TigerTedd Posted January 12, 2014 Share Posted January 12, 2014 Erm, i'll try :- Maths is an exact science, football isnt ! Despite being battered by festa ,we still have a 95.7% of being in the play offs Ah, but statistics is not an exact science, by definition. It deals in probabilities, and probabilities are never certain. But the stats say we have a very good chance of being in the play offs, a 1 in 3 chance of getting in the autos, and very little chance of winning the league. Or a huge asteroid could land in earth before the end if the season and wipe out human kind. You think I'm joking, but at this stage there's actually more chance of that happening than us getting relegated. Which, depending on your priorities, is a comforting thought. Link to comment Share on other sites More sharing options...
RamNut Posted January 12, 2014 Share Posted January 12, 2014 Can someone summarise Albert's excellent post in two sentences please?? We are going to finish 2.56th.... Or 3.14th Link to comment Share on other sites More sharing options...
nfb Posted January 12, 2014 Share Posted January 12, 2014 Ah, but statistics is not an exact science, by definition. It deals in probabilities, and probabilities are never certain. But the stats say we have a very good chance of being in the play offs, a 1 in 3 chance of getting in the autos, and very little chance of winning the league. Or a huge asteroid could land in earth before the end if the season and wipe out human kind. You think I'm joking, but at this stage there's actually more chance of that happening than us getting relegated. Which, depending on your priorities, is a comforting thought. Thats not two sentences...... Link to comment Share on other sites More sharing options...
furball Posted January 12, 2014 Share Posted January 12, 2014 I thought I would post this here to keep all the "data" in one place. If you take all the games which have not been played yet and predict the result using the reverse fixture, you can get an idea how the league will finish. I know in the Championship anyone can beat anyone, and just because Wednesday beat Leicester at home, doesn't mean they will do it again away. However, it does give a little more insight than pure points per game as it takes into account your opponents. Short story: If results go the way they have done, the league will finish like this: LEI (96) BUR (91) QPR (89) DER (85) NOT (77) WIG (72) B&H (70) LEE (69) IPS (68) REA (68) BLB (65) BLP (60) WAT (59) HUD (59) MID (58) BOU (55) BIR (54) BOL (54) CHA (47) MIL (44) SHW (40) DON (36) YEO (34) BAR (33) I'm sure at the start of the season, every Derby fan would have taken 4th, 15 points clear of 7th and only 11 points of the Champions. Link to comment Share on other sites More sharing options...
PistoldPete2 Posted January 12, 2014 Share Posted January 12, 2014 To say you 'think it's more like a 10% chance' really does crap all over Albert's work. I don't know Albert, but clearly he's got some sort of background in statistical analysis. He's based his assertions on cold hard data going back for several seasons. What facts do you base your 10% on, can we see your working out? Albert's already said that a 30% chance doesn't necessarily make it a likely result. In fact, it makes it an unlikely result. So what is the point if declaring it wrong and it should be 10% instead? It's still unlikely whichever way you look at it. We all know that there is a very unpredictable nature to football, that's why we all love it. There's still a very small chance that we could get relegated. All you can do is come up with the odds of certain possibilities based on the data available, but bookies wouldn't exist if the odds weren't fallible. Thanks I dont mean to dis Albert ; all I am saying is we have close to 0% chance of finishing ahead of Fester we are 10 points behind them and on the evidence of Friday they are way better than us. So that gives one place to go for and we are third in line for that one place. Plus we can't , though it pains me to say , rule out the possibility of Forest or one or two others finishing ahead of us , which is at least as possible as us finishing ahead of Burnley. If there are 4 teams going for one place then if all are equal thats 25%. But its not equal , we are 6 points behind burnley and 5 points behind QPR . Even if we were level with QPR most bookies would favour QPR because of the depth of their squad and their owners pockets. Fester are on target for 99 points, Burnley 92 points and QPR 90 points, which makes this season unusually tough at the top. You cant just ignore the opposition. If we get automatic promotion it would be a fantastic achievement, even a play off place is a very good achievement given the fact that other teams have more money. Link to comment Share on other sites More sharing options...
McRamFan Posted January 12, 2014 Share Posted January 12, 2014 either 2nd or 7th, I hate play-offs Link to comment Share on other sites More sharing options...
PistoldPete2 Posted January 12, 2014 Share Posted January 12, 2014 To say you 'think it's more like a 10% chance' really does crap all over Albert's work. I don't know Albert, but clearly he's got some sort of background in statistical analysis. He's based his assertions on cold hard data going back for several seasons. What facts do you base your 10% on, can we see your working out? Albert's already said that a 30% chance doesn't necessarily make it a likely result. In fact, it makes it an unlikely result. So what is the point if declaring it wrong and it should be 10% instead? It's still unlikely whichever way you look at it. We all know that there is a very unpredictable nature to football, that's why we all love it. There's still a very small chance that we could get relegated. All you can do is come up with the odds of certain possibilities based on the data available, but bookies wouldn't exist if the odds weren't fallible. "Can we see your working out"? well see my earlier post I think I shows its a lot less than 25% for reasons given. Also bookies give us about 30% chance of promotion. Our best bet of promotion at the moment is the play offs ; what does Albert say about that : a 90% chance of finishing in the top 6? If we finish 3rd to 6th then we have say a 25% chance of getting promotion via the play offs. Overall, we now have about 20% chance of going up by the play offs. If the overall chance of promotion is 30% then by deduction the chance of automatic is about 10% . Will that do? Link to comment Share on other sites More sharing options...
Albert Posted January 12, 2014 Author Share Posted January 12, 2014 We are going to finish 2.56th....Or 3.14thEssentially it's a method of creating some kind of smoothness in the function to allow better methods of analysis to be used.For those curious of the whole process involved, it was roughly outlined over the years when I put it together, but all in one place (and trying to keep it simplified):1. Using the wins, draws and losses from the season so far the average points per game and it's variation are calculated2. This is then used to create a probability density function of points per game against likelihood for the remainder of the season- Note that the above was checked and calibrated over several seasons worth of results, that is, I ran the check of probability against the entire league table for several seasons, and compared the results to the expected result, and used that as calibration. That is, the probabilities match what has happened in the past for a large sample set- I also plan to update the calibration with more data at some point, but this is more about fine details than anything else3. This is then used to "fill out" the remaining games, and the probability is as such worked out for various final points totals. This in effect gives a probability density against final points per game4. From here a points per game to league table function of some kind is produced, this allows the probability density against points per game to be converted to a probability density to final position, how the data is usually presented. This is done currently by using the method outlined in the first post. Essentially, using old data to predict how the table will evolve on average, then interpolating between positions there to give a continuous function to allow for a continuous probability density curve (this is where results such as 2.51th position come from for example,5. Using the probability density to position final position outputs are made, usually the calculated probabilities (that is, the probabilities of finishing within ranged, usually things like top 2, top 6, in first etc.) and probability density plots.So yeah, that's basically the method by which is it done. I could go through the exact measures, the data, calibration and such again if anyone is curious later in the week. I have it automated (and it needs to be, it's a lot of data to process), so basically an outline of the table as it stands, the wins, draws and losses from the team and such goes in, and the results come out. It is of course not perfect, and there are improvements that can be made still, but it should give a rough idea of the chances were taking about, rather than simply divining it from various liquids.In anycase, I think I'll chuck make a system for pie charts as well, who doesn't like piecharts? Link to comment Share on other sites More sharing options...
TigerTedd Posted January 12, 2014 Share Posted January 12, 2014 "Can we see your working out"? well see my earlier post I think I shows its a lot less than 25% for reasons given. Also bookies give us about 30% chance of promotion. Our best bet of promotion at the moment is the play offs ; what does Albert say about that : a 90% chance of finishing in the top 6? If we finish 3rd to 6th then we have say a 25% chance of getting promotion via the play offs. Overall, we now have about 20% chance of going up by the play offs. If the overall chance of promotion is 30% then by deduction the chance of automatic is about 10% . Will that do? Yep, fair enough, I'll give you that. That's good working out. Although we've got a good chance of getting in the play offs, our knack of killing off the lower league teams won't do us any good there. We really do have to up our game one way or another (either to the point that we're able to be considered a real contender for second, or to the point that we're able to be able to beat the teams around us and be a real contender for the play off champions) to stand of chance of promotion. Having said that, as others have said, if we'd have been offered a play off place at the started of the season, we would have snapped whoever was offering it's arm off. So it's all positive really. Link to comment Share on other sites More sharing options...
Albert Posted January 12, 2014 Author Share Posted January 12, 2014 "fFxRygN" alt="fFxRygN"> Link to comment Share on other sites More sharing options...
Albert Posted January 12, 2014 Author Share Posted January 12, 2014 "LreIXY7" alt="LreIXY7">Another version with the positions split into groups:First: 1Second: 2Playoffs: 3-6Top Half: 7-12Midtable: 13-18Battling: 19-21Relegation: 22-24 Link to comment Share on other sites More sharing options...
TroyDyer Posted January 12, 2014 Share Posted January 12, 2014 I suddenly fancy a pizza with a frightening amount of food colouring on it. Link to comment Share on other sites More sharing options...
RamNut Posted January 12, 2014 Share Posted January 12, 2014 I think this is over-estimating our chances of finishing in the top 2. And my guess is that 7th, 8th is not so much less likely than 4th,5th,6th. Link to comment Share on other sites More sharing options...
Leicester Ram Posted January 12, 2014 Share Posted January 12, 2014 If you could come into school and explain this to all my Leicester mates then that would be great Thanks in advance. Link to comment Share on other sites More sharing options...
TroyDyer Posted January 13, 2014 Share Posted January 13, 2014 I think we'll make the playoffs. That's a great achievement for the new management. They haven't built their own side yet and I really do think we'll see amazing things in the next 3 years. But playoffs with someone else's team, isn't to be sniffed at. Especially given the fact we haven't even been close to them since we won them. Link to comment Share on other sites More sharing options...
loughboroughRAM Posted January 13, 2014 Share Posted January 13, 2014 Our remaining fixtures are as follows with my predictions as to the outcome of the game based on current form and current league position so is purely circumstantial and down to me trying to be reasonable with predictions: Brighton (H) 1 Blackburn (A) 0 Yeovil (H) 3 Birmingham (A) 1 QPR (H) 1 Sheff Weds (A) 1 Bournemouth (H) 3 Burnley (A) 0 Bolton (H) 1 Millwall (H) 3 Reading (A) 0 Forest (H) 1 Ipswich (A) 1 Charlton (H) 3 Middlesborough (A) 0 Blackpool (A)1 Huddersfield (H) 3 Doncaster (A) 3 Barnsley (H) 3 Watford (H) 1 Leeds (A) 1 (W7 D10 L4) Thats 31 points from our remaining 21 games (average of 1.48 points a game) to leave us on 75 points come the end of the season. This prediction of results could obviously change quite dramatically if certain teams form picks up/drops off somewhat so lets say worst case scenario we end up W5 D8 L8 that would then give us 23 points from 21 games leaving us on 65 points come the end of the season. We could also experience above average form (similar to the form we have showed between Clough leaving and New Year) which could be in the region of W10 D7 L4, giving us 37 points meaning we finish up on 81 points which will give us a shout of automatic promotion (The average number of points needed for 2nd place over 11 seasons is 86). Based on each season since the Championship began the average number of points needed for the playoffs is 73. This suggests that either we'll just about make the playoffs (75 is the highest number of points a team finishing 6th have made) or miss out by what could be some way. However based on this season so far after 25 games Brighton in 6th have 39 points - an average of 1.56 points per game, taking them to 72 points by the end of the season. Based on these predictions it is hard to say whether we will make the playoffs or not, although barring poor/inconsistent form you'd bet on us making them (we hope), anything is possible in this league though remember! Link to comment Share on other sites More sharing options...
Jourdan Posted January 13, 2014 Share Posted January 13, 2014 I think we'll finish in the play-offs. We have an outside chance of automatic promotion but you need a different kind of consistency and staying power to finish in the top two, something we're yet to prove we have. I think the play-offs are a very realistic expectation at this point and an outcome that most of us would be happy with. Personally, I thought we'd finish in the top half but more likely 11th than 4th, so I'd be delighted. Our home form will be key. We have five of the current bottom eight still to play and generally you would look at games against Yeovil, Bournemouth, Charlton, Millwall, Barnsley, Huddersfield and Bolton as where most of our wins will come from, especially when you consider we won five of the seven reverse fixtures. Even games against Watford, Brighton, Forest and QPR are games where you would acknowledge the danger of the opposition but you wouldn't write us off completely either. It's going to be interesting to see whether our away form suffers, especially after Friday's result. To have won eight out of 13 away games and picked up 26 points from a possible 39 is a superb return. We've already won more away games this season than in the previous five seasons. We've also picked up more points than in four of the previous five seasons too, one point shy of the exception (27 points in 2010-11). We're on course for a record-breaking season in that respect. That said, our remaining away games do look decidedly tougher so a wobble is to be expected. Historically, we've struggled at Reading, Ipswich, Middlesbrough, Burnley, Blackburn and Blackpool and they're not going to be short of incentive for continuing the trend with all of them harbouring promotion ambitions of their own. The games I would pinpoint as ones we should pick up points in are Sheffield Wednesday, Doncaster, Leeds and Birmingham, who are markedly more inconsistent than the rest. I think 26-28 points from 21 games will be enough to squeeze us into the top six. But we should be aiming for and getting 30+ points from the remainder of the season and finishing in the top six very comfortably. Link to comment Share on other sites More sharing options...
Albert Posted January 13, 2014 Author Share Posted January 13, 2014 Below is a plot of ten years worth of performances by Championship sides in the second half of the season in comparison to the first half. The blue parts are cumulative probability for each range of performance, based on PPG scored in the second half of the season to PPG in the first:"b2LQmC0" alt="b2LQmC0">As can be seen, the fitting model fits quite well to the actual data, and in particular, if anything, it overestimates poor performance, not the other way around. Link to comment Share on other sites More sharing options...
Albert Posted January 14, 2014 Author Share Posted January 14, 2014 The above plot properly normalised:"ukG7S3a" alt="ukG7S3a">That is, the total probability of both have been set to one. I figured it would help with the comparison.Basically, that is the data for every single team to have played in the Championship from 2003-04 to 2012-13 showing how the analysis, at least to a rough extend, does predict the correct probability. Link to comment Share on other sites More sharing options...
PrivateDerby Posted January 16, 2014 Share Posted January 16, 2014 This is it now, well excited. Christmas and New Year is over and the hard work starts again. Lets do this Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.