My only doubt about ratings is that as the season progresses things can change. Teams cut adrift may just give up while those like Rotherham and Fulham may figh harder to keep clear of the drop zone. Who thought we would be saying that about Fulham at this stage?
There is that, CT and there is also the role that current form has to play. Bournemouth were top of the form table until we beat them and now they've dropped down to the bottom of the mini league on form, whereas we are at the top. We need to take advantage and get as far up the table while it's in our favour. There's no guarantee it'll stay that way.
I think this is one of the best ways of veiwing the run however, and I'm not saying you should do this JK, there are several coefficients that could be included like layering form on top of the just league position. For example leeds and Forrest both look out of position having had poor management all season which the both seem to have rectified. I'm a big fan of statistics and they can give a really good veiw if things like the relative difficulties of the remaining fixtures. What I like about football is that it doesn't bow to being played out on paper.
I think she was not alluding to stats, or even to anything allied to football. CR, do you have a problem you are keeping quiet about?
I'm back again! With more useless stats! I've made improvements (I hope) to my model for assessing the difficulty of the run-in for the top 8 clubs. Purely based on league position, it looks like this: Norwich----6.42------77.00 B'ford-------6.67------80.00 B'mouth----6.67-----80.00 Derby-------6.67------80.00 'Boro--------7.08------85.00 Watford-----7.17------86.00 Wolves------7.25------87.00 Ipswich------7.42-----89.00 The first figure is the average difficulty of remaining matches (out of 10, 10 being the hardest), and the second figure is total cumulative difficulty of all remaining fixtures. Note that the absolute lowest (easiest) these figures could possibly be (as in if a team essentially had to play all lowest-ranked teams) is 4.00 & 50.00, whilst the highest (hardest) is 9.00 & 112.00. To expand upon this, I sorted the league table firstly by home form, and secondly by away form, and assigned a 'difficulty coefficient' to each potential fixture. Hopefully, by applying this coefficient to simple league ranking, form could be somewhat factored-in. The results of this are as follows: B'ford------0.85----10.15 B'mouth---0.85----10.19 Watford----0.98----11.80 Norwich---0.99-----11.91 Derby------1.05-----12.58 Ipswich----1.07-----12.87 Wolves-----1.10-----13.17 'Boro-------1.18-----14.15 The first figure is, again, the average rating of difficulty of the remaining matches (this time with a maximum of 2.40, 2.40 being the hardest match possible). The second figure is total cumulative difficulty of all remaining fixtures. Note that the absolute lowest (easiest) these figures could possibly be (as in if a team essentially had to play all the worst-performing, out of form teams) is 0.18 & 02.21, whilst the highest (hardest) is 1.98 & 23.77. So, after factoring in 'form', it appears Brentford and Bournemouth have the 'easiest' run-ins, whereas 'Boro actually appear to have some pretty tough games to come. Although our games look the easiest simply looking at league-ranking, the form of some of the teams we still have to play makes our task seem that little bit harder.
This is what I love about statistical modelling, you can add more and more layers to create something ever more meaningful. What people tend to forget is this isn a prediction based on numbers, it is a factual statement of that task that is ahead for each team. That said there may be some interesting ways to build this into a prediction. I may have a go myself at somepoint if I get some time.
The problem with predictions is the role of luck. Whilst I think that luck probably evens out over a season it doesn't over shorter periods such as a dozen games. You only need a couple in there like us at home to Cardiff last season or Charlton this to put a pretty big skew in the mix. Ultimately a dodgy penalty either way could be the difference between second and third.
First of all, thanks again to JK for his excellent work on this which I also find fascinating. I agree with CT about 'luck' not being predictable, but at least this provides an objective framework around any deviation for luck. I'd add to that a possible drop in performance when playing struggling teams, something which City has been all too prone to in the past. It will be a test of AN to see if the level of performance can be maintained, but I suspect that it can, as he emphasizes organisation and discipline rather than 'emotional' motivation. Wigan and Millwall will be a test of that. Exciting times though.
