Jump to content
** April Poker League Result : 1st Like2Fish, 2nd McG, 3rd andybell666 **
Download the PL Forum App!
** ELO Ratings are now back **

MadnessTrebles

madnesstiger
 Share

Recommended Posts

madnesstiger

madnesstiger

Posted
Posted

This system is a tweaked version of "Fleet's Trebles" as found elsewhere on the forum. Reading his threads inspired me to construct my own version with new ratings and a variation on the staking strategy. Overview I give each fixture (from the leagues covered) a rating, based on the home teams recent performance. The teams which achieve the best ratings along with suitable odds will be placed into a 3-Fold accumulator with Bet365. The bets will be placed when there is a busy fixture period, this will usually mean every weekend and also some midweek bets, as I feel necessary. Leagues/Divisions Covered

  • English Premier League
  • English Championship
  • English League One
  • English League Two
  • Scottish Premier League
  • Scottish Division One
  • Scottish Division Two
  • Scottish Division Three
  • Turkish Superlig
The Ratings The ratings are calculated using Microsoft Excel, and are based on a complex formula which takes into account:
  • Home team goals conceded in recent games
  • Away team goals conceded in recent games
  • Both teams' number of wins draws and losses in recent games
  • Winning margins in recent games

Recent games = last 6, except in the Scottish divisions where better results were obtained using the last 5 games. Basically, the higher rating a game has, the more chance (theoretically) the home team has of winning the match. The highest rating noted from collecting past data was 49. Previous Results Analysis Results data has been collected for each division covered in the system, and a match rating given where possible since the 2001/02 season. Altogether, almost 13,500 matches have been rated and analysed. Each divisions results produce a graph showing how well the ratings system performs and is given a trendline showing the line of best fit. The graph is then set to display its R-squared equation - this shows how 'reliable' the system is, where a figure of 1 is 100% consistent with the line of best fit. From this I have produced a comparison chart showing which divisions provide the most reliable system, which have better home-win strike rates within specific groups of ratings, and also how successful the system is at picking 'home team not-to-lose' selections. For example: The English Premier League gives us the most reliable data, with a r-squared rating of 96.2% and the 2nd highest 'slope' rating of 1.07 (this rating shows how well a highly-rated match picks home winners against how well a low-rated match picks an unsuccessful home team) It also shows that in the Scottish Division 3, 93% of matches rated between 21-30 (83 matches) resulted in a home win or a draw. Selections I find ratings for each upcoming match and will generally select the 3 teams with the highest rating to be included in the 3-fold bet. Where the ratings are close to call, I will use the comparison chart to find which disivison has provided more reliable/successful data for the rating in quesiton, as well as reading match previews for any injury doubts to key players or suspensions etc. to help decide. The bookies odds also can come into play here. If the odds for a match are too low (i.e. less than 1.50 in decimal odds) I will question the inclusion of the team in question. Odds of around 1.70-2.00 would be ideal and are what I had in mind whilst working out the staking strategy. I will be expecting most selected matches to be rated at least 20. Depending on the league in question, matches rated 21-30 have a strike rate of between 58% and 72%. Matches rated over 30 have strike rates of between 60% and 100%, again depending on the league in question. Staking Plan I liked FF's idea of using a 'progression' system made up of 10 levels, where if 10 successful trebles were selected in succession, you would be loaded! However, it didn't take long for me to realise that this pretty much required selecting 30 successful home wins in a row to complete all 10 levels. With this in mind, I am going to suggest the following staking strategy initially, based on the English FA Cup for a bit of fun!... Level 1 (1st Qualifying Round) £10 stake on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return £49 Bank the £10 stake, and proceed to level 2 with the £39 profit. Level 2 (2nd Qualifying Round) 2 attempts at another treble! £19.50 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £96.00 If win, we bank £20 and proceed to level 3 with approx £76. Level 3 (1st Round Proper) 2 attempts at another treble! £38.00 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £186.00 If win, we bank £30 and proceed to level 4 with approx £156. Level 4 (2nd Round) 2 attempts at another treble! £78.00 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £382.00 If win, we bank £40 and proceed to level 5 with approx £342. Level 5 (3rd Round) 2 attempts at another treble! £171.00 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £838.00 If win, we bank £50 and proceed to level 6 with approx £788. Level 6 (4th Round) 2 attempts at another treble! £394.00 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £1930.00 If win, we bank £60 and proceed to level 7 with approx £1870. Level 7 (5th Round) 2 attempts at another treble! £935.00 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £4580.00 If win, we bank £70 and proceed to level 8 with approx £4510. Level 8 (6th Round) 2 attempts at another treble! £2255.00 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £11050. If win, we bank £80 and proceed to level 9 with approx £10,970. Level 9 (Quarter-Final) 2 attempts at another treble! £5485.00 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £26,876 If win, we bank £90 and proceed to level 10 with approx £26,786. Level 10 (Semi-Final) 2 attempts at another treble! £13393.00 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £65,625 If win, we bank £100 and proceed to level 11 with approx £65525. Level 11 (The Final) 2 attempts at another treble! £32760.00 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £160,524 If win, we either bank the lot.. or bank £110 and go on to play in the Super Cup! Level 12 (Super Cup Final) 2 attempts at another treble! £80200.00 stake (half of profit from previous level) on 3-Fold at odds of approx 4.90 (based on average home-win odds of 1.70) A win would return approx £392,980 If win, we retire! Depending how many '2nd attempts' you had to use, you could walk away with anything upto around £500,000!!! This includes unused stakes and the small amounts banked on the way. You can obviously walk away at whichever level you want, but if you can conquer the lot, I'd recommend you congratulate yourself, have a pint, buy a Reliant Robin, and walk away from gambling for a little while! I'm going straight in with a £100 starting bankroll, real money.. see where we end up! I may work on alternative staking strategies in the future, but for now, will stick to the above. Suggestions, comments.. your shout folks! :D

Link to comment
Share on other sites
madnesstiger

madnesstiger

Posted
  • Author
Posted

Re: MadnessTrebles Let's get the ball rolling then... The three matches chosen for this weekends treble are as follows: EPL: LIVERPOOL vs Sunderland [17:15] Match Rating: 21 Home Odds: 1.30 (Bet365) Sunderland without an away win all season in Premier League, and only 2 away draws. Unlikely to improve at Anfield. ENG LGE 1: BRISTOL ROVERS vs Port Vale [15:00] Match Rating: 19 Home Odds: 1.72 (Bet365) Bristol unbeaten of late, against a Port Vale squad with leaky defence away from home. Out of the 3, I think this one is most risky. SCO DIV 2: AIRDRIE vs Raith Rovers [15:00] Match Rating: 20 Home Odds: 2.25 (Bet365) Airdrie have best home record in the division, and in good recent form. Raith have good away form but have a few injuries and are missing a key midfielder through suspension. Staking Level: 1 Initial Bankroll: £100.00 Stake: £10.00 New Bankroll: £90.00 Potential Returns: £50.31 Sorry these are a little late as I've been wanting to get everything in place ready for this weekends fixture list. One more note... I may find some weeks that a team has a high ratings, but also quite high odds, in which case I could afford to bet on the win OR draw and still be able to have a 3-fold with combined odds of around 4.9. :hope:hope:hope

Link to comment
Share on other sites
madnesstiger

madnesstiger

Posted
  • Author
Posted

Re: MadnessTrebles Other high ratings for this weekends, if interested: Odds in brackets (Bet365) Leyton Orient: 20 (2.30) Rochdale: 20 (2.15) East Fife: 24 (Real low odds, 1.20)) Stranraer: 21 (1.44) Stenhousmuir: 21 (2.25) Fairly poor week in terms of ratings, expected a few more higher ratings amongst these. Also, no eligible matches in Turkey as they are playing cup games over there this weekend. Turkey's results have gone down well with this system and have produced quite a reliable graph.

Link to comment
Share on other sites
kevsul

kevsul

Posted
Posted

Re: MadnessTrebles Excellent write up Madnesstiger. I will be watching with intrest, as I have followed fleet for some time now, Will have a little flutter with you today on these to see how you go, May combine your other teams in with the 3 to do 4 or 5 folds. Wish you the best of Luck with this,:ok You have taken a lot of effort produce this so stick with it

Link to comment
Share on other sites
madnesstiger

madnesstiger

Posted
  • Author
Posted

Re: MadnessTrebles Yep... Liverpool... mmm... the girlfriend's team... fingers crossed on this one then! Just checked other scores... hope you didn't combine alot of the other teams I listed, I think only Stranraer did the business, and East Fife with their minute odds. Just watched the Spurs v Manchester Reds match online - being a Spurs fan, I thought it was going to be my lucky day, only for the latest equaliser known to man! Bloody hate 'em :@!!

Link to comment
Share on other sites
Grex

Grex

Posted
Posted

Re: MadnessTrebles Well done with the first 2 of your treble.... :hope for Liverpool. By the way.....

Just watched the Spurs v Scummers match online - being a Spurs fan' date= I thought it was going to be my lucky day, only for the latest equaliser known to man! Bloody hate 'em :@!!
..... I should avoid using derogatory names against other teams - you might offend people (e.g. myself, being a Man U fan) :spank
Link to comment
Share on other sites
madnesstiger

madnesstiger

Posted
  • Author
Posted

Re: MadnessTrebles Thanks guys. Have changed the nasty name for you.. sorry! Back to business... todays results: Level 1 Treble: Airdrie 3:0 Raith Rovers :D Bristol Rovers 3:2 Port Vale :D Liverpool 3:0 Sunderland :D Current Financial Situation: Initial Bankroll: £100.00 Stake This Bet: £10.00 Returns This Bet: £49.97 New Bankroll: £139.97 Overall Profit/Loss: +40.0% Next Bet: Attempt 1 at Level 2 Stake: £19.99 If anyone joined me with todays treble, well done, and may it be the first of many. Todays win has give me some idea as to what I have ahead of me to make this system work. Todays Notes:

  • Obviously pleased at the win at the first attempt
  • Reading match previews does help pick the selections when unsure - The Airdrie selection was a great example, with them having the same match rating as Leyton Orient (who lost incidentally) but seemed to have much more in their favour against todays opponents in terms of players available, injuries, suspensions, etc.
  • Limiting the bets at trebles rather than going for 4 or 5-folds is advisable - even if the extra teams seem certain to win (an example would be East Fife today, who only just scraped a 1-0 win when they were priced as low as 1.20)

:ok

Link to comment
Share on other sites
madnesstiger

madnesstiger

Posted
  • Author
Posted

Re: MadnessTrebles Thanks again guys, your support really will help when we come to losing weeks (should there be any!), and hopefully you can give opinions on where I may be able to improve or modify the strategy. I will try to find time to work more on the ratings formula over the next few days as there seems to be no more full fixture lists until next weekend, only a few league one/two games Tuesday I think, which I will probably leave alone. I will try tweaking copies the spreadsheets - basically I will fiddle with the scoring 'preferences' to see if I can achieve better results. I will be looking mainly at the graphs and summarys of each division for the following:

  • Improvements on the R-squared equation (the reliability of the ratings)
  • The slope equation (the number that tells us how good the system is really - the higher the number the better)
  • If there is a constant occurrence of high-rated matches from each division throughout the season. For example, the Scottish Premier League may have a 100% success rate for ratings of over '25', but, there may have only been 10 of these such matches in the past 7 years - not enough really for what we require.

And who knows... I may even add a few more leagues if I think it will provide more opportunities. Note: To create the spreadsheets, I have used data from Joe Buchdahl's football-data.co.uk website which hosts downloadable data for many popular leagues in a handy spreadsheet format. If anyone knows of similar sources for the same sort of data, please let me know. I will probably be looking for something over the summer months such as the Scandanavian leagues or South American maybe.

Link to comment
Share on other sites
gregprowt

gregprowt

Posted
Posted

Re: MadnessTrebles Just watched the Spurs v Manchester Reds match online - being a Spurs fan, I thought it was going to be my lucky day, only for the latest equaliser known to man! Bloody hate 'em :@!! Im with you there mate, love my football, I will watch any (non league etc), but Man Utd hate them to the core, Im a Saints fan and dont even dislike Pompey as much as Man U, think its more to do with being bought up in the south and being surrounded by so many 'genuine fans' great stuff on getting the treble up :ok

Link to comment
Share on other sites
madnesstiger

madnesstiger

Posted
  • Author
Posted

Re: MadnessTrebles Yeah, the first big match I ever went to see was Sheffield Weds vs Tottenham in 1989, as I grew up in South Yorkshire. At the time Spurs had Gary Lineker, Paul Gascoigne, Gary Mabbutt etc, and I just thought 'this lot are better than the local team!' and stuck with them ever since. Anyway, 'barrhead' Thankyou for your comment and thanks for joining me with the treble (wonder how many did actually play those picks - looks like there were more than I expected!). As for the staking system, after reding through Fleet's threads I thought about his staking system and thought it was a little too agressive and unrealistic, even for me! As I have mentioned elsewhere, to get the jackpot meant winning 10 trebles in a row (30 matches!). I engineered my plan so that you got 2 attempts at each level rather than just the one, and if you run out of attempts at, say, level 5, the following would occur, and you would not necessarily go back to level 1:

  • If you passed level 4 at the first attempt, you would still have 1 attempt left at level 4.
  • If it took 2 attempts at level 4, you would go back to the last level at which you had an 'attempt' left. And, if so be the case, you would go back to level 1.

Therefore, it should be a gradual step down in stakes after a loss, rather than simply starting again as it were. 1 way to reduce the risk even further could include having 3 attempts at the next level (but having more levels to reach the same goal).

Link to comment
Share on other sites
Ace123

Ace123

Posted
Posted

Re: MadnessTrebles madness, got a question for you mate about your ratings model. wot do you do with your last 6 wins draws and losses? for example if a team has 4 wins, 1 draw and 0 losses. how do you allocate points for them? do you just add on a factor of 5? or do you give 3 pts for each win and 1 pt for each draw etc.....?? have you considered taking out outliers when you look at your graph? in one case there was a match rating of 35 but only had one match outcome to that rating and resulted in a loss. so removing that lifted the Rsq from 75% to 90% in my model. Have you done something similar??

Link to comment
Share on other sites
madnesstiger

madnesstiger

Posted
  • Author
Posted

Re: MadnessTrebles Hi Ace, Ready for another big post?!.. here goes! I have been working on my ratings spreadsheets alot since the weekend and have satisfied the following:

  • Covering all 9 divisions, I have improved the average 'slope' rating from 83.0% to 85.1% (not much, but an improvement!)
  • Covering all 9 divisions, I have improved the average 'R-sq' rating from 82.1% to 88.3%
  • Total number of matches with match ratings which have a 70% or more strike rate has been improved from 304 to 463 (52% increase) - this means there should be more selections to choose from each weekend/midweek.

Now for the geeky bit.. the ratings formula I am using at present (sorry if I confuse you anywhere from now on!): Here's the main formula for the match rating: =((S+F)-R)+I Key: S = (Goals conceded by away team in last 6 away matches)-(goals conceded by home team in last 6 home matches) F = (Wins - Draws - Losses) of last 'x' home matches of home team R = (Wins - Draws - Losses) of last 'y' away matches of away team I = ('Form Points' of last 3 home matches of home team)+('Form Points' of last 3 away matches of home team) FORM POINTS = These are calculated upon the winning margin/draw/loss of the team in question, as follows: Home Win By 1 Goal: 0 points Home Win By 2 Goals: 3 points Home Win By 3+ Goals: 4 points Home Defeat: -1 points Away Win By 1 Goal: 2 points Away Win By 2 Goals: 5 points Away Win By 3+ Goals: 7 points Away Defeat: 0 points Away Draw: 1 point Values 'x' & 'y' vary from division to division as required to give the best results for the system. However, these are always a figure between 4 and 8 (inclusive). I found that the points scoring above gave good all-round results for all divisions. It was by altering the 'x' and 'y' values that I tweaked the results so that I had optimum graph readings, and a larger range of matches that fell in the higher ratings (i.e. 21+, as these generally have a 70% strike rate or better). To try answering your question about outliers, I decided to group the ratings in sets of 5 (e.g. 36+, 31-35, 26-30, 21-25, and so on). This way I usually achieve over 10 results in each ratings group in each division. And like you said, if a single match had a really high rating of, say 50, and that game was a shock win for the away side, it would look really bad on the graph, meaning a load of 75%+ ratings suddenly had a big fat 0% in the middle of them! Hope this has helped answer your query. Tried my best to keep it as simple as possible! The sample range for each division ranges from the 2001/2002 season onwards (between approx 700 games for the Scottich leagues and 2500+ for the English divisions). As a side-note... tonights results rated as follows (home team ratings in brackets): Bristol Rovers (16) 2:0 Cheltenham Macclesfield (-3) 1:2 Lincoln City Bury (-16) 2:3 Stockport All 3 pretty much went the way of the ratings ;) MT

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
 Share
Go to topic listing
×
×
  • Create New...