminster v Yeovil(+6) 1-1<br>Mansfield v  Darlington(+6) 1-1<br>4 correct out of 6 at an average priceof 12/5 giving a profit of (192-120) = 72pts<br>A profit on the day of 69pts. The bank now stands at 1004 pts(a profit of 4pts on the starting bank of 1000pts). The target profit at the outset was 5 to 10 per cent on turnover, which given an average bet of 40pts should now be about 250pts+ from 79 bets to date. There’s still time to hit the target before the English season closes. Alex Deacon also produces ratings for the Australian leagues which will be worth looking at when our season is over.<br>
<br>(Edited by Artemis at 6:26 pm on April 16, 2005)<br>
(Edited by Artemis at 6:29 pm on April 16, 2005)
Last week’s 70 rated matches produced an aggregate goal superiority of -285, an average of -4 per game. Although it is still a very small sample to draw any firm conclusions, I’m reasonably happy. The figure should ideally be positive if the formula is to succeed in obtaining an edge over bookmakers/spread/exchange odds.
Quite a few bets lined up for the weekend, just waiting to see the odds in tomorrow’s Racing Post.<br>7 possible win or LAY the opposition bets and a short list of 10 draws with ratings of +8 and below.
Weekend’s recommended bets:
LAY means betting on a team to draw or lose(i.e. not to win)
<br>LAY Aston Villa (v Bolton +9) to lose 50 pts<br>Draw 1-1 +34pts<br>LAY Ipswich(v Leeds+9) to lose 50pts<br>Draw 1-1 +37pts<br>LAY Swansea(v Bristol Rovers+14) to lose 50pts<br>Bristol Rovers won 2-0 +37pts<br>LAY Sc**t
horpe(v Grimsby+6) to lose 50pts<br>Draw 0-0 +39pts<br>Lincoln(+36) to win 25pts @15/8(Victor Chandler)<br>Lost 1-0 -25pts<br>LAY Dunfermline(v Inverness +11) to lose 50pts<br>Draw 0-0<br>The following are the five best draws according to the ratings formula:Norwich v Charlton(+1)<br>1-0 -20pts<br>Crewe v Stoke(+1)<br>0-2 -20pts<br>Port Vale v Barnsley(+2)<br>5-0 -20pts<br>Motherwell v Aberdeen(+2)<br>0-1 -20pts<br>Elgin City v Peterhead (+2) <br>0-2 -20pts<br>20pts on each at odds of around 5/2(average)
<br>A profit of 55pts over the weekend. The bank now stands at 1,059pts.
(Edited by Artemis at 10:03 am on April 25, 2005)
Last week’s 80 rated games had an aggregate goal superiority of -371, an average of -5 per game.
The average since the formula was adjusted to Trend x 2 is around -4, an improvement on the original formula using the Trend x 5. I think this is very close to the optimum figure achievable purely from using a ratings formula. If you had a formula that gave a long term average with a + figure, you would be ahead of the experts in your estimations of how well teams can perform.
The season is drawing to an end and I think that I have taken this thread about as far as I can.
The Profit and loss Account is standing at a small profit, a bit below what I was expecting, but I think that’s because I became a lot less selective than I intended to be when I started.
I believe the profits might be improved, as someone suggested early in the thread, by taking account of local knowledge and up to date team news.
These ratings, however you tweak the formula, are a very good basis for starting your analysis. We have Alex Deacon to thank for that.
Best of luck.
Just a quick update. The RFO ratings and my suggested formula – see April 4th post and start of thread (or your own variation) should be at their optimum over the next two months or so. I did a check on Saturday of the pools games and recorded a goal supremacy(1 goal =100) figure of +902 for the 44 games that were played, an average of +20.5 per game. This is a very small sample indeed, but if this favourable edge could be maintained, it would be very lucrative for spread bettors.
Anyone else using the RFO index?
This is an excellent thread Artemis. One of the best I’ve seen on Football ratings. Keep up the good work. A couple of questions if that’s OK:
1. Are the RFO ratings avilable online. It can be difficult to get the paper in the sticks in Ireland.
2. Any update on how the Alex Deacon’s "Trends" figure is arrived at.
Artemis,
yes I’m still plugging away. I tweaked it to suit me further over the summer.
I’ve also started compiling my own ratings for the four English leagues – which help me understand Deacon’s own.
I have found the RFO’s ‘true odds’ page can be a useful ‘check’ to filter out any potentially rogue ratings.
It’s been encouraging so far – 90 bets, 44 wins – 115% returns. I’ll let you know if that still holds good come the end of the season!<br>
davidbrady,
I’m afraid the ratings are not available online- it would make life easier if they were.
From time to time, Alex (Deacon) devotes his column to the Trend figures. Although he hasn’t divulged his exact formula for assessing the Trend figure, I think you can work it out by looking at the changes from week to week. I did a similar exercise last season and found that each week’s performance affected the Trend figure by between – 4 and +4 depending on the result and the strength of opposition. ÂÂÂ
tooting,
Good to hear you are still beavering away at it.
(Edited by Artemis at 9:46 am on Dec. 20, 2005)
The formula returned an aggregate goal superiority figure of +216 (1 goal =100) for the 49 pools matches this week, an average of +4.4 per game. Although this is a very small sample, it is quite encoraging. It will be pointless doing much with the ratings over the next week because in this week’s RFO, they have not been updated to take account of the Boxing Day results.
- You must be logged in to reply to this topic.