Author Topic: Top Cut Details  (Read 5831 times)

Chris

  • Guest
  • Trade Count: (0)
Top Cut Details
« on: January 06, 2013, 12:13:19 AM »
0
I am willing to allow the following trial for this tournament season:

Top Cut will be allowed (not required) for categories over 31 people.  So if the tournament host, with input from his players, wants to give it a try it will be okay with me.
 
Given that, we need to nail down the format for hosts that want to try it.

We never really hammered anything out after Rob made this post a while ago. I believe, however, that we had a relatively decent concensus in that thread: Top Cut of 16 after one player is undefeated. Is anyone particularly against this? Keep in mind this thread is not to debate the merits of top cut, rather, to hammer out the version of top cut we want to test out.

Offline Red

  • Tournament Host
  • Trade Count: (+2)
  • *****
  • Posts: 4791
  • It takes time to build the boat.
    • LFG
    • Southeast Region
Re: Top Cut Details
« Reply #1 on: January 06, 2013, 12:35:53 AM »
+2
Have a scaling top cut. At 32 people you take eight. At 64 and above take sixteen. And seed the players based on swiss records.(and pair the worst VS the best records.)
Ironman 2016 and 2018 Winner.
3rd T1-2P 2018, 3rd T2-2P 2019
I survived the Flood twice.

Offline Prof Underwood

  • Redemption Elder
  • Trade Count: (+7)
  • Hero Member
  • *****
  • Posts: 8597
    • -
    • East Central Region
Re: Top Cut Details
« Reply #2 on: January 06, 2013, 12:48:00 AM »
0
I think there should also be a way to fold the people who lose in the Top Cut back into the rest of the field for the later rounds.  If you do a top cut of 8, then there would be 2 rounds left for the first losers.  If you do a top cut of 16, then there would be 3 rounds left for the first losers.  That's a lot of time to sit around with nothing to do unless we can get them back in.

Also, I doubt that any tournament prior to Nats will actually have more than 31 players, so I'm not sure that this is really going to matter too much.

Chris

  • Guest
  • Trade Count: (0)
Re: Top Cut Details
« Reply #3 on: January 06, 2013, 12:55:13 AM »
0
I think there should also be a way to fold the people who lose in the Top Cut back into the rest of the field for the later rounds.  If you do a top cut of 8, then there would be 2 rounds left for the first losers.  If you do a top cut of 16, then there would be 3 rounds left for the first losers.  That's a lot of time to sit around with nothing to do unless we can get them back in.

Also, I doubt that any tournament prior to Nats will actually have more than 31 players, so I'm not sure that this is really going to matter too much.

I was thinking double-elimination for top cut, which would mean two off rounds for the people who get cut early, which isn't so bad.

NE and NC have a chance to, and Rob seems receptive to doing this at Nats, especially if the host is on-board with it.

Offline Prof Underwood

  • Redemption Elder
  • Trade Count: (+7)
  • Hero Member
  • *****
  • Posts: 8597
    • -
    • East Central Region
Re: Top Cut Details
« Reply #4 on: January 06, 2013, 01:01:11 AM »
0
I was thinking double-elimination for top cut, which would mean two off rounds for the people who get cut early, which isn't so bad.
I'd still like to fold the losers back in if possible.  I'd feel badly for the kid who does well early in the day, and makes the Top Cut, but then loses there.  They not only would feel badly for losing, but they would wish they lost earlier in the day instead so that they could still be playing like everyone else.  We don't want it to be a BAD thing to make it into the Top Cut.

Offline ChristianSoldier

  • Trade Count: (0)
  • Hero Member
  • *****
  • Posts: 1613
    • -
    • North Central Region
Re: Top Cut Details
« Reply #5 on: January 06, 2013, 01:15:31 AM »
0
I was thinking double-elimination for top cut, which would mean two off rounds for the people who get cut early, which isn't so bad.
I'd still like to fold the losers back in if possible.  I'd feel badly for the kid who does well early in the day, and makes the Top Cut, but then loses there.  They not only would feel badly for losing, but they would wish they lost earlier in the day instead so that they could still be playing like everyone else.  We don't want it to be a BAD thing to make it into the Top Cut.

Wouldn't Top Cut happen after all the regular rounds? If the tournament happens during the morning and afternoon, then all the rounds finish (whatever the number is) and in the evening you then have a sort of mini tournament (single or double elimination) where the top 8 or 16 play to see who wins.

This is something like how MtG tournaments work (although the one I went to they had the top cut the next day).

I'll be honest, I don't get why you would have a top cut if it happens during the regular rounds.
If you are reading this signature, thank a physicist.

Chris

  • Guest
  • Trade Count: (0)
Re: Top Cut Details
« Reply #6 on: January 06, 2013, 01:21:48 AM »
0
I was thinking double-elimination for top cut, which would mean two off rounds for the people who get cut early, which isn't so bad.
I'd still like to fold the losers back in if possible.  I'd feel badly for the kid who does well early in the day, and makes the Top Cut, but then loses there.  They not only would feel badly for losing, but they would wish they lost earlier in the day instead so that they could still be playing like everyone else.  We don't want it to be a BAD thing to make it into the Top Cut.

Wouldn't Top Cut happen after all the regular rounds? If the tournament happens during the morning and afternoon, then all the rounds finish (whatever the number is) and in the evening you then have a sort of mini tournament (single or double elimination) where the top 8 or 16 play to see who wins.

This is something like how MtG tournaments work (although the one I went to they had the top cut the next day).

I'll be honest, I don't get why you would have a top cut if it happens during the regular rounds.

The idea is that, after the cut off point, anyone who wants to quit can do so, but everyone else can feel free to finish out the tournament. There's no real harm in doing it that way, so it's kind of stuck through the debating process so far.

Underwood, I don't see any problem with putting those players back into the fold. Where would they be put though? Their records would put them towards the lower-middle pack having lost twice, which doesn't seem quite fair.

Offline lp670sv

  • Trade Count: (0)
  • Hero Member
  • *****
  • Posts: 1652
    • -
    • Northeast Region
Re: Top Cut Details
« Reply #7 on: January 06, 2013, 01:57:34 AM »
+2
Have them play for final placing out of the top whatever the cut is. Like this...

« Last Edit: January 06, 2013, 02:04:11 AM by lp670sv »

Offline EmJayBee83

  • Tournament Host
  • Trade Count: (+1)
  • *****
  • Posts: 5486
  • Ha! It's funny because the squirrel gets dead.
    • -
    • East Central Region
    • mjb Games
Re: Top Cut Details
« Reply #8 on: January 06, 2013, 09:25:52 AM »
0
Four items:

1) Rather than a straight top cut, I would prefer an X-# cut. Something where (for example) instead of taking the top sixteen players you take every player who has at most two losses at that point (an X-2 cut). The problem with a pure top cut is that (as long as you use LS differential as a tie breaker) it will retain the a chunk of the strength of schedule and the bad draw issues--you have just moved them back to the selection process. This comes at the cost of adding the likelihood of needing an extra play-in round, however.

2)
I was thinking double-elimination for top cut, which would mean two off rounds for the people who get cut early, which isn't so bad.
I also think this would be a good idea. You might want to consider the modified double elimination format used in wrestling tournaments where it is only double elimination for third place and below (wrestle-backs).

3)
I'd still like to fold the losers back in if possible.  I'd feel badly for the kid who does well early in the day, and makes the Top Cut, but then loses there.
Optimally you would have tournament pairing software that takes care of all of this automatically under the covers and not make any publicly viewable distinction between the top cut and the keep-on-keeping-on players.

4)
Also, I doubt that any tournament prior to Nats will actually have more than 31 players, so I'm not sure that this is really going to matter too much.
It is time to begin plugging the Type 2 Only tournament. This year it will be on March 1-2, and will feature even more Wild Bill (if that is possible) than ever before.


Offline Prof Underwood

  • Redemption Elder
  • Trade Count: (+7)
  • Hero Member
  • *****
  • Posts: 8597
    • -
    • East Central Region
Re: Top Cut Details
« Reply #9 on: January 06, 2013, 02:06:43 PM »
+1
1) Rather than a straight top cut, I would prefer an X-# cut. Something where (for example) instead of taking the top sixteen players you take every player who has at most two losses at that point (an X-2 cut). The problem with a pure top cut is that (as long as you use LS differential as a tie breaker) it will retain the a chunk of the strength of schedule and the bad draw issues
Any top cut in an unseeded tournament will ALWAYS have issues where some players don't get in simply because of a large number of early round "tough" matchups, and others WILL get in simply because of a large number of early round "easy" matchups.  This is one of the fundamental flaws with top cut.

Optimally you would have tournament pairing software that takes care of all of this automatically under the covers and not make any publicly viewable distinction between the top cut and the keep-on-keeping-on players.
This sounds like a GREAT idea.  And since we have some decent computer programmers on this forum, here's a project for you :)

It is time to begin plugging the Type 2 Only tournament. This year it will be on March 1-2, and will feature even more Wild Bill (if that is possible) than ever before.
This is usually one of the 3-4 biggest tournaments of the year, but as far as I know, it's never quite reached 32 players yet.

Offline EmJayBee83

  • Tournament Host
  • Trade Count: (+1)
  • *****
  • Posts: 5486
  • Ha! It's funny because the squirrel gets dead.
    • -
    • East Central Region
    • mjb Games
Re: Top Cut Details
« Reply #10 on: January 06, 2013, 06:33:50 PM »
0
1) Rather than a straight top cut, I would prefer an X-# cut. Something where (for example) instead of taking the top sixteen players you take every player who has at most two losses at that point (an X-2 cut). The problem with a pure top cut is that (as long as you use LS differential as a tie breaker) it will retain the a chunk of the strength of schedule and the bad draw issues
Any top cut in an unseeded tournament will ALWAYS have issues where some players don't get in simply because of a large number of early round "tough" matchups, and others WILL get in simply because of a large number of early round "easy" matchups.  This is one of the fundamental flaws with top cut.
True. Unfortunately the current (non-top cut) and proposed top cut formats both use LS differential as a tie breaker, which couples them both strongly to draw randomness. Draw randomness is--without a doubt--the absolute *worst* feature of the Redemption game mechanics. X-# would minimize this effect to the largest extent possible. It also gives a player (who cares) an objective measure before the tournament begins of what they need to do to advance.

It is time to begin plugging the Type 2 Only tournament. This year it will be on March 1-2, and will feature even more Wild Bill (if that is possible) than ever before.
This is usually one of the 3-4 biggest tournaments of the year, but as far as I know, it's never quite reached 32 players yet.
A guy can dream, however.  ;)

Offline Prof Underwood

  • Redemption Elder
  • Trade Count: (+7)
  • Hero Member
  • *****
  • Posts: 8597
    • -
    • East Central Region
Re: Top Cut Details
« Reply #11 on: January 06, 2013, 08:22:36 PM »
0
True. Unfortunately the current (non-top cut) and proposed top cut formats both use LS differential as a tie breaker, which couples them both strongly to draw randomness. Draw randomness is--without a doubt--the absolute *worst* feature of the Redemption game mechanics.
After looking at the results of last summer's Nats closer than I ever have in the past, I am beginning to agree with you that the HUGE variance in strength of schedule of different players is a significant problem is on the verge of simply being unfair.  However, the LS differential tiebreaker really doesn't have much to do with it.

If a player happens to run into a lot of top players and takes a few losses, they aren't going to make the top cut regardless of tie-breakers.  Meanwhile someone could cruise through easy early rounds and get in no problem.  However the only way to resolve this issue would be to follow the Olympic/World Cup formula and have the early rounds be split into small groups each with 1 top player.

But to switch to that system would require a massive amount of support which I really don't think is there, so I don't see that ever happening.

Offline Alex_Olijar

  • 16plus
  • Trade Count: (+15)
  • *
  • Posts: 8124
  • This guy is my mascot
    • -
    • Northeast Region
Re: Top Cut Details
« Reply #12 on: January 06, 2013, 09:07:23 PM »
+1
Terrible logic in support of a bad system

You keep trumpeting this idea and it keeps being ridiculous because, as I have noted before, that only works because the Olympics ARE A TOP CUT and the groups ARE LOOSELY BASED ON EQUITY BETWEEN TEAMS yet still result in repeated monikers for some groups as "GROUPS OF DEATH". Talk to Netherlands Soccer Fans about the 2012 Euros and you'll get the idea. Your proposed solution is to just judge the players skill level to put them in groups relatively fairly. I ask how do you plan on doing this? You hated my Redemption BCS thread and criticized it for being pretty much just opinion and only really having net negative potential to hurt people who aren't on the list. How is ranking people and putting them in groups any different? How can you objectively determine player skill? You have said before opinion isn't sufficient. RNRS isn't sufficient. I would enter every Nationals with less than 5 points most likely. I'm betting that would probably ruin your whole seeding and group idea since my skill level wouldn't equate with my RNRS points.

Basically, this is a terrible idea and I have no idea why on earth you keep bringing it up. To me, it shows you have literally no idea how tournament structures work or how to attempt to properly manipulate them to create a fair system. Nothing about your idea is even remotely tenable without a BCS-like structure of computers tracking game data, which everyone hates anyway.


Offline Alex_Olijar

  • 16plus
  • Trade Count: (+15)
  • *
  • Posts: 8124
  • This guy is my mascot
    • -
    • Northeast Region
Re: Top Cut Details
« Reply #13 on: January 06, 2013, 09:08:25 PM »
0
Also, since I've been playing a ton of Pokemon lately (to the point of working on entering the competitive realm) let's look at the old system (which was BCS-like in the sense that it used an ELO rating that followed you throughout tournaments) Pokemon used to invite players to the World Championships in the words of the only two-time world champion, Jason Klaczynski:

Quote
...if a player that would make the Top Cut instead dropped after the Swiss rounds completed, his or her rating would fare much better than a player in an identical situation that chose to play the Top Cut, but lost in the first round of it. Depending on whom you were paired against, it was very possible you could make the Top Cut, lose your first round, and end up with an even lower ELO rating than you had at the beginning of the tournament! In other words, you could make the Top Cut, but end up with a lower chance at qualifying for Worlds after the tournament finished. Even worse, in most cases, a player with an established ELO rating could win his first round of the Top Cut, and then be eliminated in the subsequent round, and still lose points by playing in the Top Cut. (The player would lose points from playing in the Top Cut, but not necessarily overall.) In terms of rating, that player would have been better off simply dropping after Swiss, even though he or she won his or her first match. In a City Championships with a Top 8 cut, a player with a high ELO rating that made the Top Cut often had only two possible outcomes from playing in the Top Cut:

    1) Win the tournament and gain ELO points from the Top Cut.
    2) Finish any place other than 1st and lose ELO points from playing in the Top Cut.

While the player’s gain from 1st place would almost always be higher than the loss had he or she been eliminated in Top 4 or the Finals, a player that was comfortable with his or her rating sometimes felt it was better not to take the risk.

Additional issues arose from the fact that some skilled players had chose to enter the season later, perhaps skipping City Championships and first appearing at a State or Regional Championship. These skilled players would have ratings much lower than other skilled players who had already built their rating by playing and winning in earlier tournaments. If you had a high ELO rating, playing a tournament with these latecomers was like playing Russian Roulette, take a loss to one of them and your rating would crash. Beat them, and you’d gain very little. Though your opponent may very well be a skilled player, if it is not reflected in his or her ELO rating, your ELO rating will take a big hit by losing to him or her. Areas that had a higher proportion of skilled players entering the season later would, on average, curb the ELO ratings of the highest-rated players. This problem was augmented by the fact that these State and Regional Championships would have a higher “K Rating,” meaning wins and losses would have a larger effect on your ELO rating than at previous, smaller tournaments.

It should go without saying that the constant dropping and sitting out of tournaments combined with the nerve-wracking games against a low-rated opponent produced a less fun tournament environment. (Losing on Turn 1 to a Sableye and dropping 20 ranking spots didn’t seem to please anyone.) The ELO system produced a lot of crazy situations, too. In a City Championship with a Top 8 cut, I once saw a player who finished 11th win the event as three players dropping allowed him to sneak in to the Top Cut at 8th and win the event.

Obviously there's some variables that are in Pokemon that aren't there in Redemption (like there were already top cuts) but this should suffice to show how every system is pretty terrible for fellowship, so either stop ranking players or just use a ranking system that actually works.

Offline EmJayBee83

  • Tournament Host
  • Trade Count: (+1)
  • *****
  • Posts: 5486
  • Ha! It's funny because the squirrel gets dead.
    • -
    • East Central Region
    • mjb Games
Re: Top Cut Details
« Reply #14 on: January 06, 2013, 09:43:40 PM »
0
True. Unfortunately the current (non-top cut) and proposed top cut formats both use LS differential as a tie breaker, which couples them both strongly to draw randomness. Draw randomness is--without a doubt--the absolute *worst* feature of the Redemption game mechanics.
After looking at the results of last summer's Nats closer than I ever have in the past, I am beginning to agree with you that the HUGE variance in strength of schedule of different players is a significant problem is on the verge of simply being unfair.  However, the LS differential tiebreaker really doesn't have much to do with it.
In the 2012 final rankings, places four through ten all have 21 points, with half a lost soul per game separating fourth place from ninth place. This is a fairly substantial contribution to the overall final rankings.

This tendency gets exacerbated in smaller sample sets. Taking as an example a top-8 cut that would have occurred 2012 Nats after round seven.  Martin Miller, Jonathon Greeson, Alex Olijar and Josh Brinkman would have been in by the virtue of having 18 points.  Following that we would have had 8 players with 15 victory points duking it out for the remaining four slots. Of those, Alex Lewis had a LS differential of 20 so he would have clearly been in. The remaining seven players had LS differentials ranging from 5 to 9. This is a difference of four lost souls over the course of seven games, and on this we are going to allow three through to the top cut and drop the remaining four.

Among evenly matched players, LS differential is tied almost directly to draw quality. It seems like a rather poor choice to make the first tie breaker something tied most closely to random chance.

Offline everytribe

  • Tournament Host
  • Trade Count: (+30)
  • *****
  • Posts: 1065
    • -
    • North Central Region
Re: Top Cut Details
« Reply #15 on: January 06, 2013, 10:19:57 PM »
0

It is time to begin plugging the Type 2 Only tournament. This year it will be on March 1-2, and will feature even more Wild Bill (if that is possible) than ever before.
This is usually one of the 3-4 biggest tournaments of the year, but as far as I know, it's never quite reached 32 players yet.
[/quote]

We hit 33 one year. The number isn't as important as the quality of the players. We usually average 4-5 former National Champs. And the fact that you get to play a bunch of different players other than the 2 or 3 type 2 players in your own play group.
Old Guys Rule

Offline lp670sv

  • Trade Count: (0)
  • Hero Member
  • *****
  • Posts: 1652
    • -
    • Northeast Region
Re: Top Cut Details
« Reply #16 on: January 06, 2013, 11:23:53 PM »
+2
Yes, people can get a cake walk to the top cut but the difference is that once top cut starts, only the top 8 or however many players can win it all. In the current system a player can lose and early game and play weak opponents all the way to winning the final table, without ever having to play the person that came in second at larger tournaments. In top cut you MUST play the 2nd and 3rd/4th place competitor in order to win it all. So while you can still cruise to top cut through the early rounds, you can't win it by cruising through. If you take it on it's face the problem exists, not with top cut, but with the swiss style qualifier. This can't be wholly avoided but something like a world ranking for seeding could help that out.

Chris

  • Guest
  • Trade Count: (0)
Re: Top Cut Details
« Reply #17 on: January 06, 2013, 11:49:29 PM »
0
Quote
If a player happens to run into a lot of top players and takes a few losses, they aren't going to make the top cut regardless of tie-breakers.  Meanwhile someone could cruise through easy early rounds and get in no problem.

Who makes top cut really doesn't matter. If a person had a hard schedule and didn't make it in, they weren't going to survive in top cut anyway. What matters is who ultimately places in top cut. However, the merits of top cut have been debated for pages and pages already, and that is not the point of this thread. I'd appreciate it if we can stay on topic.

I propose a double elimination with a seeded top cut of 16. Is anyone opposed to these specific points? We can work out what to do with people who get eliminated later.

Offline Gabe

  • Global Moderator
  • Trade Count: (+69)
  • *****
  • Posts: 10675
  • From Moses to the prophets, it's all about Him!
    • -
    • North Central Region
    • Land of Redemption
Re: Top Cut Details
« Reply #18 on: January 07, 2013, 12:01:07 AM »
0
I propose a double elimination with a seeded top cut of 16. Is anyone opposed to these specific points? We can work out what to do with people who get eliminated later.

I prefer top 8 and single elimination. Does someone plan to update the tournament tracker excel program to accommodate top cut?
Have you visited the Land of Redemption today?

Offline lp670sv

  • Trade Count: (0)
  • Hero Member
  • *****
  • Posts: 1652
    • -
    • Northeast Region
Re: Top Cut Details
« Reply #19 on: January 07, 2013, 01:16:03 AM »
+2
I don't think you would need it, when you do to cut it just becomes a bracket style tourney

Offline Prof Underwood

  • Redemption Elder
  • Trade Count: (+7)
  • Hero Member
  • *****
  • Posts: 8597
    • -
    • East Central Region
Re: Top Cut Details
« Reply #20 on: January 07, 2013, 07:52:16 AM »
0
We hit 33 one year. The number isn't as important as the quality of the players.
I was wrong then.  And by the way, I did NOT mean to malign the great tradition of the T2-only Tournament by saying that it probably would have less than 32 players.  It is one of the highlights of the Redemption calendar every year, and I hold it in high regard.

If we end up doing top cut, I would prefer to keep it small for the first year of trying it out, and have it affect the least number of rounds.  Therefore I would prefer a top 8, single-elimination version.

Offline everytribe

  • Tournament Host
  • Trade Count: (+30)
  • *****
  • Posts: 1065
    • -
    • North Central Region
Re: Top Cut Details
« Reply #21 on: January 07, 2013, 06:41:41 PM »
0
We hit 33 one year. The number isn't as important as the quality of the players.
I was wrong then.  And by the way, I did NOT mean to malign the great tradition of the T2-only Tournament by saying that it probably would have less than 32 players.  It is one of the highlights of the Redemption calendar every year, and I hold it in high regard.

I know I was just getting some free advertizing on another thread.  :) It would be nice if you could join us.
Old Guys Rule

Offline STAMP

  • Tournament Host
  • Trade Count: (+19)
  • *****
  • Posts: 5014
  • Redemption brings Freedom
    • -
    • Northwest Region
Re: Top Cut Details
« Reply #22 on: January 07, 2013, 07:02:54 PM »
0
Quote
(Losing on Turn 1 to a Sableye and dropping 20 ranking spots didn’t seem to please anyone.)

Okay, I'll stop complaining about 3-turn losses in Redemption.  ;)
Final ANB errata: Return player to game.

Offline EmJayBee83

  • Tournament Host
  • Trade Count: (+1)
  • *****
  • Posts: 5486
  • Ha! It's funny because the squirrel gets dead.
    • -
    • East Central Region
    • mjb Games
Re: Top Cut Details
« Reply #23 on: January 07, 2013, 07:28:03 PM »
+1
I don't think you would need it, when you do to cut it just becomes a bracket style tourney
The Tracker is a vengeful deity. It would turn on anyone who abandoned it mid-tournament.

Offline lp670sv

  • Trade Count: (0)
  • Hero Member
  • *****
  • Posts: 1652
    • -
    • Northeast Region
Re: Top Cut Details
« Reply #24 on: January 07, 2013, 07:56:06 PM »
0
I don't think you would need it, when you do to cut it just becomes a bracket style tourney
The Tracker is a vengeful deity. It would turn on anyone who abandoned it mid-tournament.

You can still use it just delete the top 8 players and use the rest to let the other people keep playing for some kind of consolation prize (since a main concern seems to be having everyone play) and do the top cut as a normal bracket style tourney.

 


SimplePortal 2.3.3 © 2008-2010, SimplePortal