Main » 2012 » May » 1 » Prediction countryranking'13 added
Prediction countryranking'13 added
22:59
OK, I have taken the teams as if all leagues had ended today and put them in the acceslist and let the matches start. Basic is that all rounds are using seeding system with the higher coefficient being seeded. The mathc is won by the home team if that's the higher coefficient team and otherwise it ens up in a draw.
Result is the Predicted Countryranking'13 We see there an extreme close fight for first spot, a clear top-9, a clear top-12, a fight for 15th spot that is won by Switzerland over Denmark because of an(other) extreme bad year for Denmark.
The last column is for Ryan, it's meant to be the maximum a country can get if it could choose their best teams on the best spots to collect as many points as possible(all other country allocations have stayed the same).
@Ryan, I was working on the maximum-points-per country. And it is possible, but you can't take it logical. I mean you have to try country by country and switch team by team to see what the consequences are. It's not that you can always use the best team always as lowest CL-spot.
I reached 12th country Turkey. And what I did sofar was to have the 2 best teams for the lowest CL and EL team, to get the furthest, behind that it was first the other CL and then the other EL. This seemed to be the best for Spain and England. For Turkey however it seemed that they were on their maximum. I could not get higher as the original calculations. (that was 7.500) Until I gave Bursaspor(the 5th in ranking) direct accest to CL and put 2 topteams in EL. Bazinga! it went up to 9.300. So question: do you also want to know which configuration leads to this best score? Do you want to use a standard or just try and pick the highest I come across.
This looks great. With this extra column it's easy to see at first glance how close each country is to its potential. England really close to optimal, Spain not so close because of Villareal and Sevilla, and so on for each country. And then you can look at the league page and see why each country is close or far to its maximum potential. I'm satisfied. You can always make more charts and graphs and include more details. If someone was paying you for this service then why not... but just as a hobby this is great. I wouldn't want you to have to take hours of calculations just for one extra column.
I'm curious if the calculation for Holland includes this week's league placements? Looks like the predicted points are close to maximum potential, but with Feyenoord not even predicted to reach the Europa League it's surprising.
In regards to finding the exact combination and placement of each team to reach the maximum potential points, it is a compromise between accuracy and your time. Personally I am satisfied with a close estimate reached by a formula or algorithm or whatever you call the automated decision rules. This way it would be calculated automatically and you would then be easily able to show maximums for all 53 countries. In cases like Turkey's where you mention that trial and error found a higher maximum coefficient personally I can live with knowing that the maximum potential is an approximation.
Another way to resolve your time constraints is to involve other visitors in a collaborative effort. If there's a way for visitors to plug in different teams into CL and EL places and see what the predicted outcome would be then visitors would be able to find the maximum potential by trial and error and you could post it without much effort on your part. Just don't ask me how to technically do this...
Hi Ryan, Actualy the base figures are indeed from before this weekend. I hadn't really played much with Netherlands. Now I did and came all the way to a 11.643. The trick is maybe to put the weakest team in the CL GS (that's what helped: put Feyenoord directly in CL-GS, that's the only way to avoid them being eliminated before GS. Sharing the work is not that easy, as I have to share my sheet with you. It's a coupled set of sheets, not easy to seperate the necessary parts.
I remember you shared your sheets with me and I couldn't figure out how to update them and make changes. At the time we were trying to get a next year's prediction in the middle of a season, and you had to do it on your own... It was difficult for me.
The trick is definitely to put the weakest team in the spot that qualifies directly. By the way, I'm sure Russia is also down a lot this week with Dynamo Moscow not qualified to the EL taking over CSKA's former spot.
From what I sense from what we have so far a team in the right spot or out of EL completely can make a difference of 1-2 points to a prediction. I thought it was more. I wonder what could be the maximum impact. Probably not more than 3 points for one team in real life. The theoretical limit would be around 6-7 but that assumes a very high ranking team with a coeffiecient sufficient to win the CL but part of a country with a low coefficient.
So just to satisfy my curiousity, how long did it take to calculate for 12 countries?
Well if you have seen my sheets, it's easy to tell. I have sorteed Bert's team ranking list per country, so I have the best teams on one sheet. What I have to do then is to go to the sheet with all the leagues, where in front of the teams is put which spot they get. If I change that, automatically on another sheet the country-total is calculated. The difficulty lies espeically in thinking which team should get which spot to get the best result.
Placing a team that would be eliminated before GS directly in the GS delivers (at least) 3 draws in the GS, so nottoo much. I was surprised by the Turkey numbers.
So if you know which teams to put where it's done in a minute or 2, but changing it, checking, changing, checking cost more time.
I'm working on it, made it a bit easier to see right away the consequences of the chances and so. Maybe I have to rework top-12. I noticed by Belarus and Croatia that they are optimal. And the only thing that matters here is that the highest coefficient team gets the CL spot. That team will get into the CL-GS then and the rest is eliminated before the GS.
I always thought that countries that place more or less the same teams in CL and EL every year have a big advantage. But it seems tht for one year the difference is not monumental. But a difference of 2-3 points from optimum every year translates into 10-15 points over a 5 year period, and that is certainly important.
Anyway, I'll be interested to see the maximum potential points for more countries.
I have updated the countryranking '13, at least their maximum. You see often te same numbers. at the bottom countries basic is what they get is a draw per team. with 4 teams that's 4*0.125 = 0.500. When 1 team reaches a next round 0.375 is added to that (0.875) and another round (or 2 teams 1 round) makes 1.250, +1 = 1.625. If a team manage to get into CL-Q3 when it loses it can make another round in EL-Q4, when lost a 0.125 is added. Sometimes it just doesn't matter which team starts when. There is 1 team that reaches a next round, wither in CL or EL , the rest just loses -> 0.875.
Problem a bit is that all this is of course also influenced by what other team qualify. If I update all qualifying teams, I can start allover for all countries again. Maybe I will do if all teams are definite.
I wonder about the Norway prediction being so close to optimum. Must be some mistake with the Fair Play team or something. Molde is predicted to go out in the Q2 round of the CL. That's such a big difference from a team that progresses to Q3 (and gets to play two games there) then at worst gets to have a chance to play for the EL Group Stage (and gets another two games). That's the one remaining big flaw in my opinion in the Platini system. These low ranked champions who drop from Q2 could have been fitted into the EL structure. I think in Molde's case they would get 0.5 points for Norway, whereas another team would get 2.5 points. Rosenborg would get into the EL Group Stage but they are projected to do that from the EL anyway so the points prediction would be the same. I haven't seen Molde play so I don't know how good they are. Would be a shame for a team like that to get a bad draw and go out after two games while other teams from Norway can progress in the EL and play more games.
Unfortunately looks like another meagre season for Denmark. The only round with seeding, unless FC København misses the championship, will be whoever starts in EL Q2. So unless FC Nordsjælland wins the league and goes straight to CL, then DK is only excepted to have one team in the group stages. A FCN championship will probably improve the coef, but as an FCK fan i really don't want that to happen :)
This is because OB has had a terrible season, which is sad after their two EL group stages in the last two seasons. Since their win in Krakow in september 2011, they have been absolutely terrible. They have yet to win a game in the spring, and is not certain of survivival yet. AaB also has a high coefficient, but they quickly threw away all of their CL money, so they were not expected to qualify for Europe. They were five minutes from relegation last season, and are four points from an EL spot now. But even with their seeding i doubt the young inexperienced side would make it to the EL group stage.
The hope is that FC Nordsjælland or the EL Q4 team manages to get a favourable draw and make it to the EL group stage. FCN has a great team, but will lose star player Andreas Bjelland to Twente.
(that's what helped: put Feyenoord directly in CL-GS, that's the only way to avoid them being eliminated before GS.