This forum is read-only now. Please use Forum 2 for new posts
| xml | No replies possible in the archive |
| probability theory! | |
| Author: aydinfatih
Date: 13-08-2004, 13:49 | this is a theory about elimination system that first match have completed. firstly we make a few assumption. ****all teams have same condition, same stadiums, same players, same coach, same fans...all conditions are equal for two teams. **** home and away goal are been took care. ****the scores are not simetrical because of home and away goals. fore example: for away 1-2 win, the qualify probability is %68 but for home win it is %60 **** this is not a statistical theory. this is come from my football experience. ****the bigger scores than 5-0 and 0-5 did not took cared. ****this percentages are changable according to teams. for example wisla krakow-real madrid. because of the first match winner is real madrid with 0-2 (you know real is real). the percantage for real is bigger than %90. not %68 %95 5-0 %90 4-0 %85 5-1 %80 3-0 %77 4-1 %75 5-2 %72 2-0 %70 3-1 %67 4-2 %65 5-3 %62 1-0 %60 2-1 %57 3-2 %55 4-3 %52 5-4 %50 0-0 %50 1-1 %53 2-2 %55 3-3 %58 4-4 %60 5-5 %63 0-1 %65 1-2 %68 2-3 %70 3-4 %73 4-5 %75 0-2 %78 1-3 %80 2-4 %82 3-5 %85 0-3 %88 1-4 %90 2-5 %93 0-4 %95 1-5 %97 0-5 %98 |
| Re: probability theory! | |
| Author: artml
Date: 13-08-2004, 14:01 | So? Well, it`s already known and not so hard to calculate because it`s very simple model. And how you will appreciate the 'percentage' for different clubs? For example, PAOC vs Maccabi? 50/50 will not be correctly, ya - Maccabi is stronger? But how much? 40/60? or 41.5972/58.4028? Well, Real is real, it`s true... But Monaco, Barcelona and even Murcia are more real, aren`t them? Sorry if I was too evil, but I don`t understand what did you want to say? |
| Re: probability theory! | |
| Author: aydinfatih
Date: 13-08-2004, 14:32 | i think it is obvious what i want to say... it is not important if it is easy or hard to say that. i am not rediscovering america.. i am just generating something. you know, the theories are very simple thing, for example the grativity theory. but the important thing is generating.. so real madrid was only an example..if you can think team by team it is easy... let's look maccabi-paok maccabi won in the first match 1-2..in table for away match 1-2 is %68(the score can be change by 0-3..but do not care for now)...this is mean maccabi will qualify with a %68 percentage. but if they would had the same condition (players, field, coach, fair referee) and you think maccabi is more strong than paok (i do not think so)..according you it will be more than %68..but according to me it will be less than %68. so this is a general and under normal circumstances theory...and you know for every theory there are some assumption.. |
| Re: probability theory! | |
| Author: artml
Date: 13-08-2004, 14:43 | Hmmm... If I will tell my mathematics teacher soluting the probability task that sought quality is 'less than 68%', probably, I couldn`t get a good mark  Well, I think that real theory is the probability theory itself and your idea is only specific task. I agree with you that the main is to generate ideas, but it`s important to not to fly in the sky... P.S. How old r u? |
| Re: probability theory! | |
| Author: artml
Date: 13-08-2004, 14:53 | P.P.S. And don`t think that theory of gravity is simple... It`s very beautiful and one of themost fundamental theories, but it`s VERY hard. No jokes. |
| Re: probability theory! | |
| Author: aydinfatih
Date: 13-08-2004, 14:57 | how old am i??..what do you think??..10 or 8 or 7.. this is not vey important..this is a idea about football... is there any machine that we can ask what is the probability of wisla-real: -what is the probabilty of paok-maccabi -hmmm, paok, maccabi let me think..hmmm..what was the first match - 1-2 -hmmm let me enter the players and referee.(3 minute later)..it is %65.98888 -thank you.. so you know football is a unpredictable game (last year u remember milan and depor matches..)..and we came across unpredictable scores..because ther are many parameters: referee, stadium, fans, players, coachs, wheather, field circumstances...so percantages are very changable, but after generating you can think in particular.. |
| Re: probability theory! | |
| Author: artml
Date: 13-08-2004, 15:10 | "so you know football is a unpredictable game (last year u remember milan and depor matches..)..and we came across unpredictable scores..because ther are many parameters: referee, stadium, fans, players, coachs, wheather, field circumstances...so percantages are very changable, but after generating you can think in particular.. " There`s an answer. Yes, it`s unpredictable - and you are speaking about probabilities... Your idea is not bad, but it`s NOT A THEORY. And it`s very crude model of reality - too crude IMHO. My question about age is not a tease - it`s just curiosity. I`m novice there - so i wanna to know smth about people i chat with. I`m 18 yrs old male if you want to know it... |
| Re: probability theory! | |
| Author: Zuberbuler
Date: 13-08-2004, 15:14 | Ok but i can`t understand what do you want to prove with this theory?You can`t use it becouse it wouldn't work.If you want to say that if the score is 0-0 the both team have 50-50 chance... Let me ask you something.If Real Madrid face Gigant Belene(one of the most unpopular team in the world  did you think that Gigant Belene have 50% chance to win? |
| Re: probability theory! | |
| Author: Ricardo
Date: 13-08-2004, 15:19 | I think you don't take enough home-side advantage in your calculations. 0-0 is absolutely better for the away team, no 50-50. But I like this approach. Next I would like to see it together with some real results. Somebody with a lot of time can dig around in all last years results and see what has happened. Maybe together with a list of all second match games. A 1-0 first leg will not have often a 4-3 second leg, I think. Who has time for this? |
| Re: probability theory! | |
| Author: aydinfatih
Date: 13-08-2004, 15:24 | but also think that there is no %100 percentages..so every think can be happy..the 1% also a probability but a very weak probability...so because of many parameters i make a few assumptions (you know in pyhsic many theory accepted in no air circumstances, is there somewhere like this.)for what??...for generating (again it) it was for make some fun..you know football is for making fun..and hove a good fun.. P.S. i am 22 years old from istanbul.. |
| Re: probability theory! | |
| Author: aydinfatih
Date: 13-08-2004, 15:39 | to ricardo yes somebody can make a statistical look for in last 5 years for example how many team of first match 1-0 winner qualified next round or 0-1 or 2-0..i have time but not for today..but this won't be very logical because this will be change every time..because parameters firstly i thought 1-1 is 50-50, but after i take 0-0 50-50 to zuberbuler i said this before but i want to say again..this is a very general approach (every circumstances are same for two teams) ..you can think in particular for real-the weaker team and it will be definetely different from this..and it is according to you. |
| Re: probability theory! | |
| Author: artml
Date: 13-08-2004, 15:41 | Well, if we are building theory, let`s look to the involved subject. Build real statistics - like: first match result - percent of success result. Then get a distribution - REAL distribution. Your hypothesis is normal distribution with average of distribution 0 (=0-0) and unknown dispersion. Comparing these two distributions we`ll check this hypothesis (it`s simple task - there is a chi-square criteria for this comparing) and then, if we`ll get a statistical distribution like P(R) where P is the probability of success and R is match result (you must bring to conformity the set of results and the set of natural integers before) we could to make a prognosis. It`s a simple probability task but NOT A THEORY... |
| Re: probability theory! | |
| Author: artml
Date: 13-08-2004, 15:46 | "but also think that there is no %100 percentages..so every think can be happy..the 1% also a probability but a very weak probability...so because of many parameters i make a few assumptions (you know in pyhsic many theory accepted in no air circumstances, is there somewhere like this." No. There is integrated probability of the results like 7-0, 6-4, 28-843 etc. Have you study the probability theory anywhere? |
| Re: probability theory! | |
| Author: aydinfatih
Date: 13-08-2004, 15:53 | last two years 2003-2004, 2002-2003 the 49 first match end with 1-0 (winner is home) and 34 of first match winner qualify so: %69.3.. mine was %62 |
| Re: probability theory! | |
| Author: aydinfatih
Date: 13-08-2004, 15:58 | also for last two years: 27 match ended with 2-1 and 15 winners qualify so: %56 mine %60, not bad.. |
| Re: probability theory! | |
| Author: artml
Date: 13-08-2004, 15:59 | Aham. And do this for every result... Better for 5 or 10 years - reliability of sample will be better |
| Re: probability theory! | |
| Author: aydinfatih
Date: 13-08-2004, 16:01 | yes i am an engineer..and i have work about it in university "probabilty and statistic"..have you i have to go know..and we will continue later...see you later.. |
| Re: probability theory! | |
| Author: artml
Date: 13-08-2004, 16:07 | Yes. In special univercity course. Where d`u think from I got these words? Well, it`s an interesting task. I`ll look for more statistics and get something - e.g. distibution diagram. If you got smth - mail artml @ inbox.ru CU again... |
| Re: probability theory! | |
| Author: EarlofBug
Date: 13-08-2004, 18:48 | I think that 1-0 home win has less than 62%. Maybe 55%. 0-0 result should be 48% to the home team, and 52% to the visitor. |
| Re: probability theory! | |
| Author: bilocan
Date: 15-08-2004, 01:01 | Aydin-- How can you call this a "theory". You can "perhaps" call this a "rule of thumb" which appears to be working. Theories require proves(even emprical validations are not enough.). I know you wanna say "hey I found something". However, not only does it qualify as a theory it also lacks(very much) the justification of statistical validity. So my suggestion is stick with your daily "engineering" job and do not go into the field of statisticians. |
| Re: probability theory! | |
| Author: aydinfatih
Date: 16-08-2004, 12:03 | i did not examine in this for years, and i did not waste my time for this. i was lying on my bed, i said trabzonspor win in kiev by 1-2.. what could be the second match and so on i write on paper and after i write here i did not do any experiement on this.. i am not descovering something never mentioned or thought. statistic and football are used for many years together.. do not take care... |
| Re: probability theory! | |
| Author: anita
Date: 17-08-2004, 12:38 | Keep on your discovering, aydinfatih!  I think your probability theories are too simple, but if you all read the original aim of this forum, it is discussions around e.g. calculations, rankings and so on, so you are spot on, aydinfatih. So guys, discuss the validity of aydinfatihs probabilities without mocking and battering. Whether he is an engineer or a bag lady is not relevant. Football is a very tricky business when it come to measuring from statistics, but nice try, aydinfatih! regards |
| Re: probability theory! | |
| Author: Ricardo
Date: 17-08-2004, 15:08 | OK guys (& girls) I spend my lunchtime on some copying and pasting in excelsheets and collected all games on Bert's site, threw out the group matches and the (one-leg-)finals. And the one leg game of the Rangers. Together 948 games. This resulted in the following: 101 times 1-leg 0-0, 34 times the home team went through, 67-away.in Short 0-0: 34-67 (101) I hope you enjoy this as much as I do |
| Re: probability theory! | |
| Author: mark
Date: 17-08-2004, 15:52 | So if we apply this data to the close/suprise ties in CL QR3 we get the following precentage Rosenborg beats Haifa - 52.5% Rangers beats CSKA - 47.5% PAOK beats Tel-Aviv - 0% Kiev beats Trabzonspor - 2.5% Juventus beats Djurgårdens - 18% Sparta beats Ferencváros - 34% Brugge beats Shakhtar - 15% PSV beats Red Star - 41% Inter beats Basel - 71% Anderlecht beats benfica - 34% Deportivo beats Shelbourne - 66% |
| Re: probability theory! | |
| Author: artml
Date: 17-08-2004, 16:40 | Sad statistics... My congradulations to Turkey but let`s see it next week |
| Re: probability theory! | |
| Author: TKW
Date: 17-08-2004, 16:53 | Thanks to Ricardo, fascinating statistics. You can have more fun with them! The team drawn at home in the first leg wins only 47% of ties - conventional wisdom is confirmed. 2-1 in the first leg is the most evenly balanced score - the home team wins the tie 53% of the time. The most common scores are: 1-0 11.8% 0-0 10.7% 1-1 8.4% 0-1 7.5% 2-0 6.9% 3-0 6.6% 2-1 6.2% No team has ever recovered from conceding 5 in the first leg, home or away. Etc.! |
| Re: probability theory! | |
| Author: Ricardo
Date: 17-08-2004, 17:05 | Yes, a goal difference of 4 is fatal, but 3 away not: 3-0, 4-1, 5-2: 87- 7. It still happened in 7.5% of the case (I don't say there is a 7.5% chance!!)that loosing away with 3 goals difference can be undone. Fresh in our memory lies ofcourse last year's CL-Quarterfinal Milan-Depor(4-1, 0-4) |
| Re: probability theory! | |
| Author: Forza-AZ
Date: 17-08-2004, 17:12 | So converted in percentages it's this way: 0-0: 34- 66 0-1: 15- 85 0-2: 0-100 0-3: 0-100 0-4: 0-100 0-5: 0-100 0-6: 0-100 0-7: 0-100 0-10: 0-100 1-0: 66- 34 1-1: 29- 71 1-2: 2- 98 1-3: 0-100 1-4: 0-100 1-5: 0-100 1-6: 0-100 1-7: 0-100 2-0: 92- 8 2-1: 53- 47 2-2: 18- 82 2-3: 0-100 2-4: 0-100 2-5: 0-100 2-6: 0-100 3-0: 95- 5 3-1: 89- 11 3-2: 59- 41 3-3: 33- 67 3-4: 0-100 4-0: 100- 0 4-1: 85- 15 4-2: 67- 33 4-3: 100- 0 4-4: 0-100 5-0: 100- 0 5-1: 100- 0 5-2: 100- 0 6-0: 100- 0 6-1: 100- 0 7-0: 100- 0 7-2: 100- 0 8-0: 100- 0 8-1: 100- 0 And sorted in the best score for the home team (of the 1st leg) to advance: 4-0: 100- 0 4-3: 100- 0 5-0: 100- 0 5-1: 100- 0 5-2: 100- 0 6-0: 100- 0 6-1: 100- 0 7-0: 100- 0 7-2: 100- 0 8-0: 100- 0 8-1: 100- 0 3-0: 95- 5 2-0: 92- 8 3-1: 89- 11 4-1: 85- 15 4-2: 67- 33 1-0: 66- 34 3-2: 59- 41 2-1: 53- 47 0-0: 34- 66 3-3: 33- 67 1-1: 29- 71 2-2: 18- 82 0-1: 15- 85 1-2: 2- 98 0-2: 0-100 0-3: 0-100 0-4: 0-100 0-5: 0-100 0-6: 0-100 0-7: 0-100 0-10: 0-100 1-3: 0-100 1-4: 0-100 1-5: 0-100 1-6: 0-100 1-7: 0-100 2-3: 0-100 2-4: 0-100 2-5: 0-100 2-6: 0-100 3-4: 0-100 4-4: 0-100 |
| Re: probability theory! | |
| Author: Ricardo
Date: 17-08-2004, 17:19 | What I find interesting to see is the change in behaviour with a 1 goal difference home win: 1-0: 66% in favor of home team 2-1: 53% 3-2: 59% 4-3: 100% especially the last 3 seems right: 1-0 is enough for the away team, but if there are more goals in the 1st leg, it's less likely that the score cen be held to just 1 goal. |
| Re: probability theory! | |
| Author: sr_sofisticacao
Date: 17-08-2004, 19:08 | This is great! Thanks Ricardo! These 948 matches are what? All the matches in the database? These include from what year on? And I think it will be fun to see how it will adapt for the next rounds! If Rosenborg, CSKA, Tel-Aviv, Trabzonspor, Djurgårdens, Ferencváros, Shakhtar, Red Star, Inter, Benfica and Deportivo go through, we have ourselves one hell of a predictor! |
| Re: probability theory! | |
| Author: SHEV
Date: 17-08-2004, 19:42 | I think it is 4 years as there is about 200 matches each year. |
| Re: probability theory! | |
| Author: JPV
Date: 18-08-2004, 11:41 | this might be even more usefull: ALL games starting from 1955, excluding one-leg finals, group games & 3-leg games (before penalties were introduced):0-0 151-346 (497) 30% / 70% these are remarkable: 5-1 66- 1 (67) 99% / 1% 6-2 6- 2 (8) 75% / 25% these are these games: 61/62 Cup Win. I FC La Chaux-de-Fonds Sui Leixoes SC Por 6-2 0-5 84/85 UEFA II Queen's Park Rangers Eng FK Partizan (Beograd) Yug 6-2 0-4 85/86 UEFA III Borussia Mönchengladbach FRG Real Madrid CF Esp 5-1 0-4 |
| Re: probability theory! | |
| Author: artml
Date: 18-08-2004, 12:29 | Hmmm... I don`t think that all-time statistics is more useful. It`s better to take last 15 years - after beginning of CL (it makes a big dfference in footbal IMHO), and also because in last years (`50-`70) there were more resultative football with less logical results (I 'remember' Spartak Moscow - Werder Bremen 4-1 2-6 match - we can`t see in in present time). But thank you for your work! Maybe Berth should put this statistics somewhere on site? |
| Re: probability theory! | |
| Author: Forza-AZ
Date: 18-08-2004, 13:21 | Nice statistics JPV. This is the most interesting: 1-0 407-319 (726) 56% / 44% 2-1 197-232 (429) 46% / 54% 3-2 61- 82 (143) 43% / 57% 4-3 9- 10 (19) 47% / 53% It clearly shows that losing away with 1 goal difference (and scoring yourself) is indeed a good result as is said so often. |
| Re: probability theory! | |
| Author: SHEV
Date: 18-08-2004, 15:28 | More interesting that you have better chances if you lose 2-4 then 1-2 |
| Re: probability theory! | |
| Author: JPV
Date: 18-08-2004, 16:23 | that's "the statistics of low data" if you only have 25 results, and 1 is a HUGE exception, you still have 4% "chance". if you'd have 200000 games, you might only get 1% chance |
| Re: probability theory! | |
| Author: Ricardo
Date: 19-08-2004, 17:02 | If anybody still cares: I took the matches out of Bert's database: that is the last 5 year. I did not include this years matches. Great work JPV, it is interesting to see the differences, and the n I especially mean the home winning with 1 goal difference: The last 5 years, the home team had an advantage, looking at your table the away team has advantage. Also funny to see is that the most popular result (1-0) is very balanced (56-44). Next step is to split the resutls up in seeded and not-seeded....... |
| Re: probability theory! | |
| Author: JPV
Date: 25-08-2004, 15:04 | and to prove "probabilities" mean shite: A club losing 4-1 in the first game, has 8% "chance" to get to the next round. However: Club Brugge lost with 3 goals difference on 4 previous occasions, and got through to the next round in 2 of the 4 occasions (against Dortmund and Ipswitch) Shakhtar Donets'k lost with 3 goals in 8 of their 34 games... Shakhtar played twice against Brugge before and never got through to the next round (but yes, most likely Shakhtar will advance) |
| Re: probability theory! | |
| Author: aydinfatih
Date: 22-09-2004, 14:17 | after searching in google i came acros a site about this link. if you like probabilities and statistics you can make a look.. http://galeb.etf.bg.ac.yu/~mirad/Odds/ecstat.txt also http://galeb.etf.bg.ac.yu/~mirad/odds.html http://galeb.etf.bg.ac.yu/~mirad/Odds/ec_odds.txt |