Five Tribes Five Tribes

Forums

Recherche
Forums » T2R Competitive Play - English » NC 2012 : Rules and discussion
Montrer: Messages du jour 
  
AuteurSujet
CIA Truckerteller
Senior Member

Pages Perso
Messages: 657
Enregistré(e) en :
October 2007
Re:NC 2012 : Rules and discussion Tue, 27 November 2012 13:16
As far as I'm concerned, the seeding should do two things:

1. Keep UEG and TuS apart in the group round, since UEG has 5 top 10 players and TuS has 2 top 5 players.

2. Keep French, German, US and maybe Italian teams spread out IF that's something these teams/countries collectively want. Maybe people from the US prefer to have some more US teams in their group, I don't know.

For the rest, who cares ? You win some, you lose some. The next 10 teams in a 4 or 5 player seeding can be all be considered nr 3 or nr 12, with fluctuating ELO. I don't think anyone can convincingly claim that some teams are clear nrs. 3-4-5. Group B was as balanced as it was because SNS (seeded 6th from the bottom) and SOS (seeded 9th from the bottom) did a little better than expected, while AMD and Red pretty much sucked. I think you will have something like this every year, and that's not a bad thing. If Rui had won a couple more matches for PRUE (and my fantasy team), there would hardly be any discussion I'd think.

I hope we have enough players next year to go back to a 5 player format and that some teams (I'm looking at you Germany) take a fresh look to get more teams with max 7-8 players.

ps. I think seeding 1-2-3-1-2-3-1-2-3 is better than 1-2-3-3-2-1-1-2-3, especially this year when teams 21 and 22 both went to group C

[Mis à jour le: Tue, 27 November 2012 13:42]

      
CIA Truckerteller
Senior Member

Pages Perso
Messages: 657
Enregistré(e) en :
October 2007
Re:NC 2012 : Rules and discussion Tue, 27 November 2012 13:38
Sysyphus wrote on Sat, 22 September 2012 20:30

The making of the groups was pretty unfortunate this year. 3 French teams together in group B, both Italian teams in the same group too...

Since ERPS'tool is also pretty obsolete -now that we use the top 4 player-, I think we could try to renew the mode of selection, to make it funner and less predictable.

RFAD was talking about an "old Patterson proposal" in the French forum before NC.
After last week's experience, I thought this proposal was making sense and could bring something new.

1)
If 3 groups, top 3 teams in the seeding are divided up into 3 groups.
#1 in group A, #2 in group B, #3 in group C.


2) 1st round of draft :
Seed #1 then chooses a team they want in their group. I.e. : team #22

Seed #2 then picks another one among the remaining ones : Team #18 (for whatever reason).

Seed #3 picks another team, ie : team #21.

2nd round of draft :
In order to make it more random, and also more fun for the community, the team just picked would then choose their next opponent.
To preserve balance, Group C (team #21) would be first to pick in that round, then followed by B and A.

3rd round of draft :

Group A gets first pick, then group B etc...


It'd not be that hard to organize, we would just need one player of each team for the draft night.

Would solve the non-representative ELO issue, the ERPS' tool issue and the unfortunate groupings.



Bad proposal. Who's to say team A, B and C (presumably the three strongest teams, but that's debatable to begin with) will come up with a balanced draft ? Maybe Elric is in a particularly funny mood picking the strongest 6 opponents for Red to all play in a group together. Maybe teams vote because they don't want to meet / have to schedule with players they realy don't like.
      
SY/\/ \/\/ill
Senior Member
Maître des citrouilles

Messages: 382
Enregistré(e) en :
June 2010
Re:NC 2012 : Rules and discussion Tue, 27 November 2012 13:41
Pretty much agree with trucker, but im sure he meant SNS, not UEG. Smile

If you're going to use ELO there was one easy way of adjusting it this year...

Why use top 4 ELO for a 4-player clash? You didn't use top 5 last year... Use top 3 - that's all you need to win a clash. Shows the strength and chance of you winning more accurately in this format against smaller nations like mine where myself, Steve and Robin are playing every match. That figure more accurately shows our potential.

Or take an average of all the players in your team, but this is less accurate as the strongest players will play more frequently I guess...

I haven't got time to add them, but maybe top 3 gives more accurate seeding?
      
CIA Truckerteller
Senior Member

Pages Perso
Messages: 657
Enregistré(e) en :
October 2007
Re:NC 2012 : Rules and discussion Tue, 27 November 2012 13:48
I would argue for the opposite, a top 6 seeding or average when n<6 If OLE, TuS, CAT, decide to play with a broad range of players, the seeding should be based on more than 4 players. Since we don't know who is going to play how many games and this may depend upon the standings at the time, who knows.
      
AAA_dea1
Senior Member
Vainqueur Nation Cup AdR 2007

Messages: 1928
Enregistré(e) en :
September 2005
Re:NC 2012 : Rules discussion Tue, 27 November 2012 14:31
OLE Masimo schrieb am Tue, 27 November 2012 12:42


Dea - if this captain/rating system is better than ELO - doesn`t that mean that ELO is not correct reflecing the players strength?


Yes, that's what it means.

As you say yourself "NC is special".
Why is it so special?

Reason #1: Because every tournament is different from casual play.
People take more time to think about their moves, you make sure you have good playing conditions (play with the PC you are used to, make sure you are undisturbed, ...) and that you are in good shape (get enough sleep, don't drink alcohol, ...). People probably don't try new ideas in a tournament game.
For some players the difference may be less than for others.
I consider mine pretty huge, and I think my tournament results are a lot better than my average ELO.

Reason #2: In NC you have tons of spectators, and you have your team expecting a good performance. If you mess something up, everyone sees it, you have to apologize to your team, you may get some "funny" comments in the lobby. In other words, the pressure is much higher than it usually is.
Some people can cope with pressure quite well and may even get better under pressure, others don't.

Summary: Yes, I state that ELO is not correct in reflecting a player's tournament strength and especially a player's NC strength.

OLE Masimo schrieb am Tue, 27 November 2012 12:42


Then why do we use ELO at all? We could do a weekly survey on who to be #1 Cool


Honestly, being among the top 10 players in such a survey (which I wouldn't be) would be worth more to me than being ELO #1 (which I was).

Anyhow, it depends what you expect from it. Reaching #1 is not easy, you're certainly a good player if you achieve it. But no, I don't think it automatically means you are the best player for as long as you hold the position.
A player with ELO 1700 is certainly a better player than somebody with 1100.
But - in the upcoming multi tournament I could easily win or lose 100 points ... will my skills have increased or decreased afterwards? I guess not.

I think, ELO gives a good hint at the average casual play strength of a player (as opposed to tournament strength).
I don't think it can be used to compare 2 specific players, unless the ELO difference is really huge.
      
Sysyphus - Pommard
Senior Member
Vainqueur AdR All Around Tournament 2011

Pages Perso
Messages: 2409
Enregistré(e) en :
December 2007
Re:NC 2012 : Rules and discussion Tue, 27 November 2012 15:16
Truckerteller wrote on Tue, 27 November 2012 05:38

Sysyphus wrote on Sat, 22 September 2012 20:30



2nd round of draft :
In order to make it more random, and also more fun for the community, the team just picked would then choose their next opponent.
To preserve balance, Group C (team #21) would be first to pick in that round, then followed by B and A.





Bad proposal. Who's to say team A, B and C (presumably the three strongest teams, but that's debatable to begin with) will come up with a balanced draft ? Maybe Elric is in a particularly funny mood picking the strongest 6 opponents for Red to all play in a group together. Maybe teams vote because they don't want to meet / have to schedule with players they realy don't like.


Bad proposal because you read it bad Wink !

Pretty much like what you propose. 3 teams are top 3 seeds.
Following your idea, UEG, TuS to which you add the winner or the runner-up from previous NC.

Then they start the draft of their group by picking their first opponent, probably a weak one, for instance teama #21, #22, #18. When each group has picked an opponent, then teama #21, #22, #18 have to pick a team in their group.

Quite close to top 3 teams and then "who cares" that you propose + it has the advantadge to add the perception factor of the strength of a team.
If you add some constraints (number of teams of the same country by group for instance), then you have the flexibility that rigid ELO can not offer.

Moreover, since EPRS' tool is obsolet, the seeding is not readable nor transparent...

Quote:


and SOS (seeded 9th from the bottom) did a little better than expected, while AMD and Red pretty much sucked.



When I'd agree with "For the rest, who cares", I had to comment on this.

You then have the great role. SOS is seeded #13 and gets the qualification, so they played way better than AMD seeded #5 (with me, Drake and onyx (!) way overrated for whatever reasons), not qualified, who sucked because they were one match shorter than the team 1st in the group !

Yes, we didn't play well overall... but SOS with a similar lineup (and many would think stronger with Bean, you and Patterson, against Drake, ommie and me...) won one more match and less games than AMD, and you say that you did better than expected and that we sucked in RR ?

In the end SOS sucked as much as AMD !

[Mis à jour le: Tue, 27 November 2012 15:18]

      
CIA Truckerteller
Senior Member

Pages Perso
Messages: 657
Enregistré(e) en :
October 2007
Re:NC 2012 : Rules and discussion Tue, 27 November 2012 15:19
Sysyphus wrote on Tue, 27 November 2012 15:16



In the end SOS sucked as much as AMD !




Yes, especially when you add the round of 12
      
Elric - Sancerre
Senior Member
Vainqueur Nation Cup AdR 2005

Pages Perso
Messages: 1619
Enregistré(e) en :
May 2005
Re:NC 2012 : Rules and discussion Tue, 27 November 2012 15:38
*** Dietlind, Martin, Sven, Julien... come on kids, it's school time again ! *** (training for Madame Capi, lol)
      
AGT-DN
Senior Member

Pages Perso
Messages: 833
Enregistré(e) en :
November 2006
Re:NC 2012 : Rules and discussion Tue, 27 November 2012 17:28
OLE Masimo schrieb am Tue, 27 November 2012 11:38

More interesting to me here would be to see how the tied games might have changed the picture in case these games would have been bo5 as played in previuos NCs - would that have changed the picture?


YES, especially for OLE,WASA & SNS!

The reason is not only that tied games are missing here but even more that tied matches now have a winner (the one that was leading 3-0/3-1/3-2):

Assuming the matches would have been best-of-5 only and there would have been no ties in games, following clashes would have ended a different way (each clash half a clash point in difference to our tables, except for AGT-TIC where the clash result totally would have turned):

A:
UEG vs SMP : 2 - 2
PRUE vs OLE : 2 - 2
PRUE vs SMP : 2 - 2
OLE vs White : 2 - 2
PRUE vs FC : 2 - 2
OLE vs SMP : 1 - 3

Difference in Clash points:

UEG   -0.5
CAT     =
OLE   -1.5
SMP   +0.5
FC    +0.5
PRUE  +0.5
White +0.5

New table and difference in places:

CAT   5.5 18-06 +1
UEG   5.0 19-05 -1
SMP   3.0 12-12 +1
FC    2.5 09-15 +1
PRUE  2.0 10-14 +1
OLE   2.0 09-15 -3
White 1.0 07-17  =



B:
AMD vs SNS : 3 - 1
Red vs WASA : 3 - 1
AMD vs SOS : 2 - 2
AMD vs Champ : 2 - 2
Red vs AMD : 2 - 2
Red vs SNS : 2 - 2

*** special case: AMD vs WASA : 1 - 3 (06-10)*** two ties in Vballman20-kostasss, so only 2-2 - WASA would have won the clash assuming kostasss would have won the outstanding game

Difference in Clash points:

WASA  -0.5
SOS   -0.5
Champ +0.5
SNS   -1.0
AMD     =
Red   +1.5
Rosé    =

New table and difference in places:

Champ 4.0 14-10 +2
Red   4.0 14-10 +4
SOS   3.5 15-09 -1
AMD   3.5 13-11 +1
WASA  3.5 10-14 -4
SNS   2.5 11-13 -2
Rosé  0.0 05-19  =



C:
TUS vs AT : 2 - 2
TIC vs ITA : 2 - 2
TUS vs ITA : 2 - 2
AGT vs YWN : 3 - 1
AGT vs TIC : 1 - 3
AT vs YWN : 2 - 2
TIC vs YWN : 2 - 2

Difference in Clash points:

AGT -0.5
BIP   =
TUS -1.0
AT    =
TIC +1.0
YWN -0.5
ITY   =
ITA +1.0

New table and difference in places:

AGT 6.0 21-07  =
BIP 5.5 19-09  =
AT  4.0 17-11 +1
TUS 4.0 15-13 -1
TIC 3.5 14-14  =
YWN 2.0 11-17  =
ITY 2.0 09-19  =
ITA 1.0 06-22  =  
      
CIA GenuineFauxFarm
Senior Member

Pages Perso
Messages: 333
Enregistré(e) en :
February 2009
Re:NC 2012 : Rules and discussion Tue, 27 November 2012 18:00
While I appreciate the discussion and the reasons for it all. Let me remind everyone of three things that keep events like Nation's Cup strong.

1. A clear set of rules.
Regardless of how the rules fall, if they are clear, we can all abide by them. We can decide how we will strategize for them and we can each have good or bad luck depending on how they are written.

2. Consistency in rules and playing environment over time.
While it is important to respond to real issues, it is a good idea to not change the rules too often or too drastically. Otherwise, you may change the character of the event in a way that may remove interested persons. Multiple changes also reduce the clarity of the rules in the collective mind of the group (see #1).

3. A community of interested people that care about the event.

Opinion regarding seeding - Keep It Simple

Rob
GFF
      
Sysyphus - Pommard
Senior Member
Vainqueur AdR All Around Tournament 2011

Pages Perso
Messages: 2409
Enregistré(e) en :
December 2007
Re:NC 2012 : Rules and discussion Tue, 27 November 2012 18:43
I do agree with you Rob here.
However, we were at a turning point this year, so drastic changes had to be made, which overall were quite successful and assimilated considering the conditions (all made in a hurry but with a strong discussion and understanding from everybody).

If this year was the year of change, we can now adapt/correct the format and make it stronger so that it can last a few more years. Change often needs consolidation.

3 points in my opinion that need to be discussed, rest are details (some have been already discussed) :

a) size of a team : I'd limit a team to 6 or 7 players to promote consistent lineups for a consistent competition.

b) if we keep ELO in the seeding, broaden it. Best 5 ELO in a 6/7 player team.
Keep it simple ? Sum of best 4 is not simple unless you have an Excel sheet. We don't have a clear ERPS' tool like before.
If we change the seeding, bring flexibility to avoid same country clashes, and/or add some players' opinion with a draft.

c) Keep a consistent format. Either on a long shot, keep 6 games in RR and bo7 in KO. Or keep a consistent format all through NC, i.e. keeping 6 games in KO too.

[Mis à jour le: Tue, 27 November 2012 18:44]

      
SY/\/ \/\/ill
Senior Member
Maître des citrouilles

Messages: 382
Enregistré(e) en :
June 2010
Re:NC 2012 : Rules and discussion Wed, 28 November 2012 02:39
OK, putting my sensible hat on now: Christmas

Sysyphus wrote on Tue, 27 November 2012 17:43


we were at a turning point this year, so drastic changes had to be made, which overall were quite successful.


Personally, as I have already voiced, I think the change for smaller teams has benefitted the competition with more 1-nation entries which is a good thing. For me, I really detest that I can draw a game, for example against Trucketeller. He may be jumping for joy at the prospect of not losing against a superstar Wink but for me there is no fun in the 2-2 outcome. I don't mind a nation vs nation clash to be a draw, but please go back to a best of 5 format in the round robin phase. Sven, out of interest, did you feel happy/sad/indifferent after a 2-all draw against me? I'd have rather played a decider, giving myself a chance to win, and lost to 42 + 10 locos and start then settled for a draw... Is that just me? Doesn't seem to be much feedback on this point...

Sysyphus wrote on Tue, 27 November 2012 17:43


a) size of a team : I'd limit a team to 6 or 7 players to promote consistent lineups for a consistent competition.

Couldn't agree more. SNS were forced to play with 5 players for a 4-player clash. TuS had 10, why weren't they forced to play with 2 teams of five... Don't get me wrong I have really enjoyed my matches but we've had scheduling problems quite a lot and one week I could have done without playing...

Sysyphus wrote on Tue, 27 November 2012 17:43


b) if we keep ELO in the seeding, broaden it. Best 5 ELO in a 6/7 player team.Keep it simple ? Sum of best 4 is not simple unless you have an Excel sheet. We don't have a clear ERPS' tool like before. If we change the seeding, bring flexibility to avoid same country clashes, and/or add some players' opinion with a draft.

Asking for complication. I like that idea mentioned earlier of top 3 seeds picking groups, but obviously who the top 3 are is open for debate. SNS have got further than TuS this year. Just thought I should mention that again! Obliterated them. Well done to Felix for winning a tough match, I had plans to kick his backside later on in the tournament though, shame... I think the top 3 seeds (or 4 if their are more teams/groups) should be this years' winner, runner-up, both losing semi's teams, and play a bronze medal match between the semi's losers if just 3 groups... Takes any ELO nonsense/arguments out of the picture and I don't think luck can get you to the finals (no comments about ukb please!)

UKB were seeded 7th last NC, the two teams with their players are both still in the competition with a chance. A month or two later... (when only 5 players had 1700+)
http://img198.imageshack.us/img198/425/picture17qd.png
There's no way we were accurately seeded for NC by ELO, use the previous years' results as a guideline. It's unlikely a few new superstars will emerge and start playing next year, so it should give a reasonably accurate result. It's also a privilege/reward for teams that did well. If SNS go out I wouldn't mind if the teams in semi's/finals chose groups according to the method Sysy mentioned... If Elric (not sure why the example??) wanted to choose a tough group then let him. His team would've earned the right to. Don't worry, he won't have the chance next year after Wink

Sysyphus wrote on Tue, 27 November 2012 17:43


c) Keep a consistent format. Either on a long shot, keep 6 games in RR and bo7 in KO. Or keep a consistent format all through NC, i.e. keeping 6 games in KO too.

Already mentioned that I didnt like round robin draws... better last year in that respect... I'd really not like to see it in knockouts either... Smile

Trucker wrote


I would argue for the opposite, a top 6 seeding or average when n<6 If OLE, TuS, CAT, decide to play with a broad range of players, the seeding should be based on more than 4 players. Since we don't know who is going to play how many games and this may depend upon the standings at the time, who knows.

Your maths teacher wouldn't be happy! I assume you mean n>6 Laughing For a small team, Me Robin and Steve are playing every match. A smaller ELO sample gives you a much clearer representation of what you are up against when playing us. Perhaps with a team of 10 like TuS this doesn't, as they will be playing probably 2 or more of their other players each week... but the smaller ELO sample would at least give you a more accurate viewpoint of their potential maximum strength... Ie the chance of them winning if they get to the final. I'd still rather use a method of perceived strength to get rankings, or preferably my method based on last years' results. You wouldnt need to spend hours individually ranking all the teams either...
      
SY/\/ \/\/ill
Senior Member
Maître des citrouilles

Messages: 382
Enregistré(e) en :
June 2010
Re:NC 2012 : Rules and discussion Wed, 28 November 2012 02:48
Sorry, that turned into an essay... In summary...

1. I like Sysy's/Pattersons proposal of top seeds picking groups
2. Make things simple. Teams who win/get to semi's this year are top seeds next year. Really simple, results based! Everyone knows what'll happen next year, no need for any ELO/ranking etc.
3. No ties in a match!

Surely they are really easy changes that could be implemented...

Edit: I just noticed that my profile now looks like Sean Connery is holding a pumpkin hostage! Smile

[Mis à jour le: Wed, 28 November 2012 02:49]

      
SY/\/ \/\/ill
Senior Member
Maître des citrouilles

Messages: 382
Enregistré(e) en :
June 2010
Re:NC 2012 : Rules and discussion Wed, 28 November 2012 03:55
This is a FUN idea that could make the line-ups very interesting! It could give a balanced set of groups and at least guarantee that some funny clashes happen. I was almost upset to hear TuS disappeared before we got to them! Wink Slightly different variation of the Patterson/Sysy idea. Pretty much like what they proposed. 3 teams are top 3 seeds. I don't think it'll catch on, but it would be amusing certainly! Any takers? Even if you don't like this, I still really like my idea of this years' top teams being seeded top next year and using some sort of method for them to select groups. Removes all ELO crap...

Group A - seed #1 2012 NC Winner
Group B - seed #2 2012 NC Runner-up
Group C - seed #3 2012 NC Bronze


#1 picks a team that has to play in group B
#2 picks a team that has to play in group C
#3 picks a team that has to play in group A

Then return the favours...
#3 picks a team to play in group B
#2 picks a team to play in group A
#1 picks a team to play in group C

Rotate which team picks for where etc.
i.e. Seed 3, UEG, decide to put TuS in SNS's group.

makes things a bit more fun, creating bad-ass clashes in the group stages! If you don't want to meet your fellow countrymen, you put them in other groups... Prioritise Smile I think the original picking teams for your own group will be more popular, but I thought I'd at least mention the idea...

[Mis à jour le: Wed, 28 November 2012 04:12]

      
onyx puffin MAD
Senior Member

Messages: 977
Enregistré(e) en :
January 2005
Re:NC 2012 : Rules and discussion Wed, 28 November 2012 04:25
Interesting assumption that for NC 2013, there will be 3 groups.
Remember we had 22 groups this year, hmm .. rules state:

17-22: Round robin, 3 groups, with QF, SF and F.
23/24: four groups (1,8,9,16,17,24)(2,7,10,15,18,23)(3,6,11,14,19,22)(4,5,12,13 ,20,21) - groups split according to ELO team rank.

Maybe a discussion on what we do if more than 24 teams. Shocked

By the way, I am one of AMD writing with opinions. AMD may not have made it thru the Round Robin, but I think we might win team with most opinions.
Also of note, I am in agreement that any system for selecting a ranking is flawed, and really question ELO factors. In previous years I have maintained AT was ranked too low because their top 3 were so high, and the others were not; but all you needed then was a top 2 that won consistently and then win 1 of the other 3. So using top 6 ELO was bogus as well. My real opinion is all ways of choosing a grouping is questionable so live with whatever rules are set. And yes, perhaps that sets me not on same page as my teammate about group choice factors. But what is new about teammates disagreeing. And his note on my ElO being higher than true to tournaments is a fair statement since my ELO is built from multi games.

[Mis à jour le: Wed, 28 November 2012 04:47]

      
stemayf
Senior Member

Pages Perso
Messages: 352
Enregistré(e) en :
May 2006
Re:NC 2012 : Rules and discussion Wed, 28 November 2012 13:16
Hi,

Please correct me if I'm wrong.

If WASA Geoff's ELO (was it Geoff?) had been taken into account prior to the seedings, would not WASA have been seeded 6th overall? Therefore in group A? Therefore transforming the makeup of the other groups?

Perhaps Group B would not have been so tight and have been less likely to present upsets.

Why were WASA seeded 11th? Why did we group prior to the finalisation of the seedings?

Steven


      
Sysyphus - Pommard
Senior Member
Vainqueur AdR All Around Tournament 2011

Pages Perso
Messages: 2409
Enregistré(e) en :
December 2007
Re:NC 2012 : Rules and discussion Wed, 28 November 2012 15:32
SNS stemayf-Thunderballs wrote on Wed, 28 November 2012 05:16



If WASA Geoff's ELO (was it Geoff?) had been taken into account prior to the seedings, would not WASA have been seeded 6th overall? Therefore in group A? Therefore transforming the makeup of the other groups?



Yes indeed, but Geoff could even have more ELO points than right now.

If ELO is an instant photography : who thinks that onyx puffin with 3698 is a better 2er player than Geoff with 3610 ? Onyx just boosted his USA score by playing multis.

But you also have the case with FC, who would have been seeded 18th instead of 12th, without wasser and pacmon who didn't play in group A.


Maybe we should just take into account the 2er score as a reference in the ELO seeding, which is closer to perception of one's strength. Multi map players are way less numerous than multi players, so it'd affect a smaller range of players. And they are less likely to boost their ELO : easier to inflate/deflate your USA score with multis, than inflating your 2er score by playing an alternative map.

[Mis à jour le: Wed, 28 November 2012 15:41]

      
AAA_dea1
Senior Member
Vainqueur Nation Cup AdR 2007

Messages: 1928
Enregistré(e) en :
September 2005
Re:NC 2012 : Rules and discussion Wed, 28 November 2012 17:29
My view on the format:

- 6 games in RR is very good & I like the possibility of draws
--- if you can't win by more than 1, is that really a justified win?
--- 6 games instead of bo5 means everybody gets the chance to play 6 games. With bo5 your match may be over after 3 games - that's frustrating as we participate because we like to play (at least for the loser). Additionally, as games won are a tiebreaker, I want to have a right to play all games possible.

- I'd like to play 6 games in KO, too
--- see the reasons above
--- the possibility of draws in matches reduces the amount of draws in clashes. I think that would be good, because I think the tiebreaker match is the worst idea we ever had.


- I'd like to change to 6 games in KO + games won as tiebreaker
--- the tiebreaker match puts too much emphasis on a team's best player
--- the tiebreaker match can be a timing problem
--- If games won are important the performance of weaker players will be more important and thus more valued (instead of "you lose, you don't count")



- for the seeding I prefer anything that's "non ELO"
--- variants mentioned already
--- Will's idea would be fun, too
      
ATN Drake
Senior Member

Messages: 812
Enregistré(e) en :
March 2006
Re:NC 2012 : Rules and discussion Wed, 28 November 2012 18:49
OLE Masimo wrote on Mon, 26 November 2012 01:05


I am not sure I understand the whole thing .... Drake and Sysy are trying to make us believe that they are out of the game cause to bad grouping?

I looked up their this years NC stats, Drake is up with 1 win, 2 losses and 2 ties, Julien with 1 win, 2 losses and 3 ties (sorry for warming that up Twisted Evil ) - I would say that just isn`t enough to make it to KO phase unless the rest of your team can make up for that.




I am not going to say me and Sys played well, but maybe the reason our records were so bad is because OUR GROUP WAS HARDER!

And the rest of our team DID make up for it. We won 3.5 clashes, but it was not enough to make it to the KO phase. SMP won 2.5 clashes and it was enough to get to KO. We even did better than OLE as a 3rd seed with more matches and games won. Again, our group was harder.

BUT that's NOT the point. I don't care that our group was harder or that we didn't make it to the KO round. The point is, I want the formation of the groups to not be RANDOM. ELO is more or less random - I can be anywhere in the top 10 to top 100 depending on what I'm playing. If 1 or 2 other people on a team are similar, then ELO is meaningless.

OLE Masimo

The grouping itself is not the problem - the question is whether the ranking is correct, ie. whether a 7th ranked team is on its apropriate place or whether it should be ranked differently.


This I disagree with. I do NOT care if groups are uneven, as long as they are uneven for a tactical reason (i.e. they were drafted that way). So the problem is the grouping.

And people against drafting say ELO is the "easiest". It is WAY harder and TIME CONSUMING to calculate than a simple 3 minute draft. You have to log everyone who is on a team, calculate 2P + USA scores, we have to argue whether top 3-4-5 members of team should be used.

OLE Masimo

because NC is special - I saw soooo many upsets - just scroll through the history of NC ... In NC it may happen to all of us that be get beaten by underdogs.


I like the idea of upsets and underdogs winning clashes, BUT all those upsets might not really be upset if the seeding is wrong. I mean SOS beating AMD on paper was an upset #5 over #2, but in reality it probably wasn't. IF AMD was drafted as a #2 seed and SOS a #5 seed, then I would say it was more of an upset. But I bet if teams were drafted SOS would be a higher seed and AMD lower.

For me personally it would be more fun and more incentive to do well. If there was a draft and AMD got drafted as a #5 seed, I would like to prove people wrong and try to do better than place 5th in our group.

      
Cromze - Pouilly-fumé
Senior Member

Messages: 372
Enregistré(e) en :
September 2009
Re:NC 2012 : Rules and discussion Wed, 28 November 2012 20:28
WHITE team report (mainly RFAD, Tokar and my opinions) :

- 6 games / BO5 / BO7 :

For RR we're all for 6 games match, much better than BO5. Same reasons as Dea, see 2 posts above. We all think that reform was essential.
For KO and final, no definitive opinion. Let's see the end of the tourney.


- Group composition / seeding :

RFAD is 100% for the draft : no seeding problems, more fun and surely it would be interesting to analyze the different choices.

My opinion is that having harder and easier groups is not a problem and whatever the rule is it will be impossible to have a total equilibrium. Any tournament has a "death group" right ?
The main problem is having too much teams from the same country in a group. One more time whatever is the rule we won't solve this problem. Maybe it could be better to change the ORDER of the clashes : teams from the same country should play at the beginning of the RR, then we avoid the RED vs CHAMP situation we had this year. I don't know if it's possible to include this point in the rules...

I just learned above that ELO seeding was calculated with the USA score and not 2er score ! Anomaly ! If we keep the seeding rule it must be with the 2er score as reference !

With that change I think that ELO seeding or draft are both okay, but draft is finally more attractive cause it's more fun. Will's idea sounds nice too.
==> WHITEs are for the draft


- TD / hierarchy :

All OK. Good job guys !


I may add another post later about team sizes and ELO seeding if we keep it (best 3/4/5/6 ELO). I have to discuss these points with my team first.

Greets
Cromze
      
Mr Bean
Senior Member

Pages Perso
Messages: 788
Enregistré(e) en :
May 2006
Re:NC 2012 : Rules and discussion Wed, 28 November 2012 20:55
Cromze - Pouilly-fumé wrote on Wed, 28 November 2012 20:28

I just learned above that ELO seeding was calculated with the USA score and not 2er score ! Anomaly ! If we keep the seeding rule it must be with the 2er score as reference !




Actually the seeding is done on the added totals of your USA score + your 2-player score Cool
      
OLE Masimo
Senior Member
Vainqueur FEUd Cup AdR 2009

Pages Perso
Messages: 809
Enregistré(e) en :
November 2004
Re:NC 2012 : Rules and discussion Thu, 29 November 2012 11:09
Drake Storm schrieb am Wed, 28 November 2012 18:49

OUR GROUP WAS HARDER!


ok ok - to make you feel happy I agree ... (still not knowing why you`d have prefered Tus, Ueg, Cat or AGT in your group - or what makes u so sure your team would have beaten us but not SOS)


When I look at the data provided by DN above (thanks a lot for that, Dom, really interesting) I feel like you`re out rather to a rule change than to the grouping. Playing 6 games instead of bo5 kicked you and red and saved us!
This rule was to our favour, indeed - but I did not vote on it ... you made this up and once there is a rule you have to live with it. I am no real fan of playing 6 games, but a point for it may be that it favours competition players, that don`t give up after a sure loss but turn the game to a tie vs those that give the win away ... anyway - I don`t think that discussion leads anywhere.

I very much agree with Dale, any way of choosing a grouping is questionable! And whatever will be agreed on will be questioned soon after. I have no problem with that - what disturbs me is that at the end some cry "we would have done much better IFFFFFF ..."

And I think we have to face that any rule change will favour some teams and be disadvantageous to others (as were the 6 games this year) - thats something we have to live with.


I was surprised by Wills idea - that sounds like fun Very Happy - finally you will know who your REAL friends are Twisted Evil - but seriously - why not try something like that - gives a bit of tactical touch already to the grouping. And teams really have to evaluate how strong other teams might be - nice ...
But I think the teams that got picked in the first round should get the chance to do the next pick instead of the three/four top seeded teams picking ALL their opponents (of course then picking rd 4 be done by the teams picked in rd3 and so on) - but well in the order supposed by Will.
Else all the teams will just have to sit and wait what the "stars" come up with.
      
AAA_Angel6
Senior Member
Vainqueur Nation Cup AdR 2007

Messages: 1035
Enregistré(e) en :
November 2004
Re:NC 2012 : Rules and discussion Thu, 29 November 2012 12:53
Imho, DoW didn't do a very good job when implementing the new rankings.
In the 2 player ranking, the different maps are mixed.
In the US ranking, 2, 3, 4 and 5 player games are mixed.
So, when we are taking US ranking and 2 player ranking, the best seeded player might be a 5 player US spezialist, and a 2 player EU spezialist, who never played a single US 2 player game Twisted Evil
What we really wanted for NC seeding is a US 2 player ranking, which we can't get at the moment.

Another thing is unsteady rankings. When a strong player only does serious tourney for some time, he reaches 1800. Then he plays "drunken fun" for some time and drops to 1600. However, for NC ranking, we should take 1800, because that's the level he will play in NC. So I would do an "ELO history curve" for every player, then lay a slightly dipping tangent on it, and the end point (of the tangent) for todays date is the ELO value I would take for seeding.
E.g. the line drops 50 points a year, and a player had 1800 a year ago, and 1600 now. Then I would take 1750 for seeding!


For "my perfect ranking system", I'd see 2 possible solutions:

A) DoW gives me access to the game results database and I'd calculate the real rankings.

B) Someone writes a script for reading ALL game histories (including player ID, player name, date, map, result and if ranked) and gives me the resulting database.

I love statistics, so I'd really like to get all the game history data Very Happy

[Mis à jour le: Thu, 29 November 2012 13:05]

      
GSV3
Senior Member
Vainqueur Nation Cup AdR 2012

Pages Perso
Messages: 673
Enregistré(e) en :
May 2009
Re:NC 2012 : Rules and discussion Thu, 29 November 2012 13:11
AT_Angel6 écrit le Thu, 29 November 2012 12:53

I love statistics, so I'd really like to get all the game history data Very Happy


Let's check it Surprised

What's the probability for each team to win this NC?
      
SY/\/ \/\/ill
Senior Member
Maître des citrouilles

Messages: 382
Enregistré(e) en :
June 2010
Re:NC 2012 : Rules and discussion Thu, 29 November 2012 14:42
My method... The proper sensible one...
Finalist teams from this year will surely have played the best in the competition. They get to choose next years' groups. Think of it as a reward to the best performing teams. It actually gives very little or no advantage to them but a little tactical fun - groups should really be similar strength but you can avoid or fight your enemies depending on your mentality. They should be no more 'unfair' than this years' distribution.

The system is without individual player ratings, any hassle and discussion about ELO being ineffective, and any players who are drastically underscored (e.g haven't played since ipad changed the rankings considerably). We can save hours upon hours of adding up scores to calculate inaccurate seedings, and you know what is happening next year! After entries are in it'd take no time at all to select the groups! The details for numbers of teams/groups could be fine-tuned like in current rules, but the idea is just so nice and easy! Sysy really liked it even though he hasn't replied yet... We should assume (as an example) that we can get four groups of 6 next year when the Germans split up a little etc, but obviously it makes little difference if there are 3 groups. Bronze is quick match of some sort between losing semi-finalists.

Group A, Winner
Group B, Runner-up
Group C, Bronze winner
]i]Group D, Bronze loser[/i]

There are then two different methods of these seeds choosing teams, either choosing a team to play in your group (original Patt/Sysy idea), or choosing a team to play in another group (my slightly tactical fun idea, see previous post). Either way, in the first round of picking, Group A should have the advantage (either choosing a team for their group first, or being the last group to accept a team from another groups nomination, ie group D wants TuS to play in group A)

Going back to Patterson's/Sysy's idea, say group A chooses AGT, then next turn AGT chooses... This is adding more complication, you need a representative from each team online at the same time. I would suggest that the four seeds would sort out the groups by themselves, keep it simple!

However you do the group choosing afterwards is not something I am hugely interested in. I just believe using last years top teams as next years' top seeds is just so incredibly easy! I can't see any team moaning - and if they do... then they should win NC to get first pick next year Wink No debate! Team selection will be based on perceived strength. If TuS think they should be top seeds, well we have a few results (grats OLE) to say otherwise! Let the four seeds choose the groups - they should come out even! As I mentioned, if SNS were knocked out I'd have no problem with being chosen in this way.
http://www.thepinksweetshop.co.uk/lg_images/Meerkat_Jelly_Sweets.jpg

P.s. Vincent, due to UKB's immense strength and TuS disappearing...
49% SNS/49% WASA/1% AGT/1% UEG or AT
      
SMP-JenAck
Senior Member

Pages Perso
Messages: 456
Enregistré(e) en :
February 2008
Re:NC 2012 : Rules and discussion Thu, 29 November 2012 14:47
just a small point for the ELO"thing"

the last 3 year (since i play NC)
we play OLE,PRUE,UEG and CAT twice or more in RR.

i would like to play also other teams wich we never played,
so i`m interest at any idea to change that.
      
Cromze - Pouilly-fumé
Senior Member

Messages: 372
Enregistré(e) en :
September 2009
Re:NC 2012 : Rules and discussion Thu, 29 November 2012 14:58
About grouping :

Thx for the precision Bean there's still some points I ignore...

I agree with Angel : we don't have the good ranking information for a 2 player USA map tourney. Since we don't have this statistic, grouping with ELO is questionable and there would be some "upsets for bad reason".
Seeding done on the 2er score ONLY would be better than the current formula (Sysy explained above : there's more multi players than alternative maps players) but still not perfect.

Now let's imagine we have the USA 2 player statistic. I guess we would not be in a definitive safe place. What would happen if we have another Ipad inflation, this time during the NC grouping phase ? More generally, who thinks the ranking system is totally safe and perfect ? It is surely good, not perfect...

RFAD/Patterson draft would be a solution except for one problem : calculate the top 3 seed !
Then Will's idea solve all the problems : replace top 3 seed by previous NC top 3.
There's just one point I dislike with Will's proposal, which is giving absolute grouping power to top 3. Let's have a mix between Will's idea and RFAD/Patterson draft, like Masimo just suggested.

This is a personal opinion, I still have to add another WHITE report about the other points.
      
Sysyphus - Pommard
Senior Member
Vainqueur AdR All Around Tournament 2011

Pages Perso
Messages: 2409
Enregistré(e) en :
December 2007
Re:NC 2012 : Rules and discussion Thu, 29 November 2012 15:34
Why would gather one person of each team would be that hard for a draft ?

Instead of using the perception of 3/4 persons, we would have a mixed perception at least. And why would the perception of that 3/4 persons be more "accurate" than some other players in weaker teams -there I agree with Martin :p- ?

We manage to gather one person of each team for both QT! It's then not impossible ! Will's idea is good because it goes in the right direction according to me because it's fun and uses perception over ELO.

Lobby is at his best when crowded for an event.

Keep things simple ? Have you ever attended to CL's draft ? It's not simple, not to say uber-complicated but it's fun ! Formation of pots balances the draw.

About Angel's idea ? How many players have gotten really serious in the training before tourneys lately ? Not very many.
Not many 1500+ or top 200 games have been played compared to past years.
About the curve ? The ELO line has been up 50 points over the past year : Would that make a 1800 player 1850 when he's actually 1600 at the moment ?
Is the proportion of games played drunk that huge to justify a low score ? Being underrated or overrated can be just explained by good or bad streaks that we all experience any time.
And it's not very readable and transparent !





[Mis à jour le: Thu, 29 November 2012 15:40]

      
SY/\/ \/\/ill
Senior Member
Maître des citrouilles

Messages: 382
Enregistré(e) en :
June 2010
Re:NC 2012 : Rules and discussion Thu, 29 November 2012 19:14
Sysyphus wrote on Thu, 29 November 2012 14:34


Is the proportion of games played drunk that huge to justify a low score?



In my case, hell yes! Smile
      
CAT-pegaso
Senior Member
Vainqueur Nation Cup AdR 2010

Messages: 240
Enregistré(e) en :
March 2006
Re:NC 2012 : Rules and discussion Sat, 01 December 2012 13:14
AT_dea1 wrote on Wed, 28 November 2012 17:29

My view on the format:

- 6 games in RR is very good & I like the possibility of draws
--- if you can't win by more than 1, is that really a justified win?
--- 6 games instead of bo5 means everybody gets the chance to play 6 games. With bo5 your match may be over after 3 games - that's frustrating as we participate because we like to play (at least for the loser). Additionally, as games won are a tiebreaker, I want to have a right to play all games possible.

- I'd like to play 6 games in KO, too
--- see the reasons above
--- the possibility of draws in matches reduces the amount of draws in clashes. I think that would be good, because I think the tiebreaker match is the worst idea we ever had.


- I'd like to change to 6 games in KO + games won as tiebreaker
--- the tiebreaker match puts too much emphasis on a team's best player
--- the tiebreaker match can be a timing problem
--- If games won are important the performance of weaker players will be more important and thus more valued (instead of "you lose, you don't count")




Too many issues open in this thread.
I'll start about the number of games. This is the easier one as I only want to say that agree 100% with dea Smile
Only will change that for the KO I would prefer 8 games rather than 6.
      
CAT-pegaso
Senior Member
Vainqueur Nation Cup AdR 2010

Messages: 240
Enregistré(e) en :
March 2006
Re:NC 2012 : Rules and discussion Sat, 01 December 2012 13:34
Sysyphus wrote on Tue, 27 November 2012 18:43

I do a) size of a team : I'd limit a team to 6 or 7 players to promote consistent lineups for a consistent competition.




I strongly disagree with this.
I'll put some examples of why:
a) There are players that are more one the mood of being part of the team and support them team mates that in playing. As they're happy playing very few games. Should this players be excluded in order to get consistent lineups?
b)Large teams is a good way to attract new players. Maybe new players may feel that they're not ready to play many games, but they can play some, feel the NC charm and want to keep playing and improve for the next year
c)This year 4 members of our team were on holidays at the same time (it was not coincidence, most of us are friend in real life). With the 7 players limit, this year would not have been possible for us to play in the same team.
      
CAT-pegaso
Senior Member
Vainqueur Nation Cup AdR 2010

Messages: 240
Enregistré(e) en :
March 2006
Re:NC 2012 : Rules and discussion Sat, 01 December 2012 16:39
About the number of players per match:

I personally prefer 5 rather than 4.

But anyway I think is the most important is decide it right now.
In this way if we decide now that the next year we will come back to the 5 players format teams will have almost a whole year for search for the new player.
What I think will be very bad is decide in august we will come back to 5 players don't giving teams with problems to find players the time to search for a new one.
      
CAT-pegaso
Senior Member
Vainqueur Nation Cup AdR 2010

Messages: 240
Enregistré(e) en :
March 2006
Re:NC 2012 : Rules and discussion Sat, 01 December 2012 17:22
Now the big issue, the seeding.

First some considerations:

Are some groups stronger than others?

I take the last three years of the NC, give 3 points to the winner 2 to the runner-up and 1 to the 3rd and 4th teams.

TRUE points are splitted between PRUE and WASA
UKB points are splitted between SNS and WASA

The final result is:

Group A: 9 points (4+4+1)
Group B: 7 points (3+2+1+1)
Group C: 5 points (5)

Group B (the group with more players whinning) is just in the medium being (in theory) the better chance to be first and the most difficult to make it to the top 12.

Is ELO basically random?

Well if you take a look at the erps tool right now, the three best ranked teams are the last 4 years winners. Coincidence?? Rolling Eyes Rolling Eyes
      
CAT-pegaso
Senior Member
Vainqueur Nation Cup AdR 2010

Messages: 240
Enregistré(e) en :
March 2006
Re:NC 2012 : Rules and discussion Sat, 01 December 2012 17:40
All this said, I point of view is that ELO is a good hint of how strong a team is, but not perfect, and maybe other teams perception will be better.

If we want to use teams perception, some methods have been proposed right now:
A) 3 teams choosing all the groups
B) 3 teams choose, then the 3 teams choosed choose the next 3 teams etc..
C) Each teams submits his opinion about the strenght of the teams.

With A and B splitting between you choose your own group teams or you put teams in other groups.

A) is the more likely to give unbalanced groups. Let's imagine that the strongest teams (UEG or TUS) decide that is better to have the stronger teams in your group. In this way if you loose against them is not that hard that if it happen in the KO round and you can kick out some of them.

B) seems ok to me, but could give again unbalanced groups if some teams give more importance to play or not play again your own country team or your friends that to their perception of how strong other teams are.

C) Is my favorite, maybe is not that funny neither that 'tactical' but is the one that gives more chances of balanced groups (from my point of view) and will be also interesting to see how each team is voting.

      
Sysyphus - Pommard
Senior Member
Vainqueur AdR All Around Tournament 2011

Pages Perso
Messages: 2409
Enregistré(e) en :
December 2007
Re:NC 2012 : Rules and discussion Wed, 05 December 2012 03:42
CAT-pegaso wrote on Sat, 01 December 2012 05:34


a) There are players that are more one the mood of being part of the team and support them team mates that in playing. As they're happy playing very few games. Should this players be excluded in order to get consistent lineups?



If they play then, should they not be counted in the ELO average of the team ?

If you want to keep large teams, then we need to broaden the ELO average.

But with 12 players, is it not possible to create 2 teams ?

Quote:


b)Large teams is a good way to attract new players. Maybe new players may feel that they're not ready to play many games, but they can play some, feel the NC charm and want to keep playing and improve for the next year



I do agree with this. But someone can step up and build a new team (i.e. French and American this year), keeping both teams under the same captain for example like Italy did.

The more players, the better. The more teams, even better.

Quote:


c)This year 4 members of our team were on holidays at the same time (it was not coincidence, most of us are friend in real life). With the 7 players limit, this year would not have been possible for us to play in the same team.



Well you have Safe Exit rule ! With 7 players, you still have 3 available to play and 1 can use Safe Exit ! Most TGV are also friends in real life, are working together to build teams under the same entity, but are just not in the same team.


CAT-pegaso wrote on Sat, 01 December 2012 09:22



Are some groups stronger than others?

I take the last three years of the NC, give 3 points to the winner 2 to the runner-up and 1 to the 3rd and 4th teams.(...)


Then you deny the strength of the teams seeded 3rd, 4th, 5th, 6th.
The point of Round Robin is not to be 1st, but to be qualified in KO.

Quote:


Is ELO basically random?

Well if you take a look at the erps tool right now, the three best ranked teams are the last 4 years winners. Coincidence?? Rolling Eyes Rolling Eyes


Yes, but we can read ERPS tool a different way.

Since the Round Robin is meant to reach KO, that reading does matter only if you want to reach 1st in RR.
Reaching 1st in RR is a bit less important considering the number of upsets we had during this NC.

If the ERPS tool is that accurate, let's look at the average of the whole team.
Only 2 teams have an average above 3590 in group A, when only 1 team has an average under 3590 in group B.

Since the lineups were including almost every player in each team, I think that is quite representative.

If we rank the teams by their average score, that's what we would have currently.


Group A

1, 6, 11, 15, 17, 19, 21


Group B
3, 4, 7, 8, 9, 14, 18


Group C
2, 5, 10, 12, 12, 16, 20, 22


Like Truck, I think there's not much discussion about the top 2 seed : UEG and CAT.

After, CAT 15th is an heresy obviously if they can put they best lineup every week. But for the rest, it's still pretty close to my perception, considering that teams 3-14 are quite close, and so are teams 17-21.

[Mis à jour le: Wed, 05 December 2012 04:34]

      
Sysyphus - Pommard
Senior Member
Vainqueur AdR All Around Tournament 2011

Pages Perso
Messages: 2409
Enregistré(e) en :
December 2007
Re:NC 2012 : Rules and discussion Sun, 09 December 2012 17:56
Community decided:



1. Should we fix a maximum number of players per team ?

A. YES : 5 votes

B. NO : 13 votes


2. Should we keep the ELO score ELO-related as a basis for the seeding of the teams in Round Robin ?

C. YES : 11 votes

D. NO : 7 votes


3. Should we modify the format of KO round and/or Round Robin ?
(i.e : keep the same format in RR and KO)

E. YES : 11 votes

F. NO : 7 votes




So the debate about a max number of players is over.

About ELO now:
2 questions :

1. Should we keep ELO 2p USA as a basis for our seeding ? Or go for a 2p ELO only ? Or anything out of the box like Angel ?

2. Since we don't have a max number of players : Should we broaden the seeding from the sum of the 4 best players to 5 ? 6 ? the whole team ?

I'd go for the whole team since most of players that registered played. Would be more representative of the strength of a team in RR. Seeding is meant for RR, not for KO where the best 4 lineup is more likely to play.

About the format

1. Keep 6 games in RR and go for 6 games in KO ? 8 games ?
2. bo5 in RR ? bo7 in KO ?
3. Keep 4 matches ? Go back to 5 matches ?
4. If we keep 4 ? Should we keep the current tie breaker which is hard to implement ? and does it give too much importance to one player in the outcome of a match ?

Discussion is open, a vote would be good before Christmas. So that everybody can be ready for next year.

Thanks all
      
AGT-DN
Senior Member

Pages Perso
Messages: 833
Enregistré(e) en :
November 2006
Re:NC 2012 : Rules and discussion Mon, 10 December 2012 00:47
Sysyphus schrieb am Sun, 09 December 2012 17:56

Should we keep ELO 2p USA as a basis for our seeding ? Or go for a 2p ELO only ? Or anything out of the box like Angel ?


Whatever the proposals will be, we should keep the question in mind how tho collect the ELOs easily!

The last erps-tool working is fixed to US+2player and the average of the top 6 players per team...changing to a different basic no. of players/to average team is no big problem at all with a calculating sheet but the tool will be useless for e.g. 2players-ELO as long as nobody is able to modify the code.

I collected ELOs for SPWC this year (60 players) with different tools to get them just in time at the end of the deadline which took a lot of hours working. This year over 150 players were registered for NC, so we maybe also should discuss about the technical way to implement a change in seeding!
      
ATN Drake
Senior Member

Messages: 812
Enregistré(e) en :
March 2006
Re:NC 2012 : Rules and discussion Mon, 10 December 2012 07:32
AGT-DN wrote on Sun, 09 December 2012 15:47



Whatever the proposals will be, we should keep the question in mind how tho collect the ELOs easily!

The last erps-tool working is fixed to US+2player and the average of the top 6 players per team...changing to a different basic no. of players/to average team is no big problem at all with a calculating sheet but the tool will be useless for e.g. 2players-ELO as long as nobody is able to modify the code.

I collected ELOs for SPWC this year (60 players) with different tools to get them just in time at the end of the deadline which took a lot of hours working. This year over 150 players were registered for NC, so we maybe also should discuss about the technical way to implement a change in seeding!



I think Alot of people voting just assume that the current system is the easiest, when in fact it is not.

I think people who voted for keeping the seeding the same with ELO, should be the ones to calculate all the scores, or change their vote to another easier seeding method..i.e. draft.
      
AAA_dea1
Senior Member
Vainqueur Nation Cup AdR 2007

Messages: 1928
Enregistré(e) en :
September 2005
Re:NC 2012 : Rules and discussion Mon, 10 December 2012 11:17
I learned last weekend that some people only voted for "keep ELO" because the vote didn't say what possible alternatives for "no ELO" could be.
So they feared that "no ELO" might result e.g. in "random" or "all good teams in 1 group" or something else they would consider worse than ELO.

Knowing this, we maybe should allow to discuss "non ELO variants", too, despite the vote, and then vote again in 2 phases, this time the other way around:

Phase 1:
2 Questions:

Q1:
If we were to keep ELO, which of the following variants would you prefer?
V1 / V2 / V3 (everybody votes here, wheter you like ELO or not)

Q2:
If we were to abandon ELO, which of the following variants would you prefer?
V11 / V12 / V13 (everybody votes here, wheter you like ELO or not)

now assemble the answers of phase 1 (say in Q1 the majority went for V3, in Q2 V12 was the winner) for phase 2
1 Question:
Shall we
A) keep ELO with V3
B) abandon ELO and do V12
      
Qorlas
Senior Member

Pages Perso
Messages: 1378
Enregistré(e) en :
January 2008
Re:NC 2012 : Rules and discussion Mon, 10 December 2012 11:25
If we remain with ELO: we could use average score. Erps tool gives it too.
      
Pages (6): [ «  <  1  2  3  4  5  6  >  » ]     
Sujet précédent:NC 2012: Schedule / Line-ups / Pairings
Sujet suivant:MPWC - group D
Aller au forum: