Forum

xml	No replies possible in the archive

Author: bert.kassies
Date: 05-10-2004, 19:38

So, what was expected by some (see thread UEFA - are pot A teams guarenteed to be at home?), appears to be true. English teams in the UEFA Cup group draw were pre-determined to become "Team B" in the computer draw. I repeat it would be very interesting to know all constraints implemented in the computer algorithm. I can imagine that some teams are not happy with the current match schedule. Maybe next year they should also make claims for a certain schedule.

Do you know more active constraints of the computer draw?

Re: UEFA s computer
Author: scoti Date: 05-10-2004, 20:02	The computer only lets you play home then away or vice versa. I.e you can't play two away games in a row. Also i don't think that the second strongest team in the group can play at home to the strongest. I.e Auxerre hosting Rangers.

Re: UEFA s computer
Author: bert.kassies Date: 05-10-2004, 20:30	scoti, the match schedule is given by the UEFA Cup regulations. The computer only determines which team will be Team A, B, C, D, or E. So, your first remark is not a real constraint of the computer draw. Your second remark contains a strong observation. No 2nd ranked team plays at home against a 1st ranked team. Coincidence, or is this really a constraint?

Re: UEFA s computer
Author: calvon Date: 05-10-2004, 21:14	One more rule to that famous algorithm is that clubs from tha same city not to play home both

Author: Dinamo-Zagreb
Date: 05-10-2004, 22:03

One of the more interesting constrains was tat computer managed not to put too hard or too easy home/away schedules for almost every team. There are exceptions however...

Utrecht - home - Dnjipro, Austria Wien
Dinamo Zagreb - home - Heerenveen, Beveren
AZ Alkmaar - home - Auxerre, Glasgow R.
Aigaleo Athens - home - Middlsbrough, Lazio (although this may be the result of a 'group of death'.

But if you had in mind that solution, that 'mistakes' would happen and you knew it. I tried to find some logic within it, but for now I can't find any. For now it seems that UEFA League is more 'liberal' than CL...

Re: UEFA s computer
Author: Ricardo Date: 05-10-2004, 22:29	Yes, it looked to have been een quite 'open' draw. There is also definitely a division over the first(A-D) and last groups(E-H) Home/away is not arranged: Dutch teams are evely spread, Spanish have 3 away games last day

Author: Ricardo
Date: 05-10-2004, 22:44

If you add the original pot-numbers to the teams in the rgoups order A-E you get:
34512
51234
13524
31245
23451
13524
41352
12345
So th order 12345 is followed 4 times, 3 times is 13524 and D is the exception, but there is Newcastle which had to have B.

There also are a lot A in high/low - D in low/high group couples (high/low division is A-D/E-H):
Standard/Brugge (B/C) - Beveren(G)
Socheaux(D) - Auxerre(F)
Villareal(E) - Bibao/Zaragoza(B/C)
Heerenveen(G) - Feyenoord(A)
=> This might cause the failaing Rangers(F) - Hearts(A)!!
AEK(H) - Panionis(D)

Author: bert.kassies
Date: 06-10-2004, 09:13

Ricardo, your observation of the sequence of the original pot numbers is remarkable indeed. It indicates a pattern. Remains the question if this is UEFA intended behaviour, or just the result of a lousy computer code.

Your remark about the division of clubs in A-D and E-H cannot be due to the computer draw. The computer draw just determined the sequence in the groups. But I did not see the live TV coverage of the draw. Do you suggest that the draw of clubs from pots to groups was biased?

Re: UEFA s computer
Author: Ricardo Date: 06-10-2004, 10:33	I didn't see the TV coverage either, but looking at the results you can be sure that it was biased. It fits for all 11 countries with more than 1 participant. Did anybody see the draw? did anyboody read or heared somthing offical about this?

Re: UEFA s computer
Author: scoti Date: 06-10-2004, 10:50	Ricardo, I don't understand your comment: => This might cause the failaing Rangers(F) - Hearts(A)!! could you explain this?

Author: bert.kassies
Date: 06-10-2004, 11:26

The pieces of information fit together. The observation of scoti and the sequences of Ricardo lead to simple constraints for the computer draw. In fact the draw has very little to do with luck. Maybe at most one random choice in each group.

scoti observed that "No 2nd ranked team plays at home against a 1st ranked team". Written in Ricardo's rank sequences this means that only the sequences 12xxx and 1xx2x can apply. Or shifted versions of this cyclic sequences (e.q. x12xx). So the 2nd ranked team is next in the sequence, or shifted right 2 extra positions.

Ricardo observed two dominant sequences: 12345 and 13524. These sequences are even more general. Any team always plays at home against the next lower ranked team, and (as a consequence) plays away against the next higher ranked team (in a cyclic way think of team 1 as next lower ranked for team 5).

This holds true for all groups, except for the group of Newcastle which was pre-determined to become Team B. But the sequence 12453 in this group is a combination of the NEXT and SKIP2 sequences. The rules above almost apply. With one exception: Newcastle plays at home against Dynamo Tbilisi instead of the other way around. It's not clear to me yet why it was necessary to make this change. Even with the above sequences Newcastle could have been Team B. So there must be another constraint (or a stupid algorithm).

So given the order of the teams the computer only had to allocate which team got e.g. position A. Now again: was that a random choice? Or can we see other active constraints. We need only some more pieces of information for a complete reverse engineering of the UEFA computer code

Author: Ricardo
Date: 06-10-2004, 11:59

Bert, I came to the same conclusion:
Every team plays at home with the team from the next seeding group. For this to work there are only 2 possible numberlists: 12345 and 13524.
In the split(A-D/E-H) each side has 2 of these(except for group D that is an exception. Logically this should be 13524. Without "Newcastle=B" rule I guess Newcastle and Tblisi would have been changed places.
Other rules, I think, taken into account are 'couples'
between the 2 split sides. I found:
AUT C-B
BEL A-D
ESP A-D and C-D
FRA A-B
GER E-B
GRE A-D
ITA E-B
NED A-D and C-B
POR C-E
SCO A-B
All are ahaving on 1 side ACE and on the other DE. 1 Exception:POR, but that is probably because Sporting is in the exception group D.
It even looks like teams on the same split-side are in the same ACE/BD group. Only 1 exception:GRE, they have an A and a D on the E-H-side.

Still the question is how deos this draw work. Does it start with group A assign Feyenoord an A-E, and then next, or are then the couples of group A(Feyenoord-Heerenveen, Schalke-Stuttgart, Hearts-Rangers) taken into account and are that groups now decided.

Author: duncshine
Date: 06-10-2004, 12:15

Can't claim to be keeping up with the genius of Bert and Ricardo, but if the draw was this 'pre-destined' they made a pig's ear of it!

As per Orange's thread elsewhere, they have managed to get two teams from the North East of England to Athens for matches on the same day.

I don't know how close Aigaleo and Panionios' grounds are, but they are playing Newcastle and Middlesbrough on the same day, with thousands of fans travelling over at the same time, and rival fans in Athens all that day.

No-one can condone football hooliganism, but if there are problems in Athens that week (Arsenal are there on the 20th too), UEFA must take a lot of the blame.

Cheers

Dunc

Re: UEFA s computer
Author: Adam(L) Date: 06-10-2004, 12:27	Another coincidence. Clubs from North-East (Poland, Ukraine, Russia, Romania) countries play away in 1st December and do not play at 15th of December.

Author: Ricardo
Date: 06-10-2004, 12:37

Yes, Uefa mentioned that one of the reasons of the computer draw was to take into account that some Nordic countries could not play at home mid-december. Placing these teasm as "E" is not even letting them play at home at all in december.

This was also the problems that caused roup D to be an exception. There are 2 hard locations Newcastle(1) = B and Dinamo Tblisi(5) = E. This is not compatible with either 12345 or 13524.

Re: UEFA s computer
Author: wojtek Date: 06-10-2004, 12:42	I think this system will be proved to be very wrong. Too much importance is put into drawing. If you have two of next to your spot`s teams at home you are like 30 % more likely to succeed.

Author: Orange
Date: 06-10-2004, 12:46

Bigger problem in the new system is that teams will be like the UEFA computer calculating how many points they want. Even more than in the CL.

Final match when both teams need a point I think the game will end in a draw and the team sitting at home could be the one going out.

How silly to come up with a 5-team group. Those "important" figures have made studies about the UEFA Cup with a system like this anyone knows that it's not competitive in the way it should be.

Re: UEFA s computer
Author: bert.kassies Date: 06-10-2004, 13:37	Ricardo, I can hardly imagine that Newcastle is a "hard" location. Might be true for some teams that were 4th or 5th ranked like Dnipro, Tbilisi, Amica, and Zenit (and maybe Steaua). But all these clubs can become "Team E" (and obviously did) by assigning the proper sequence 12345 or 13524. So, I still don't see the need for the exception in the "Newcastle group".

Re: UEFA s computer
Author: duncshine Date: 06-10-2004, 13:51	Bert, I guess the English teams need to be 'B' because matchday 4 clashes with the English League Cup quarter finals on 1st December. Tbilisi need to be E because of hard location. That's why this group must be an exception... Dunc

Re: UEFA s computer
Author: bert.kassies Date: 06-10-2004, 14:43	Duncan, Ricardo, excuse me for this blind spot. Until now I assumed both conditions (Newcastle="B" and Tbilisi="E") could be met by one of the two sequences. Of coarse, that's not the case, and this is the reason for the exception.

Re: UEFA s computer
Author: anita Date: 06-10-2004, 15:01	I thought that Zenith were the only exception when it comes to winter location, and maybe D.Tbilisi, but one thing is rather obvious; the third seeded teams (UEFA Team Ranking) have got a (suspiciously) more favourable draw (home matches) compared to fourth seeded. But maybe accidental and not conspirative G5, since this goes for teams as Zagreb, Utrecht, GAK and Basel.

Re: UEFA s computer
Author: Forza-AZ Date: 06-10-2004, 15:05	On TV they said that Zenit, Amica and Dnjepr were the 3 teams that had the "winter"-restriction. They didn't mention Dinamo Tbilisi.

Re: UEFA s computer
Author: vakho Date: 06-10-2004, 15:13	I am surprised to see Tbilisi as "hard winter" location. Average temperature in Tbilisi during december is around +9 C....

Re: UEFA s computer
Author: anita Date: 06-10-2004, 15:26	Thx, vakho. I just wonder, do Wronki have so low temperatures in December, compared to e.g.Dresden or Praha?

Re: UEFA s computer
Author: mark Date: 06-10-2004, 15:44	Wow. You have done an amazing work of reverse engineering and I wonder if you could be sued under the american DMCA law Now that you all are in such a great mental shape, maybe it is time to tackle the 0.001 issue again

Author: Adam(L)
Date: 06-10-2004, 15:44

I'm sure that avg temperature in Wronki is lower then in Dresden or Prague. The general rule in Central Europe is that the south-west is the wormiest north-east is the coldest. Wronki is about 400km north of Prague, so there are lower temperatures. Polish leagues end their competition in the mid of November while Czech Republic league ends at the beginning of December. So, you may assume Wronki is not a good place to play in December.

Re: UEFA s computer
Author: Forza-AZ Date: 06-10-2004, 16:42	But is it on 25 November? I think I start packing my warm winter clothes already.

Author: ralfinho
Date: 06-10-2004, 23:57

If Tbilisi was not seeded as "E", the sequence "31245" in group D would make no sense at all. Thus I suppose that UEFA regarded Tbilisi as a "winter location". Would not be the first funny UEFA idea.

Bert's statement "We need only some more pieces of information for a complete reverse engineering of the UEFA computer code

" infected me.
For just one hour, I'm looking now at the schedule and the seedings, but I can't find something that's going beyond the aspects Ricardo and Bert already mentioned.

For example, in group H Zenit (pot 5) had to be "E". So, two sequences were possible, 12345 or 24135. I see no reason why the outcome must have been 12345 (as it was) and not 24135. Thus, it just seems to me that the schedule draw really was randomized.

The result was worse than it had been with some additional constraints. Especially the "city clash" in Athens could easily have been prevented by any other sequence in group E than the real one.

Maybe, we guys should make a joint offer to UEFA to develop a better algorithm next year.

Author: viulo
Date: 07-10-2004, 17:29

Boy, I hadn't had so much fun reading a thread since the 0.001 discussion. I considered the possibility that Dinamo's E allocation might have to do with a winter stop in their domestic competition (with UEFA, for instance, avoiding arranging matches for teams that had no local competition for weeks), but Georgia's championship has a round on December 5. Well, they might just be ignorant of the weather...

Re: UEFA s computer
Author: macaskil Date: 07-10-2004, 18:09	It's uncannily like the surrealist panel game "Mornington Crescent" which features in the BBC programme "I'm sorry I haven't a clue". If you hear it the 1st time you bust your brain trying to work out the rules - but it eventually becomes clear - there ARE no rules....

Author: macaskil
Date: 07-10-2004, 18:14

Here's the definitive version -

you play at home on matchday 1 if the temperature in your home city is less than -5 degrees in December, if your club plays within 200 km of another one, if you have a player whose name begins with Z, if your club restaurant serves mousakka and lasagne, if you have never lost to a Scottish team, if you have hooligans who have beaten up fans from at least 4 other countries, and if you are closer to the North magnetic pole than the equator - unless your groundskeeper owns a cat.

Re: UEFA s computer
Author: macaskil Date: 07-10-2004, 18:15	And if your groundskeeper DOES own a cat your coefficient will (if it is a Tuesday) be reduced by 0.001

Re: UEFA s computer
Author: anita Date: 07-10-2004, 22:52	You may be into something there, macaskil. I have written your suggestions down and will check them out. We who have tried to encrypt UEFA's masterminds for some years now, appreciate new ideas. And, as you may observe, we have all got paranoia. Goes with the territory.

Re: UEFA s computer
Author: ralfinho Date: 08-10-2004, 00:12	Oh yes, macaskil, and the answer to the Great Question is "Forty-two".

Re: UEFA s computer
Author: areksen Date: 08-10-2004, 06:35	earlier you were talking about temps in Dresden, Wronki, Prague, Tbilisi... here are some average temps in NOv & Dec Wronki 3.6 0.1 Dresden 4.2 1.2 Munich 2.4 -1.0 (or 4.3 2.5 for Munich-Riem[don't know the difference>) Prague 3.8 0.3 Dneprpsk 1.0 -1.7 Tbilisi 12.8 7.5 (or 7.6 2.9 two readings came up! maybe reading for a Tbilisi state[if there is one>)??

Re: UEFA s computer
Author: areksen Date: 08-10-2004, 06:37	reader friendly version: Wronki.... 3.6, 0.1 Dresden... 4.2, 1.2 Munich.... 2.4,-1.0 (or 4.3 2.5) Prague.... 3.8, 0.3 Dneprpsk.. 1.0,-1.7 Tbilisi.. 12.8, 7.5 (or 7.6 2.9)