Champions League 2013/14
Nice piece: -
How Uefa's seeding system helps Arsenal & hinders Celtic
http://www.bbc.co.uk/sport/0/football/23680545
How Uefa's seeding system helps Arsenal & hinders Celtic
http://www.bbc.co.uk/sport/0/football/23680545
-
giulio2010
- Posts: 962
- Joined: Sun Sep 12, 2010 9:11 am
Group A: Manchester United, Shakhtar Donetsk, Bayer Leverkusen, Real Sociedad
Group B: Real Madrid, Juventus, Galatasaray, Copenhagen
Group C: Benfica, Paris St-Germain, Olympiakos, Anderlecht
Group D: Bayern Munich, CSKA Moscow, Manchester City, Viktoria Plzen
Group E: Chelsea, Schalke, Basel, Steaua Bucharest
Group F: Arsenal, Marseille, Borussia Dortmund, Napoli
Group G: Porto, Atletico Madrid, Zenit St Petersburg, Austria Vienna
Group H: Barcelona, AC Milan, Ajax, Celtic
Group B: Real Madrid, Juventus, Galatasaray, Copenhagen
Group C: Benfica, Paris St-Germain, Olympiakos, Anderlecht
Group D: Bayern Munich, CSKA Moscow, Manchester City, Viktoria Plzen
Group E: Chelsea, Schalke, Basel, Steaua Bucharest
Group F: Arsenal, Marseille, Borussia Dortmund, Napoli
Group G: Porto, Atletico Madrid, Zenit St Petersburg, Austria Vienna
Group H: Barcelona, AC Milan, Ajax, Celtic
-
giulio2010
- Posts: 962
- Joined: Sun Sep 12, 2010 9:11 am
Very tough for the italians...
Very lucky for Benfica, PSG, Bayern, Man city and Chelsea!
Very lucky for Benfica, PSG, Bayern, Man city and Chelsea!
I like matches where a bigger team lose. Especially after yesterday carnage: Galatasaray 1-6 Real MadridEuler wrote:Mare for Chelsea!
if you watch only the first half you'll never guess the final score. I think if Ancelotti was at Chelsea and Mourinho at Real we would have had a reverse
.Will see how they will play the next match.
26/11 Basel - Chelsea
Champions league ratings for this evening, forecast variation to long term average: -
Most to least goals: -
Arsenal Vs. Dortmund - 118%
Austria Vienna Vs. Atl Madrid - 110%
Schalke Vs. Chelsea - 108%
AC Milan Vs. Barcelona - 108%
Steaua Bucharest Vs. FC Basel - 106%
Porto Vs. Zenit St Petersburg - 101%
Celtic Vs. Ajax - 100%
Marseille Vs. Napoli - 99%
Most to least goals: -
Arsenal Vs. Dortmund - 118%
Austria Vienna Vs. Atl Madrid - 110%
Schalke Vs. Chelsea - 108%
AC Milan Vs. Barcelona - 108%
Steaua Bucharest Vs. FC Basel - 106%
Porto Vs. Zenit St Petersburg - 101%
Celtic Vs. Ajax - 100%
Marseille Vs. Napoli - 99%
Peter, Is the percentage you include the chance of goals above the average?Most to least goals: -
Arsenal Vs. Dortmund - 118%
Austria Vienna Vs. Atl Madrid - 110%
I am not sure if there is anything in this, but if you took the average implied odds of the "over 2.5 goals market" for each of last nights games it was 50%.
2.5 times 8 = 20 goals and there were 20 goals last night. Not sure if my maths or logic is right there. I noticed your system predicted 22.
Last night, I headed down the pub with my laptop to test a theory with a friend. The theory is that the best chance of a swing is to lay the team that score the first goal for matches where the pre-match odds are competitive (i.e. win & loss odds are similar) and the over and under markets point towards a decent number of goals "likely" to be scored. We green-up whichever team scores the next goal.
We picked out the following matches:
Arsenal v Dortmund (W2.56, D3.6) - Profit
Celtic v Ajax (W2.5, D3.5) - Loss
Marseille v Napoli (W2.9, D3.4) - Loss
Schalke v Chelsea (W3.5, D3.55) - Loss
Steaua Bucharest v Basel (W3.1, D3.5) - Profit
Clearly we need to look at a lot more matches, but I'd be interested if anyone has any feedback on the following:
1. The theory maybe flawed in that although the odds suggest they aren't sure who will win and there will be goals, statistically the first team who scores may score all the goals in the match. The first goal maybe defining.
2. The match that swung the most from the selection was the Arsenal game, it was also had the second most competitive odds out of the eight games last night W:39%/D:28%/L:33%. It also had the best odds in the over 4.5 goals market. Although, more previous data is needed as I don't want to fall into statistical deja-vu.
3. Other markets could be added to selection criteria. For example "Both teams to score", "Home & away teams total goals" could project the likely hood of a team coming back and scoring a goal.
4. If this was a league, say the English Championship, I could also maybe factor in statistics on the likely hood of a "comeback" based on previous matches, however I can't see how I could factor that in the Champions League.
5. The trades could be altered so that the rather than green-up after the initial lay they run or I could lay and back for every goal including laying the draw from the start.
Steve.
We picked out the following matches:
Arsenal v Dortmund (W2.56, D3.6) - Profit
Celtic v Ajax (W2.5, D3.5) - Loss
Marseille v Napoli (W2.9, D3.4) - Loss
Schalke v Chelsea (W3.5, D3.55) - Loss
Steaua Bucharest v Basel (W3.1, D3.5) - Profit
Clearly we need to look at a lot more matches, but I'd be interested if anyone has any feedback on the following:
1. The theory maybe flawed in that although the odds suggest they aren't sure who will win and there will be goals, statistically the first team who scores may score all the goals in the match. The first goal maybe defining.
2. The match that swung the most from the selection was the Arsenal game, it was also had the second most competitive odds out of the eight games last night W:39%/D:28%/L:33%. It also had the best odds in the over 4.5 goals market. Although, more previous data is needed as I don't want to fall into statistical deja-vu.
3. Other markets could be added to selection criteria. For example "Both teams to score", "Home & away teams total goals" could project the likely hood of a team coming back and scoring a goal.
4. If this was a league, say the English Championship, I could also maybe factor in statistics on the likely hood of a "comeback" based on previous matches, however I can't see how I could factor that in the Champions League.
5. The trades could be altered so that the rather than green-up after the initial lay they run or I could lay and back for every goal including laying the draw from the start.
Steve.
You do not have the required permissions to view the files attached to this post.
My methodology is based on something I did 20-30 years ago. I was trying to work out what created a draw so I figured two teams with similar scoring rates will most likely end in a draw. The closer the scoring rates are to an even number the more likely the draw.stevequal wrote:Peter, Is the percentage you include the chance of goals above the average?
I am not sure if there is anything in this, but if you took the average implied odds of the "over 2.5 goals market" for each of last nights games it was 50%.
2.5 times 8 = 20 goals and there were 20 goals last night. Not sure if my maths or logic is right there. I noticed your system predicted 22.
So I started working on scoring rates. But you realise that each footy market is a step function per goal. So 1 to 2 nil is a big jump and in reality you can't exactly predict that. So the only way to make a sensible judgement, over time, is to say if this match was played 100 times how many goals would be scored on average. So you end up with impossible score lines like 2.48 to 0.74 for a total of 3.22 goals. On individual matches you will be above or below that but in aggregate, especially over the course of a season, you can be very accurate. Each season I'm within a fraction of the actual number against forecast.
Once you have that data you can work out all the other stuff, such as time of first goal.
Peter, I wonder if you were trying to win the Pools? 
By scoring rates do mean how many goals a team scores on average? Or how often a goal is scored?
Liverpool score on average 2.65 goals a game.
Liverpool score on average one goal every 28 minutes.
By scoring rates do mean how many goals a team scores on average? Or how often a goal is scored?
Liverpool score on average 2.65 goals a game.
Liverpool score on average one goal every 28 minutes.
I won hundreds of dividends in the end including a first on Littlewoods.
There isn't a direct correlation between goals and time, but that's more or less right. It depends how refined you want to get it. I managed to work out key variables that accounted for the vast majority of difference between teams which can go right back to individual players. But because of the variability, it's really tricky to put some of that to use in the short term, it only plays out over vast numbers of matches. Information losses it's value the deeper you go as well.
You also get a few curves balls thrown in. You price up something, only for the team to change or the manager to change the shape. Whatever you do footy tends to produce large variations to mean unless you apply it consistently over long periods.
A lot of the trading or betting systems out there on footy tend to be high % processes that dupe people into thinking its the selection process that is wrong. It's quite possible to run a profit on football (and tennis) for long periods of time without achieving and underlying profitability.
There isn't a direct correlation between goals and time, but that's more or less right. It depends how refined you want to get it. I managed to work out key variables that accounted for the vast majority of difference between teams which can go right back to individual players. But because of the variability, it's really tricky to put some of that to use in the short term, it only plays out over vast numbers of matches. Information losses it's value the deeper you go as well.
You also get a few curves balls thrown in. You price up something, only for the team to change or the manager to change the shape. Whatever you do footy tends to produce large variations to mean unless you apply it consistently over long periods.
A lot of the trading or betting systems out there on footy tend to be high % processes that dupe people into thinking its the selection process that is wrong. It's quite possible to run a profit on football (and tennis) for long periods of time without achieving and underlying profitability.