- This topic has 78 replies, 10 voices, and was last updated 20 years, 4 months ago by
.
- Posts
Further analysis of goal supremacy(for this weekend’s games) gave a total figure of +698 for 55 games. The higher the (+) figure for goal supremacy, the better the ratings have performed. The totals so far for 76 games(still a very small sample to draw any conclusions from) is +408, an average of +5 per game. If this figure stays positive in the long term, it means the ratings are likely to produce a steady stream of value bets and a small profit. 5 per cent on turnover is the target.
The average stake is about 40pts, and after about 45 bets (1,800pts) the profit is 74pts, a little short of the target profit of 90pts.
The ratings for tonight’s games indicate that there are a number of value bets:<br>              ratings odds    actual odds<br>Cardiff(+9)       6/4           2/1
Coventry(+3)     13/8          11/4
Preston(+38)     6/5            9/4
Brentford(+46)    1/1            8/5
Elgin City (+27)    5/4           5/2
<br>Recommended bets:
LAY Ipswich to lose 50pts@2.54(Betfair)<br>Ipswich won 1-0 -50pts<br>LAY Millwall to lose 50pts@2.78(Betfair)<br>Draw 1-1 +27pts
Preston to win 20pts@9/4 (Tote)<br>Draw 1-1 -20pts<br>Brentford to win 30pts@8/5(Tote)<br>won 2-0 +48pts
Elgin to win 15pts@5/2, generally available.<br>Draw1-1 -15pts
<br>165 pts at risk. lose 10pts<br>The bank is now 1,064 pts(profit 64pts)
(Edited by Artemis at 10:20 am on Mar. 16, 2005)
Two value bets tonight, judged by the ratings.
Notts County (+12) should be 13/8, but are currently on offer at 5/2. With a (+) rating of below 30, I’m inclined to keep the draw on my side and lay Boston Utd.<br> Boston Utd won 4-0<br>The encounter between Nottingham Forest and Leeds is level at 0, and looks a fair bet for a draw at 12/5.<br> Draw 0-0<br>The live TV game between Liverpool(+3) and Blackburn may also end up level, but doesn’t look as good a bet as the Nottingham/Leeds game.<br> Draw 0-0<br>Recommended:
LAY Boston Utd to lose 50pts@2.26 (Betfair)<br> -50pts<br>Nottigham Forest and Leeds Utd to draw 15pts@12/5, best odds available +36pts
65pts at risk.<br>A loss of 14pts. Bank now 1,050pts(50pts profit)
(Edited by Artemis at 9:15 am on Mar. 17, 2005)
The following weekend games are short-listed as possible bets:
ratings odds actual odds
Gillingham(+15) 11/8 2/1<br>Burnley(+4) 13/8 10/3!<br>Sheff Wed(+56) 5/6 6/5<br>Wycombe(+67) 4/5 13/10<br>Southend(+34) 5/4 7/4<br>Notts Co(+24) 6/5 7/4<br>Ross County(+24) 11/8 12/5 <br>
hi there artemis. great thread. <br>i would like to do some system backtesting on the rfo index, and wondered if you had the indices for each game.? they are unavailable on line (unlike scores, shots, corners etc)
Hi muppet,
I’m afraid I don’t keep a record of the ratings except to update the profit/loss figures and monitor the goal supremacies as a fair measure of the accuracy of the ratings. As far as I know, the ratings are no longer available on line.
(Edited by Artemis at 6:42 pm on Mar. 18, 2005)
Recommended Bets:
LAY Ipswich to lose 50pts@2.25(Betfair)<br> Draw 0-0 +38pts<br>LAY Sheff Utd to lose 50pts@1.79(Betfair)<br> Sheff utd won 2-1 -50pts<br>Wycombe to win, 40pts@ 13/10(Bet365)<br> Lost 2-1 -40pts<br>Southend to win, 30pts@7/4 (Ladbrokes)<br> won 3-0 +52pts
Notts County to win, 25pts@9/5 (Bet Direct)<br> lost 3-2 -25pts<br>Ross County to win, 20pts@12/5(Skybet)<br> won 2-0 +48pts
A profit of 23 points<br>The bank now stands at 1,073points(73pts profit)
(Edited by Artemis at 6:52 pm on Mar. 19, 2005)
I’ve looked at the total goal supremacies for Tuesday to Saturday’s games in terms of (+) or (-) to see how the ratings have performed in comparison with the predicted outcome. The higher the (+) figure for goal supremacy, the better the ratings have performed.
The aggregate figure for 71 games was -1,041.
After analysing a total of 147 games, the aggregate figure is -632, an average of -4 pts per game. This is still a very small sample, so no conclusions can yet be drawn. In the long run, an average figure of close to zero would be strong evidence that the formula is reliable.
A handful of midweek games and a decent selection of lower league games on Friday and Saturday. I can only find two games where the ratings suggest that the actual odds may not be right.
In tonight’s match, Stenhousemuir(+13)(=13/8) are a stand out 11/5 with Ladbrokes, but the general price is 7/4. I would normally lay the opposition where the (+) figure is below 30, but in this case the betting exchanges are unlikely to be out of line.
Recommended bet:
Stenhousemuir 20pts win@11/5 (Ladbrokes).<br> won 2-0 +44pts<br>The bank now stands at 1,117pts(117pts profit)<br>On Friday, Macclesfield(+44)(=6/5) look attractive at 15/8 with Skybet.
Alex Deacon also produces ratings for International Teams based on competitive matches played. I’ve looked through Saturday’s games and the only value I can see is Romania(+2)(=13/8) at home to Holland. Skybet have them at 11/5- I might be tempted to lay the Dutchmen.
(Edited by Artemis at 8:17 am on Mar. 23, 2005)
Tomorrow’s recommended bet:
Macclesfield(+44)(=6/5), 25pts win@ 15/8, generally available.<br> lost 2-0 -25pts<br>Bank now stands at 1,092pts(92pts profit)<br>Romania lost 2-0 to Holland
(Edited by Artemis at 8:39 am on Mar. 27, 2005)
Last weekend’s aggregate goal supremacy total (actual goal supremacy – ratings prediction of goal supremacy) was -502 for 27 games. The total for 174 games is -1134, an average of -7 per game. After 500 games, which is a fair sample, I will possibly adjust the formula to reflect this.
There appear to be a number of possible bets for the coming weekend, although not all bookmakers have priced up the games at this stage.
ratings odds actual odds
Bolton(+7) 6/4 10/3
W. Brom(+26) 5/4 7/4
Nottingham<br>Forest(+21) 11/8 7/4
Derby(+13) 6/4 3/1
MK Dons(+84) 8/11 6/5
Grimsby (+2) 13/8 5/2
Lincoln(+49) 1/1 11/8
Macclesfield(+30) 5/4 5/2
Oxford Utd(+16) 6/4 7/4
Kidderminster(+8) 6/4 7/2
<br>Recommended bets on Fri/Sat.
Recommended bets:
The following are all LAYS to lose 50pts (Betfair odds at 9.40 a.m.)
Liverpool 2.02<br>Liverpool won 1-0      -50pts<br>Everton 2.62<br>West Brom won 1-0 +29pts <br>Crewe 2.84 <br>Draw 1-1             +26pts<br>Ipswich 1.93<br>Ipswich won 3-2       -50pts<br>Mansfield  2.8<br>Mansfield won 2-0      -50pts<br>Shrewsbury 2.64<br>Shrewsbury won 3-0    -50pts<br>Other bets:
MK Dons to win 40pts@5/4  Skybet   <br>Drew 1-1             -40pts<br>Lincoln to win 30pts@6/4 Stan James<br>Lost 1-0              -30pts<br>Macclesfield to win 20pts@5/2 Skybet<br>Drew 0-0             -20pts<br>Kidderminster to win 15 pts@7/2 Skybet, Stan James<br>Lost 3-0              -15pts
A loss of 250pts over the weekend. The bank now stands at 842pts (158pts loss). 8 losing bets out of 10 can be disappointing, but statistically it is likely to occur more often than most people might expect.[/ color]
(Edited by Artemis at 6:41 pm on April 2, 2005)<br>
(Edited by Artemis at 7:24 pm on April 3, 2005)
The goal supremacy total for the 56 games rated last week was -1,730. The aggregate total for 230 games is now -2,864, an average of -12 per game. Although I would like to have looked at the totals and averages after 500 games, I think there is fairly clear evidence that the ratings formula needs an adjustment. It was suggested earlier on the thread(by tooting) that the trend figure(T) was set too high and I am now inclined to agree with this in view of the evidence from goal superiority.
I am now adjusting the formula so that the trend is multiplied by a factor of 3 rather than 5. The net effect will be that the effect of the trend(T) will be reduced by 40%.
Revised Formula:
Final Rating = [M + H(or A) +(Tx3)]/2
(Edited by Artemis at 9:42 am on April 4, 2005)
Hi Artemis,
I’ll be interested how the revised formula works for rest of season.
I’ve been playing with x2 Trends, and since I changed I’ve come out just slightly in front. This way gives fewer bets (5 this weekend).
Whether profitable or not it is a fantastic means of providing rough and ready prices for loads of matches, and prices that are really amazingly accurate compared to those available. For which I remain genuinely grateful.
One thing that stands out from my variant formula is that nearly all the bets are for away teams (or lays of home teams), often against teams with strong home form. I’m guessing that the master rating is modifying the home form more than it is the away form – in particular for teams who are virtually impregnable at home.
I think I may look for a further modifier based around teams with overly large differences between home and away ratings.
tooting,
If you only use the Master ratings, you will find that on average the advantage enjoyed by home teams is between 30 and 40 points(0.3-0.4 goal). This has been shown quite clearly by Alex Deacon’s analysis and is also confirmed by Kevin Pullein in the Racing Post. I think using the formula tries to look beyond these average figures, but there is always the possibility that it is underestimating the importance of home advantage.<br>Although I have adjusted the Trend figure, I am aware that the other two variables could be the reason the goal superiority figures were showing a negative figure. At the end of this ‘formula’ approach, we will probably have to conclude that the experts don’t make very many mistakes pricing up games.
The only price out of line this evening is Sunderland(+36) who are generally 7/4 to beat Wigan. The ratings suggest they should be around 5/4.
Recommended bet:
Sunderland to win 20pts@7/4 general(15/8 Stan James)<br>Sunderland won 1-0     +35pts. The bank is now 877pts(loss 123pts)
<br>(Edited by Artemis at 11:13 am on April 6, 2005)<br>
(Edited by Artemis at 12:01 pm on April 7, 2005)
The formula adjustment has (as expected) brought the ratings odds closer to the bookmaker’s prices, so bets are going to be more difficult to find. This weekend’s games of interest are:
Friday<br>Cheltenham(0) worth considering at 12/5 for the draw
Saturday<br>Leeds(+30= 5/4) on offer at 13/8
Colchester(+19=11/8) on offer at 11/8,- a Lay of Hartlepool
MK Dons(+34=5/4) available at 7/4. They keep cropping up as value, but have let me down a few times.
Lincoln(0=13/8) worth considering for a draw at 9/4
Notts Co(+1=13/8) on offer at 5/2 look the best value bet of the day.
Clyde (+48=1/1), look value at 6/4.
Recommended bets Friday/Saturday.
Weekend’s bets:
Cheltenham to DRAW with Shrewsbury 20pts@12/5 generally available.<br>Draw 1-1 +48pts<br>Leeds to beat Watford 30pts@7/4 (Stan James, Tote)<br>Leeds won 2-1 +52pts<br>LAY Hartlepool to lose 50pts@2.72(Betfair)<br>Draw 1-1 +28pts<br>MK Dons to beat Torquay 30pts@7/4, generally available<br>Torquay won 1-0 -30pts<br>Lincoln to DRAW with Swansea 20pts@9/4, gen av.<br>Lincoln won 1-0 -20pts<br>Notts County to beat Oxford 20pts@5/2,gen av<br>Oxford won 2-1 -20pts<br>170pts at risk.  <br>A profit of 58pts. The bank is now 935 pts( loss 65pts.<br>I’m now up to 70 bets at an average stake of around 40pts(Turnover, 2,800pts) and showing a loss of about 2 percent against a target profit of 5 per cent(140pts). I haven’t analysed the goal supremacy figure for the week’s matches yet.
(Edited by Artemis at 6:56 pm on April 9, 2005)
- You must be logged in to reply to this topic.