ELO & Poisson rating system

[giraldi]

I have some experience on betting and stats and I would like to present you my strategy based on my own rating system. Description of the system I have at my disposal 35 European leagues with results for at least 5 seasons. Unfortunately I don-t have the odds but for 12 main leagues England 1,2,3,4 Italy Spain France and Germany first 2 leagues for each. Each team started with an value of 2000 points 3 seasons ago and depending on their results, according an ELO formula (a little bit modified) they won or lost points. Than for a given game I calculate the probabilities. This probability is than affected by the probability calculated with poison formula. for the past few games at home respectively away. In my opinion this correction will highlight recent form. After that I apply few more correction to take into account the type of the game for example - big local derby, ...etc I work to improve this correction to take into account more and more facts. For example I noticed that the games rescheduled have a different pattern than normal games and so on The final step is to compare my probabilities with the odds given by pinnacle to study a little bit head to head the teams to see what is about ( I will never pass to robots my money :) ) Next step is to include the real facts as - missing offensive or defensive good players -will affect with ? percent and so on I tested for my self this system and the results are encouraging. Last month (November) I tested on another forum with yield + 15% and 105 units invested. I can easy prove that. Staking plan Initial bank=100 units 1 unit for each bet. Last month were around 10 bets / day on weekend and maximum 1-2 on the other days I do not recommend to play real money yet I am open to any discussion about my rating. I do not believe in Santa Claus. I know I am not he first who try this but my results are as I have said encouraging. English is not my native language. I have already use all the English words I know :). I will try to show you good bets and good results.

[chikilicuatre]

Re: ELO & Poisson rating system Good morning, can I ask what is your native language? I'm Spanish. Thanks and Luck

[Datapunter]

Re: ELO & Poisson rating system

- same 2000 points to start (regardless of strength at that moment) - starting 3 seasons ago (no matter half the players are now different) - results, according an ELO formula (a measurement created for individuals who remain the same person, applied to teams that change all the time) - probability is than affected by the probability calculated with poison formula (???????) - for the past few games at home respectively away. (i assume disregarding the strength of the opponant) - few more correction to take into account the type of the game (??????????) - more and more facts, rescheduled have a different pattern than normal games and so on (makes all the previous obsolete) - to study a little bit head to head the teams to see what is about ( as subjective as it gets and also makes all the previous obsolete ) - Next step is to include the real facts as - missing offensive or defensive good players -will affect with ? percent and so on ( see previous )

Well, it appears to me you don't have a system, looks more like chaos. If the results are good you'll never know why as you have no idea what works and what doesn't and more importantly the subjective elements in above list mean your 'system' isn't reproducable. If the results are bad you'll have no idea where to adjust. Anyhow, just my 2 cents, i wish you good luck :ok

[giraldi]

Re: ELO & Poisson rating system Thank you Datapunter for your comments. I am very impressed that you payed such attention to a new user but, on the other hand I am sure that my poor English is a handicap for me in explaining my system so you think that is chaos here. I don-t want to expose a lot of explanation but I assure you that I know very well what is ELO and Poisson. You are right, ELO was created for 2 players game but the name ELO stands for any rating system based on winning or loosing according with opponent strength. FIFA, UEFA and many others use this kind of system . I am sure you know. To calibrate such a system you need at least 30-40 games that is the reason I choose to start with calculation 3 years ago. All the team in Premier started with 2000 points, Championship with 1700..and so on. For promoted or relegated teams they maintained their rating..etc. Never the teams will be the same, of course, but the team rating is "alive". The team changes but the relative value is the same. If the strength is rising the rate will rising after few good games so doesn-t matter the players name. The final result for a game is given by the relative strength between opponents and not by the opponent strength. The 1-0 is the same between City and Utd or between 2 League Two teams. Only the difference is important. About Poisson. On Pinnacle site you can find an article about Poisson and how they calculate the probabilities (of course without any in deep suggestions). I choose to calculate probabilities for goals scored with Poisson formula considering last games for each team. A team with may goals scored at home on the last 5 games is likely to score again. Check the stats. So ..the probabilities given by the ELO difference and the probabilities given by the last scored games (to highlight recent form) are combined with different percents to result the final probabilities. You will be surprised what close I am to bookies probabilities for the majority of games. There are few games where the probabilities are some differences (sometimes big). Those games are my target. I have already said, I have some experience on betting. I tested all the probabilities against the real odds past 2 seasons and I choose only patterns that were with profit last 2 seasons and this season. For example - Segunda Division with the probabilities over 33% (according my rating) for draw was with profit last years and this season. This doesn-t mean that will be profitable for ever but worth to try. Of course I have a filter and don-t play whatever the rating present. I study the game head to head and than make my decision. Good to play or not. I think is common sense to filter all the games proposed based on my experience. And finally I want to assure you that I really know what to adjust. If you think that I use a paper and a pencil (or EXCEL) to calculate you are wrong. I can wait for Saturday (just because is a new month) to start my predictions. I wish you good luck too and even started with your 2 cents (I hope I understood ) bet on me, but be patient and give me an entire month .

[chikilicuatre]

Re: ELO & Poisson rating system Goodnight My English is not perfect, but I think from what little we understand, and google translator, we will be bringing something forward. If I can help you, count on me. Greetings.

[giraldi]

Re: ELO & Poisson rating system Thank you chikilicuatre for your support. I really don-t think that here are users that judges a book by its cover so I hope the not perfect English is not a big obstacle to see the core. I can not wait until 1st December. I will start my prediction based on my rating system today. The results of course will count for December (only if are winning :) ) As I said the bank is 100 units, I play 1 unit for each bet, Pinnacle (if not, I will write the bookmaker) For the first month the main objective is to not loose entire bank :) So, let-s unleash the.."chaos". I also present the probabilities given by my rating system for the event. 1. Ita1. Catania - Milan Over 2.5 @ 2.06 (49.65%) 2. Spa1. Osasuna - Vallecano Under 2.5 @ 1.99 (62.17%) 3. Spa2. Murcia - EM Castilla Under 2.5 @ 2.25 (59,63%) 4. Fra1. Ajjacio - St. Etienne Draw @ 3.19 (35.25%) 5. Fra2. Monaco - Sedan Draw @ 4.06 (34.26%) 6. Fra2. Arles - Le Mans Over 2.5 @ 2.64 (50.10%) In my opinion good probabilities and good prices. At least 2 good bets won (I would prefer the draws) should be enough for the beginning.

[Datapunter]

Re: ELO & Poisson rating system http://en.wikipedia.org/wiki/My_two_cents :ok

[robotzel78]

Re: ELO & Poisson rating system Welcome giraldi, and i hope to see a very good system here. I'll be watching you...

[chikilicuatre]

Re: ELO & Poisson rating system Good Luck¡¡¡

[giraldi]

Re: ELO & Poisson rating system Datapunter, Roger that ! Robotel78, chikilicuatre Thank you !

1. Ita1. Catania - Milan Over 2.5 @ 2.06 (49.65%) 2. Spa1. Osasuna - Vallecano Under 2.5 @ 1.99 (62.17%) 3. Spa2. Murcia - EM Castilla Under 2.5 @ 2.25 (59,63%) 4. Fra1. Ajjacio - St. Etienne Draw @ 3.19 (35.25%) 5. Fra2. Monaco - Sedan Draw @ 4.06 (34.26%) 6. Fra2. Arles - Le Mans Over 2.5 @ 2.64 (50.10%)

Good start yesterday: 4 won out of 6 and +3.49 units. For today I will play another 11 bets with good probabilities good prices and most important with value (according with my rating - don-t blame me :)) 7. Eng2. Huddersfield - Leeds Over 2.5 @ 1.93 (54.75%) 8. Ger2. Union - Bochum Under 2.5 @ 2.23 (61.08%) 9. Ger1. Augsburg - Freiburg Freiburg @ 2.79 (43.95%) 10. Ger1. Furth - Stutgart Draw @ 3.52 (31.57%) 11. Ger1. Furth - Stutgart Under 2.5 @ 1.97 (61.91%) 12. Eng2. Bristol - Wolves Under 2.5 @ 2.10 (77.74%) 13. Spa2. Gijon - Las Palmas Draw @ 3.42 (32.47%) 14. Spa2. Lugo - Barcelona B Draw @ 3.54 (32.68%) 15. Spa2. Cordoba - Xerez Under 2.5 @ 1.90 (60.37%) 16. Fra1. Evian - Nancy Over 2.5 @ 2.10 (50.08%) 17. Ger1. Bayern - Dortmund Under 2.5 @ 2.39 (59.32%)

[analyzer]

Re: ELO & Poisson rating system Hi giraldi. If I understood you correctly you have historical odds for 12 leagues. Have you tried to test your system on historical data. It would be interesting to know what are the results after thousands of games.

[giraldi]

Re: ELO & Poisson rating system Hi, analyzer. Of course I tried. For example Step 1. I tested my system against played games this season and write down what leagues, type of bet (Home win, Draw, Away win, Under, Over 2.5), probability given by my system (over 55% for example), odds given by bookmakers (odds over 2.10 for example) were profitable If I had played those let-s day patterns. So the patterns looks like this: Italy 1, Home Win, probability over 55%, odds over 2.10 . Just an example. For this pattern the profit had been 15% for example this season. For all 12 leagues I obtained for example 18 patterns for current season. Step 2. Each pattern is tested against real odds for last 2 seasons. Only pattern with profit each season will go further. Remains only 8 or 9 This is just an example but I did this and I have a set of patterns where I choose from This does not mean that the patterns will be winning for the rest of current season but there are great chances in my opinion. I tested "live" this system in November on big forum with good results 115 games played, 19 units profit Yield - 16.5% (I may prove this but I don-t want to put link here - only private for moderators if insist) I would like to present my system here just because here are very good punters and very good theoreticians on the same time. I am willing to present my system details to improve this. Please don-t say that many good guys did it before. Contrary datapunter thinks the system is very easy adjustable but I have no reason so far because works fine for me. If something is going obviously very wrong I just change one of many variables (for example how many games behind the system consider when compute current form..etc.) for fine tuning and try again. On the other hand, I must say I created this system first just for help me to find at a glance good games to study. In 10 seconds I filter all the games and I have only 20 to look closely. Maybe 9-12 will remain good to play. It is a big advantage. Further, I admit that this point of view should be only the first step. One very good punter will consider last events (players..injuries..etc.) to make final decision. Personally skip this step and it is very wrong. But I am to lazy as long as it works so far. :) I need a long row of bad run to do that. (I hope not so soon but it is inevitable - the god of statistics tell me that :)) I think that my system is good I will prove that. Anyhow, just my 2 cents :D :ok

[giraldi]

Re: ELO & Poisson rating system

7. Eng2. Huddersfield - Leeds Over 2.5 @ 1.93 (54.75%) 8. Ger2. Union - Bochum Under 2.5 @ 2.23 (61.08%) 9. Ger1. Augsburg - Freiburg Freiburg @ 2.79 (43.95%) 10. Ger1. Furth - Stutgart Draw @ 3.52 (31.57%) 11. Ger1. Furth - Stutgart Under 2.5 @ 1.97 (61.91%) 12. Eng2. Bristol - Wolves Under 2.5 @ 2.10 (77.74%) 13. Spa2. Gijon - Las Palmas Draw @ 3.42 (32.47%) 14. Spa2. Lugo - Barcelona B Draw @ 3.54 (32.68%) 15. Spa2. Cordoba - Xerez Under 2.5 @ 1.90 (60.37%) 16. Fra1. Evian - Nancy Over 2.5 @ 2.10 (50.08%) 17. Ger1. Bayern - Dortmund Under 2.5 @ 2.39 (59.32%)

+0.61 units. Small profit but still profit. 17 units played, 4.1 units profit, yield 24.12% I hope I will find something good to play for tomorrow. Usually Sunday my rating give me 7-10 games to select from.

[giraldi]

Re: ELO & Poisson rating system Finally I will play 14 games today. Few surprises, with good prices so I will try with the hope the correct probabilities will work for me on long term. 12. Ita1. Udinese - Cagliari Udinese @ 2.19 (52.34%) 13. Ita1. Genoa - Chievo Draw @ 3.41 (50.48%) 14. Ita1. Siena - Roma Draw @ 3.97 (33.22%) 15. Ita1. Siena - Roma Under 2.5 @ 2.65 (63.23%) 16. Ita1. Lazio - Parma Over 2.5 @ 2.02 (54.78%) 17. Ita1. Napoli - Pescara Under 2.5 @ 2.38 (63.10%) 18. Spa1. Granada - Espanyol Espanyol @ 3.81 (44.30%) 19. Spa1. Granada - Espanyol Over 2.5 @ 2.29 (61.79%) 20. Spa1. Deportivo - Betis Betis @ 3.56 (34.95%) 21. Spa2. Elche - Villareal Elche @ 2.33 (52.02%) 22. Spa2. Alcoron - Girona Draw @ 3.40 (32.71%) 23. Fra1. Lorient - Toulouse Draw @ 3.26 (33.40%) 24. Fra1. Brest - Marseilles Draw @ 3.26 (34.48%) 25. Ger1. Wolfsburg - Hamburg Hamburg @ 4.14 (39.41%)

[giraldi]

Re: ELO & Poisson rating system

12. Ita1. Udinese - Cagliari Udinese @ 2.19 (52.34%) 13. Ita1. Genoa - Chievo Draw @ 3.41 (50.48%) 14. Ita1. Siena - Roma Draw @ 3.97 (33.22%) 15. Ita1. Siena - Roma Under 2.5 @ 2.65 (63.23%) 16. Ita1. Lazio - Parma Over 2.5 @ 2.02 (54.78%) 17. Ita1. Napoli - Pescara Under 2.5 @ 2.38 (63.10%) 18. Spa1. Granada - Espanyol Espanyol @ 3.81 (44.30%) 19. Spa1. Granada - Espanyol Over 2.5 @ 2.29 (61.79%) 20. Spa1. Deportivo - Betis Betis @ 3.56 (34.95%) 21. Spa2. Elche - Villareal Elche @ 2.33 (52.02%) 22. Spa2. Alcoron - Girona Draw @ 3.40 (32.71%) 23. Fra1. Lorient - Toulouse Draw @ 3.26 (33.40%) 24. Fra1. Brest - Marseilles Draw @ 3.26 (34.48%) 25. Ger1. Wolfsburg - Hamburg Hamburg @ 4.14 (39.41%)

Only 4 out of 14 but only 3.9 units lost. Could be worst but, easier, could be much more better. Anyhow if is true that on long term lucky/unlucky is 50/50 (I believe remains only skill), than I am ready to win some unexpected units next weeks :D. 31 played , +0.2 units (it is still profit :D ) let-s see if it is something good to play today.

[giraldi]

Re: ELO & Poisson rating system I will try today 32. Fra2. Guincamp - Istres Draw @ 3.21 (33.28%) 33. Ger2. Cottbus - Hertha Hertha @ 2.13 (46.87%) The value for the second game (if exists) is minimum. Home results for Cottbus are good but rating difference between the two teams is around 200 points. Hertha is Bundesliga I team since Cottbus started the season with poor rating, (now is 100 points higher) Maybe is interesting for you what are the ratings for all the teams - Let-s say for England 1,2,3,4 I remember that rating started on the season 2008/2009. After calibration the rating has pretty big inertia so only very good or very bad games changes it significantly. You can compare with your own ratings and I would be happy to talk about that. [TABLE=width: 322]

[TR] [TD=align: right]1[/TD] [TD]Eng1[/TD] [TD]Man City[/TD] [TD=align: right]2194[/TD] [/TR] [TR] [TD=align: right]2[/TD] [TD]Eng1[/TD] [TD]Man Utd[/TD] [TD=align: right]2189[/TD] [/TR] [TR] [TD=align: right]3[/TD] [TD]Eng1[/TD] [TD]Tottenham[/TD] [TD=align: right]2018[/TD] [/TR] [TR] [TD=align: right]4[/TD] [TD]Eng1[/TD] [TD]Everton[/TD] [TD=align: right]2005[/TD] [/TR] [TR] [TD=align: right]5[/TD] [TD]Eng1[/TD] [TD]Chelsea[/TD] [TD=align: right]1988[/TD] [/TR] [TR] [TD=align: right]6[/TD] [TD]Eng1[/TD] [TD]Arsenal[/TD] [TD=align: right]1986[/TD] [/TR] [TR] [TD=align: right]7[/TD] [TD]Eng1[/TD] [TD]Swansea[/TD] [TD=align: right]1940[/TD] [/TR] [TR] [TD=align: right]8[/TD] [TD]Eng1[/TD] [TD]Liverpool[/TD] [TD=align: right]1937[/TD] [/TR] [TR] [TD=align: right]9[/TD] [TD]Eng1[/TD] [TD]W.B.A.[/TD] [TD=align: right]1929[/TD] [/TR] [TR] [TD=align: right]10[/TD] [TD]Eng1[/TD] [TD]Stoke[/TD] [TD=align: right]1922[/TD] [/TR] [TR] [TD=align: right]11[/TD] [TD]Eng1[/TD] [TD]West Ham[/TD] [TD=align: right]1882[/TD] [/TR] [TR] [TD=align: right]12[/TD] [TD]Eng1[/TD] [TD]Norwich[/TD] [TD=align: right]1869[/TD] [/TR] [TR] [TD=align: right]13[/TD] [TD]Eng1[/TD] [TD]Fulham[/TD] [TD=align: right]1856[/TD] [/TR] [TR] [TD=align: right]14[/TD] [TD]Eng1[/TD] [TD]Newcastle[/TD] [TD=align: right]1849[/TD] [/TR] [TR] [TD=align: right]15[/TD] [TD]Eng1[/TD] [TD]Wigan[/TD] [TD=align: right]1844[/TD] [/TR] [TR] [TD=align: right]16[/TD] [TD]Eng1[/TD] [TD]Sunderland[/TD] [TD=align: right]1824[/TD] [/TR] [TR] [TD=align: right]17[/TD] [TD]Eng1[/TD] [TD]Aston Villa[/TD] [TD=align: right]1802[/TD] [/TR] [TR] [TD=align: right]18[/TD] [TD]Eng1[/TD] [TD]Reading[/TD] [TD=align: right]1781[/TD] [/TR] [TR] [TD=align: right]19[/TD] [TD]Eng1[/TD] [TD]Southampton[/TD] [TD=align: right]1773[/TD] [/TR] [TR] [TD=align: right]20[/TD] [TD]Eng2[/TD] [TD]Millwall[/TD] [TD=align: right]1758[/TD] [/TR] [TR] [TD=align: right]21[/TD] [TD]Eng2[/TD] [TD]Cardiff[/TD] [TD=align: right]1744[/TD] [/TR] [TR] [TD=align: right]22[/TD] [TD]Eng2[/TD] [TD]Leicester[/TD] [TD=align: right]1727[/TD] [/TR] [TR] [TD=align: right]23[/TD] [TD]Eng2[/TD] [TD]Crystal Palace[/TD] [TD=align: right]1726[/TD] [/TR] [TR] [TD=align: right]24[/TD] [TD]Eng2[/TD] [TD]Watford[/TD] [TD=align: right]1719[/TD] [/TR] [TR] [TD=align: right]25[/TD] [TD]Eng1[/TD] [TD]QPR[/TD] [TD=align: right]1700[/TD] [/TR] [TR] [TD=align: right]26[/TD] [TD]Eng2[/TD] [TD]Middlesbrough[/TD] [TD=align: right]1667[/TD] [/TR] [TR] [TD=align: right]27[/TD] [TD]Eng2[/TD] [TD]Hull[/TD] [TD=align: right]1660[/TD] [/TR] [TR] [TD=align: right]28[/TD] [TD]Eng2[/TD] [TD]Blackpool[/TD] [TD=align: right]1660[/TD] [/TR] [TR] [TD=align: right]29[/TD] [TD]Eng2[/TD] [TD]Bolton[/TD] [TD=align: right]1659[/TD] [/TR] [TR] [TD=align: right]30[/TD] [TD]Eng2[/TD] [TD]Charlton[/TD] [TD=align: right]1657[/TD] [/TR] [TR] [TD=align: right]31[/TD] [TD]Eng2[/TD] [TD]Brighton[/TD] [TD=align: right]1656[/TD] [/TR] [TR] [TD=align: right]32[/TD] [TD]Eng2[/TD] [TD]Blackburn[/TD] [TD=align: right]1653[/TD] [/TR] [TR] [TD=align: right]33[/TD] [TD]Eng2[/TD] [TD]Birmingham[/TD] [TD=align: right]1635[/TD] [/TR] [TR] [TD=align: right]34[/TD] [TD]Eng2[/TD] [TD]Leeds[/TD] [TD=align: right]1616[/TD] [/TR] [TR] [TD=align: right]35[/TD] [TD]Eng2[/TD] [TD]Nottingham Forest[/TD] [TD=align: right]1615[/TD] [/TR] [TR] [TD=align: right]36[/TD] [TD]Eng2[/TD] [TD]Derby[/TD] [TD=align: right]1607[/TD] [/TR] [TR] [TD=align: right]37[/TD] [TD]Eng2[/TD] [TD]Burnley[/TD] [TD=align: right]1604[/TD] [/TR] [TR] [TD=align: right]38[/TD] [TD]Eng2[/TD] [TD]Wolves[/TD] [TD=align: right]1601[/TD] [/TR] [TR] [TD=align: right]39[/TD] [TD]Eng3[/TD] [TD]MK Dons[/TD] [TD=align: right]1561[/TD] [/TR] [TR] [TD=align: right]40[/TD] [TD]Eng3[/TD] [TD]Sheffield Utd[/TD] [TD=align: right]1557[/TD] [/TR] [TR] [TD=align: right]41[/TD] [TD]Eng2[/TD] [TD]Ipswich[/TD] [TD=align: right]1554[/TD] [/TR] [TR] [TD=align: right]42[/TD] [TD]Eng2[/TD] [TD]Huddersfield[/TD] [TD=align: right]1537[/TD] [/TR] [TR] [TD=align: right]43[/TD] [TD]Eng3[/TD] [TD]Notts[/TD] [TD=align: right]1535[/TD] [/TR] [TR] [TD=align: right]44[/TD] [TD]Eng3[/TD] [TD]Brentford[/TD] [TD=align: right]1531[/TD] [/TR] [TR] [TD=align: right]45[/TD] [TD]Eng3[/TD] [TD]Doncaster[/TD] [TD=align: right]1529[/TD] [/TR] [TR] [TD=align: right]46[/TD] [TD]Eng3[/TD] [TD]Tranmere[/TD] [TD=align: right]1495[/TD] [/TR] [TR] [TD=align: right]47[/TD] [TD]Eng3[/TD] [TD]Swindon[/TD] [TD=align: right]1494[/TD] [/TR] [TR] [TD=align: right]48[/TD] [TD]Eng3[/TD] [TD]Bournemouth[/TD] [TD=align: right]1480[/TD] [/TR] [TR] [TD=align: right]49[/TD] [TD]Eng3[/TD] [TD]Stevenage[/TD] [TD=align: right]1470[/TD] [/TR] [TR] [TD=align: right]50[/TD] [TD]Eng2[/TD] [TD]Sheffield[/TD] [TD=align: right]1468[/TD] [/TR] [TR] [TD=align: right]51[/TD] [TD]Eng2[/TD] [TD]Bristol City[/TD] [TD=align: right]1451[/TD] [/TR] [TR] [TD=align: right]52[/TD] [TD]Eng3[/TD] [TD]Coventry City[/TD] [TD=align: right]1448[/TD] [/TR] [TR] [TD=align: right]53[/TD] [TD]Eng2[/TD] [TD]Barnsley[/TD] [TD=align: right]1441[/TD] [/TR] [TR] [TD=align: right]54[/TD] [TD]Eng2[/TD] [TD]Peterborough[/TD] [TD=align: right]1441[/TD] [/TR] [TR] [TD=align: right]55[/TD] [TD]Eng3[/TD] [TD]Preston[/TD] [TD=align: right]1426[/TD] [/TR] [TR] [TD=align: right]56[/TD] [TD]Eng3[/TD] [TD]Crewe[/TD] [TD=align: right]1424[/TD] [/TR] [TR] [TD=align: right]57[/TD] [TD]Eng4[/TD] [TD]Southend[/TD] [TD=align: right]1415[/TD] [/TR] [TR] [TD=align: right]58[/TD] [TD]Eng4[/TD] [TD]Port Vale[/TD] [TD=align: right]1413[/TD] [/TR] [TR] [TD=align: right]59[/TD] [TD]Eng4[/TD] [TD]Gillingham[/TD] [TD=align: right]1395[/TD] [/TR] [TR] [TD=align: right]60[/TD] [TD]Eng3[/TD] [TD]Leyton[/TD] [TD=align: right]1376[/TD] [/TR] [TR] [TD=align: right]61[/TD] [TD]Eng3[/TD] [TD]Crawley[/TD] [TD=align: right]1374[/TD] [/TR] [TR] [TD=align: right]62[/TD] [TD]Eng3[/TD] [TD]Yeovil[/TD] [TD=align: right]1361[/TD] [/TR] [TR] [TD=align: right]63[/TD] [TD]Eng3[/TD] [TD]Portsmouth[/TD] [TD=align: right]1360[/TD] [/TR] [TR] [TD=align: right]64[/TD] [TD]Eng3[/TD] [TD]Bury FC[/TD] [TD=align: right]1352[/TD] [/TR] [TR] [TD=align: right]65[/TD] [TD]Eng3[/TD] [TD]Colchester Utd[/TD] [TD=align: right]1349[/TD] [/TR] [TR] [TD=align: right]66[/TD] [TD]Eng3[/TD] [TD]Oldham[/TD] [TD=align: right]1349[/TD] [/TR] [TR] [TD=align: right]67[/TD] [TD]Eng4[/TD] [TD]Torquay[/TD] [TD=align: right]1335[/TD] [/TR] [TR] [TD=align: right]68[/TD] [TD]Eng4[/TD] [TD]Rotherham[/TD] [TD=align: right]1332[/TD] [/TR] [TR] [TD=align: right]69[/TD] [TD]Eng4[/TD] [TD]Fleetwood[/TD] [TD=align: right]1324[/TD] [/TR] [TR] [TD=align: right]70[/TD] [TD]Eng3[/TD] [TD]Shrewsbury[/TD] [TD=align: right]1322[/TD] [/TR] [TR] [TD=align: right]71[/TD] [TD]Eng3[/TD] [TD]Scunthorpe[/TD] [TD=align: right]1315[/TD] [/TR] [TR] [TD=align: right]72[/TD] [TD]Eng3[/TD] [TD]Carlisle[/TD] [TD=align: right]1315[/TD] [/TR] [TR] [TD=align: right]73[/TD] [TD]Eng3[/TD] [TD]Walsall[/TD] [TD=align: right]1309[/TD] [/TR] [TR] [TD=align: right]74[/TD] [TD]Eng4[/TD] [TD]Chesterfield[/TD] [TD=align: right]1308[/TD] [/TR] [TR] [TD=align: right]75[/TD] [TD]Eng4[/TD] [TD]Cheltenham[/TD] [TD=align: right]1308[/TD] [/TR] [TR] [TD=align: right]76[/TD] [TD]Eng4[/TD] [TD]Bradford[/TD] [TD=align: right]1306[/TD] [/TR] [TR] [TD=align: right]77[/TD] [TD]Eng4[/TD] [TD]Dag & Red[/TD] [TD=align: right]1298[/TD] [/TR] [TR] [TD=align: right]78[/TD] [TD]Eng4[/TD] [TD]York[/TD] [TD=align: right]1286[/TD] [/TR] [TR] [TD=align: right]79[/TD] [TD]Eng4[/TD] [TD]Northampton[/TD] [TD=align: right]1283[/TD] [/TR] [TR] [TD=align: right]80[/TD] [TD]Eng4[/TD] [TD]Exeter[/TD] [TD=align: right]1279[/TD] [/TR] [TR] [TD=align: right]81[/TD] [TD]Eng4[/TD] [TD]Rochdale[/TD] [TD=align: right]1262[/TD] [/TR] [TR] [TD=align: right]82[/TD] [TD]Eng4[/TD] [TD]Morecambe[/TD] [TD=align: right]1239[/TD] [/TR] [TR] [TD=align: right]83[/TD] [TD]Eng4[/TD] [TD]Burton[/TD] [TD=align: right]1234[/TD] [/TR] [TR] [TD=align: right]84[/TD] [TD]Eng4[/TD] [TD]Wycombe[/TD] [TD=align: right]1225[/TD] [/TR] [TR] [TD=align: right]85[/TD] [TD]Eng4[/TD] [TD]Oxford Utd[/TD] [TD=align: right]1224[/TD] [/TR] [TR] [TD=align: right]86[/TD] [TD]Eng4[/TD] [TD]Accrington[/TD] [TD=align: right]1208[/TD] [/TR] [TR] [TD=align: right]87[/TD] [TD]Eng3[/TD] [TD]Hartlepool[/TD] [TD=align: right]1188[/TD] [/TR] [TR] [TD=align: right]88[/TD] [TD]Eng4[/TD] [TD]Aldershot[/TD] [TD=align: right]1179[/TD] [/TR] [TR] [TD=align: right]89[/TD] [TD]Eng4[/TD] [TD]Bristol Rovers[/TD] [TD=align: right]1177[/TD] [/TR] [TR] [TD=align: right]90[/TD] [TD]Eng4[/TD] [TD]Plymouth[/TD] [TD=align: right]1176[/TD] [/TR] [TR] [TD=align: right]91[/TD] [TD]Eng4[/TD] [TD]Wimbledon[/TD] [TD=align: right]1144[/TD] [/TR] [TR] [TD=align: right]92[/TD] [TD]Eng4[/TD] [TD]Barnet[/TD] [TD=align: right]1131[/TD] [/TR] [/TABLE]

[kovomaster]

Re: ELO & Poisson rating system Hi giraldi, nice work done :) I know about that how much effort does it take... I have something even bigger (more than 10000 lines of code) and calculations took hours on thousands of historical matches. From your predictions I can see that you use mostly poisson distribution to count the propabilities of goal scoring. This is nice of course. Have you tried to convert ELO into 3 probabilities? :) Isn't this the hardest thing? Or do you take the same approach as Joseph Buchdahl in Fixed Odds sports betting??? Just for curiosity...

[f0r3st727]

Re: ELO & Poisson rating system how do you calculate your probability? how do you get this ratings? is there any good formula to use to do it? :)

[Matt_Trakker]

Re: ELO & Poisson rating system

Eachteam started with an value of 2000 points 3 seasons ago and dependingon their results, according an ELO formula (a little bit modified)they won or lost points. Than for a given game I calculate theprobabilities. This probability is than affected by the probabilitycalculated with poison formula. for the past few games at homerespectively away. In my opinion this correction will highlightrecent form.

Hi....sothe numbers you've added below for the teams should be putinto >ELO or Poisson calculators andthen you'll be given the probability? Then you take into accountgoalscoring record, head to head injuries etc? Howdid you take away or add points....just like league scoring...3 ptswin, 1 draw, nothing for win?

[giraldi]

Re: ELO & Poisson rating system Hi, kovomaster I did not read Joseph Buchdahl's book. As I have said, I started on 2008/2009 season with 2000 rating points for each Premier team, 1700 for Championship, 1700 for League 1 and 1400 for league 2. So, on this moment each entry is possible only on League 2 with 1400 points. A good Championship team will be promoted on Premier with around 1800 points that-s look fine comparing with relegated Premier teams. Of course I transform rating difference on probabilities for 1,x,2, for each game taking into account the home advantage. Next because I would like to see the influence of last games I calculate with poisson probabilities for 1,x,2, considering last games. (taking into account for example last 10 games - the strength of the opponent is not so important for the average because all teams, usually play with very strong and very weak teams on the last 10) The base is the first set of probabilities but I consider a variable percent for the second set. (for example 75%-25% or 80%-20%) There could be a lot of discussion but the system can be easily adjusted if something is going real wrong. I also have an entire database and a lot of code lines.... I am sure I don-t take into account many aspects that I am sure the bookies do. (the importance of the game, important players missing, local derby...etc.). On short time I will take into account other few aspects. f0r3st727, The formula I used for lost/win points is a classic ELO formula ( you can google it) a little bit modified by me. After 4 seasons the rating are what I presented. On my opinion reflects pretty well real situation.

[aek fc]

Re: ELO & Poisson rating system Hi there i am new in punterslouge is there any way to share this system? i am interesting

[giraldi]

Re: ELO & Poisson rating system Hi, my rating system proves nothing so far. I think it is too early to speak about sharing. On the other hand I am willing to speak about systems with other users in order to improve our performances. I would like to speak about principles instead of performances. If the principle is good the performances will come. My opinion is that never will be invented a system to beat mathematical the bookmakers because they make the rules. We play only inside their cage. What we can do is to make "something" to help us select 5 games out of other 50 and focus on them. We will have than an big advantages over other players that have not our tools but play based on "feeling" That is only my point of view.

[giraldi]

Re: ELO & Poisson rating system

32. Fra2. Guincamp - Istres Draw @ 3.21 (33.28%) 33. Ger2. Cottbus - Hertha Hertha @ 2.13 (46.87%)

For today I will play 18 games. The larger number so far. All the games seem to have good probabilities and good prices. Will see. 34. Eng3. Swindon - Doncaster Doncaster @ 4.69 (47.26%) 35. Eng3. Hartlepool - Stevenage Stevenage @ 2.39 (72.73%) 36. Eng3. Yeovil - Notts Notts @ 2.47 (45.02%) 37. Eng4. Plymouth - York Over 2.5 @ 2.20 (61.37%) 38. Eng4. Bristol - Dag&Red Over 2.5 @ 1.74 (68.28%) 39. Ita1. Roma - Fiorentina Roma @ 2.00 (50.18%) 40. Ita2. Brescia - Reggina Brescia @ 1.99 (51.74%) 41. Ita2. Ternana - Stabia Stabia @ 2.95 (39.22%) 42. Ita2. Novara - Padove Padova @ 3.36 (36.54%) 43. Spa2. Guadalajara - Numancia Draw @ 3.33 (33.20%) 44. Spa2. Barcelona B - Elche Under 2.5 @ 2.14 (56.83%) 45. Spa2. Almeria - Lugo Under 2.5 @ 1.93 (60.88%) 46. Spa2. Las Palmas - Huesca Under 2.5 @ 1.87 (64.06%) 47. Fra1. Sochaux - Lille Draw @ 3.36 (35.65%) 48. Fra1. Rennes - Brest Draw @ 3.73 (32.09%) 49. Fra1. Troyes - Nice Over 2.5 @ 2.20 (71.67%) 50. Ger2. Paderborn - Hertha Hertha @ 1.99 (54.72%) 51. Ger2. Dresden - Bochum Under 2.5 @ 1.91 (58.59%)

[giraldi]

Re: ELO & Poisson rating system

34. Eng3. Swindon - Doncaster Doncaster @ 4.69 (47.26%) 35. Eng3. Hartlepool - Stevenage Stevenage @ 2.39 (72.73%) 36. Eng3. Yeovil - Notts Notts @ 2.47 (45.02%) 37. Eng4. Plymouth - York Over 2.5 @ 2.20 (61.37%) 38. Eng4. Bristol - Dag&Red Over 2.5 @ 1.74 (68.28%) 39. Ita1. Roma - Fiorentina Roma @ 2.00 (50.18%) 40. Ita2. Brescia - Reggina Brescia @ 1.99 (51.74%) 41. Ita2. Ternana - Stabia Stabia @ 2.95 (39.22%) 42. Ita2. Novara - Padove Padova @ 3.36 (36.54%) 43. Spa2. Guadalajara - Numancia Draw @ 3.33 (33.20%) 44. Spa2. Barcelona B - Elche Under 2.5 @ 2.14 (56.83%) 45. Spa2. Almeria - Lugo Under 2.5 @ 1.93 (60.88%) 46. Spa2. Las Palmas - Huesca Under 2.5 @ 1.87 (64.06%) 47. Fra1. Sochaux - Lille Draw @ 3.36 (35.65%) 48. Fra1. Rennes - Brest Draw @ 3.73 (32.09%) 49. Fra1. Troyes - Nice Over 2.5 @ 2.20 (71.67%) 50. Ger2. Paderborn - Hertha Hertha @ 1.99 (54.72%) 51. Ger2. Dresden - Bochum Under 2.5 @ 1.91 (58.59%)

0.46 units lost The bank is -0.13 units with 51 invested. It is a lot of work to keep the bank around 0 , believe me :D. Anyway I am sure I will manage to get profit this month. I just need few good days. This is the secret. Loose few units on bad days but win a lot on good days.:D Anyway I will make a good assessment to see what patterns/leagues/probabilities were profitable so far Today I play another 10 bets. 52. Ita1. Inter - Napoli Inter @ 2.36 (51.29%) 53. Ita1. Palermo - Juventus Draw @ 3.85 (33.37%) 54. Ita1. Torino - Milan Milan @ 2.18 (55.59%) 55. Ita1. Torino - Milan Over 2.5 @ 2.19 (52.18%) 56. Spa1. Betis - Barcelona Under 2.5 @ 2.88 (65.64%) Bet365 57. Spa1. Bilbao - Celta Vigo Under 2.5 @ 2.22 (62.06%) 58. Spa2. Xerez - Girona Draw @ 3.30 (33.58%) 59. Spa2. RM Castilla - Santander Under 2.5 @ 2.33 (51.64%) 60. Fra1. St. Etienne - Lyon Draw @ 3.37 (32.05%) 61. Fra1. Reims - Bordeaux Over 2.5 @ 2.42 (55.95%)

[giraldi]

Re: ELO & Poisson rating system

52. Ita1. Inter - Napoli Inter @ 2.36 (51.29%) 53. Ita1. Palermo - Juventus Draw @ 3.85 (33.37%) 54. Ita1. Torino - Milan Milan @ 2.18 (55.59%) 55. Ita1. Torino - Milan Over 2.5 @ 2.19 (52.18%) 56. Spa1. Betis - Barcelona Under 2.5 @ 2.88 (65.64%) Bet365 57. Spa1. Bilbao - Celta Vigo Under 2.5 @ 2.22 (62.06%) 58. Spa2. Xerez - Girona Draw @ 3.30 (33.58%) 59. Spa2. RM Castilla - Santander Under 2.5 @ 2.33 (51.64%) 60. Fra1. St. Etienne - Lyon Draw @ 3.37 (32.05%) 61. Fra1. Reims - Bordeaux Over 2.5 @ 2.42 (55.95%)

only 4 bets won out of 10, -1.05 units lost and this month I have 1.18 units lost considering 61 invested. I've been in worse situations :D. Today I play 62. Ita1. Bologna - Lazio Over 2.5 @ 2.15 (46.63%) 63. Spa1. Vallecano - Zaragoza Under 2.5 @ 2.01 (58.02%)

[giraldi]

Re: ELO & Poisson rating system

62. Ita1. Bologna - Lazio Over 2.5 @ 2.15 (46.63%) 63. Spa1. Vallecano - Zaragoza Under 2.5 @ 2.01 (58.02%)

Considering last week performance yesterday I had big profit +0.01 units. Anyway it is good for morale.;) Today I play 7 games. Every bet has small value on my rating "opinion".:unsure 64. Fra1. Lille - Toulouse Draw @ 3.50 (34.62%) 65. Fra1. Nicce - Rennes Draw @ 3.35 (30.32%) 66. Fra1. Valenciennes - PSG Draw @ 3.72 (31.89%) 67. Fra2. Nimes - Chateaureux Over 2.5 @ 2.42 (51.84%) 68. Fra2. Guincamp - Arles Draw @ 3.67 (31.91%) 69. Fra2. Laval - Angers Draw @ 3.18 (31.28%) 70. Fra2. Tours - Lens Draw @ 3.27 (30.30%)

[giraldi]

Re: ELO & Poisson rating system

64. Fra1. Lille - Toulouse Draw @ 3.50 (34.62%) 65. Fra1. Nicce - Rennes Draw @ 3.35 (30.32%) 66. Fra1. Valenciennes - PSG Draw @ 3.72 (31.89%) 67. Fra2. Nimes - Chateaureux Over 2.5 @ 2.42 (51.84%) 68. Fra2. Guincamp - Arles Draw @ 3.67 (31.91%) 69. Fra2. Laval - Angers Draw @ 3.18 (31.28%) 70. Fra2. Tours - Lens Draw @ 3.27 (30.30%)

2 bets won yesterday out of 7 and 0.06 units lost I said yesterday that 0.01 units is big profit. Than 0.06 units lost is a real disaster. ;) OK, 2 days ago I won 0.01, yesterday I lost 0.05, last week -0.2,+0.3...etc 70 units invested, 1.26 units lost. I feel like I play chess. :) The real launch is coming. :) Today I try 71. Fra1. Ajjacio - Reims Draw @ 3.40 (31.41%) 72. Fra1. Brest - Montpellier Draw @ 3.25 (34.86%) 73. Fra1. Lyon - Nancy Over 2.5 @ 1.94 (54.17%) 74. Fra2. Sedan - Dijon Over 2.5 @ 2.27 (58.48%

[giraldi]

Re: ELO & Poisson rating system

71. Fra1. Ajjacio - Reims Draw @ 3.40 (31.41%) 72. Fra1. Brest - Montpellier Draw @ 3.25 (34.86%) 73. Fra1. Lyon - Nancy Over 2.5 @ 1.94 (54.17%) 74. Fra2. Sedan - Dijon Over 2.5 @ 2.27 (58.48%

The best part is that Sedan was postponed so I lost only 3 instead of 4 . Now I have -4.23 units lost and 74 played on December. This month seems to not be able to beat the bookmakers margin but I have more 15 days. Today I try 75. Spa2. Lugo - RM Castilla Under 2.5 @ 2.21 (60.11%) 76. Fra1. Rennes - Valenciennes Draw @ 3.43 (29,52%) 77. Fra2. Angers - Auxere Draw @ 3.21 (32.80%) 78. Fra2. Monaco - Nimes Draw @ 3.90 (32.53%) 79. Fra2. Monaco - Nimes Under 2.5 @ 1.94 (32.53%)

[giraldi]

Re: ELO & Poisson rating system

75. Spa2. Lugo - RM Castilla Under 2.5 @ 2.21 (60.11%) 76. Fra1. Rennes - Valenciennes Draw @ 3.43 (29,52%) 77. Fra2. Angers - Auxere Draw @ 3.21 (32.80%) 78. Fra2. Monaco - Nimes Draw @ 3.90 (32.53%) 79. Fra2. Monaco - Nimes Under 2.5 @ 1.94 (32.53%)

Only 2 out of 5 but 0.84 units profit. I have to recover 3.39 units in December (Yield - 4.29%) Today I try 80. Eng2. Barnsley - Sheffeld Under 2.5 @ 2.05 (75,68%) ?? 81. Eng3. Walsall - Yeovil Yeovil @ 3.12 (41.79%) 82. Eng3. Shreswsbury - Carisle Over 2.5 @ 1.97 (63.34%) 83. Eng4. Aldershot - Rochdale Over 2.5 @ 2.05 (60.28%) 84. Ita1. Lazio - Inter Lazio @ 2.48 (50.57%) 85. Ita1. Lazio - Inter Over 2.5 @ 1.99 (57.55%) 86. Spa1. Sevilla - Malaga Under 2.5 @ 1.92 (58.23) 87. Spa2. Alcoron - Barcelona B Alcoron @ 2.47 (51.96%) 88. Spa2. Murcia - Gualalajara Draw @ 3.23 (37.00%) 89. Fra1. Reims - Lille Draw @ 3.40 (30.19%) 90. Fra1. Troyes - Ajaccio Over 2.5 @ 2.24 (82.04%) ?? 91. Fra2. Dijon - Istres Draw @ 2.96 (44.15%) 92. Fra2. Nantes - Caen Draw @ 3.06 (37.91%) 93. Ger1. Furth - Augsburg Draw @ 3.30 (44.18%) Few very high probabilities. I am very curious what will happen there. I check manually for Troyes games and everything seems to be OK. Ajaccio received at least 2 goals away this season (except 2 games out of 9). More they scored 12 goals away, more than many top teams (and Troyes has not the best defense at home). As I said I will try to watch live this game. Of course if the final score will be 0-0 but will have 20/20 shots on goal, 2 missed penalties ...etc, my system is right, ;)

[giraldi]

Re: ELO & Poisson rating system

80. Eng2. Barnsley - Sheffeld Under 2.5 @ 2.05 (75,68%) ?? 81. Eng3. Walsall - Yeovil Yeovil @ 3.12 (41.79%) 82. Eng3. Shreswsbury - Carisle Over 2.5 @ 1.97 (63.34%) 83. Eng4. Aldershot - Rochdale Over 2.5 @ 2.05 (60.28%) 84. Ita1. Lazio - Inter Lazio @ 2.48 (50.57%) 85. Ita1. Lazio - Inter Over 2.5 @ 1.99 (57.55%) 86. Spa1. Sevilla - Malaga Under 2.5 @ 1.92 (58.23) 87. Spa2. Alcoron - Barcelona B Alcoron @ 2.47 (51.96%) 88. Spa2. Murcia - Gualalajara Draw @ 3.23 (37.00%) 89. Fra1. Reims - Lille Draw @ 3.40 (30.19%) 90. Fra1. Troyes - Ajaccio Over 2.5 @ 2.24 (82.04%) ?? 91. Fra2. Dijon - Istres Draw @ 2.96 (44.15%) 92. Fra2. Nantes - Caen Draw @ 3.06 (37.91%) 93. Ger1. Furth - Augsburg Draw @ 3.30 (44.18%)

+5.41 units won yesterday and finally I am back to little profit +2.02 units. for 93 units played. (Yield +2.17%). I would be very happy to not loose this month. As I said I can prove very easy my profit last month was around 16%. The second month on profit would be great and encouraging. Of course it is still a lot to work with the system. Very interesting those big probabilities for Barnsley and Ajaccio with very very good prices. Today I try another 10 94. Ita1. Parma - Cagliari Parma @ 2.02 (62.63%) 95. Ita1. Genoa - Torino Draw @ 3.21 (50.06%) ?? 96. Ita1. Catania - Sampdoria Over 2.5 @ 2.09 (51.96%) 97. Ita1. Fiorentina - Siena Over 2.5 @ 2.04 (52.92%) 98. Ita1. Parma - Cagliari Over 2.5 @ 2.24 (47.26%) 99. Spa1. Zaragoza - Levante Levante AH(0) @ 2.70 100. Spa1. Barcelona - Atl. Madrid Under 3 @ 2.16 101. Fra1. PSG - Lyon Draw @ 4.01 (34.92%) 102. Fra1. Nancy - Bordeaux Over 2.5 @ 2.54 (54.85%) ?? 103. Ger1. Bremen - Nurnberg Under 2.5 Draw @ 2.17 (60.82%)