SuperLadle Discussion - cont. from time change thread

[Concord]

just so no one get's confused, the idea isn't to add a league or to change the Ladle, it's to add a one day tournament

[Lucifer]

just so no one get's confused, the idea isn't to add a league or to change the Ladle, it's to add a one day tournament

Ah, but you end up with a de facto league at that point, by seeding a new one-day tournament with results from the Ladle. I think that's a great idea, because then there will be a higher tournament for people who do well in the ladle to go at it again, and with the ladle always available to anybody who wants to play, there's a path for any team to get its way to the higher tournament.

Keeping in mind that "league" is generally considered a round robin tournament, so is still a tournament. I'm just turning the word around to refer to a series of non-round robin tournaments to seed another non-round robin tournament.

[sinewav]

just so no one get's confused, the idea isn't to add a league or to change the Ladle, it's to add a one day tournament

Ah, but you end up with a de facto league at that point, by seeding a new one-day tournament with results from the Ladle.

Well, that's just awesome. I like how this is shaping up.

As we've seen before, the best way to introduce potent change is little by little. Tiny, focused effort. The prospect of a seasonal Bowl can be just what Fortress needs right now. Sometime later we should set a tentative date for the event and a time span for a season. Let's chat this idea up on the grid and see if we can get some more input from the community.

[Lucifer]

As we've seen before, the best way to introduce potent change is little by little. Tiny, focused effort. The prospect of a seasonal Bowl can be just what Fortress needs right now. Sometime later we should set a tentative date for the event and a time span for a season. Let's chat this idea up on the grid and see if we can get some more input from the community.

If you run a seasonal bowl, say 4 times a year (because in some areas, there really are 4 seasons), then I'd like to see added (later, after a year or two of successful bowls) a yearly tournament seeded from the bowls.

[Phytotron]

Alright, I just threw this together. This is only an illustrated example of one approach you all could take to a SuperLadle based on one year's-worth of Ladles (2011).

IMPORTANT NOTE: I looked only at the team names. I did not take into account the rosters of those teams. Indeed, I think the prime challenge to the "legitimacy" of a super-tournament based on prior Ladles is the inconsistency of teams—in terms of participation, team names, and rosters; even the number of teams per Ladle. That's why I think some sort of simple league/season would have more competitive value and meaning. Doesn't have to replace the Ladle; they could be separate entities. But before I digress, on with it....


This is based on Ladles 41-52. Initial qualification for the tournament is a victory in the final of at least one Ladle in 2011. Only teams in the WIN column qualify. That resulted in 7 teams (I consolidated some; see below).

Code: Select all

WIN                      LOSS

you in danger, girl		Team.uNk
Crazy Tronners Alpha	  ~*Speeders*~
Rogue Tronners		     Speeders
Twi¦×¦ted ¦×¦ats		   Team.uNk 
Twi¦×¦ted ¦×¦ats		   Rogue Tronners 
Speeders			        Twi¦×¦ted ¦×¦ats
Team.uNk			        Speedhax
uNk.Team			        _~Rogue Tronners~_
\\MeetYourMaker		    ~*SP*~ 
~*SP*~ shows [...]		 \\MeetYourMaker
\\MeetYourMaker		    Speeders
Speeders			        \\MeetYourMaker

Seeding begins with the number of Ladle wins. As you can see, there was one team with 3 (SP); three teams with 2 (TX, MYM, uNk), and three teams with 1 (CT, RT, YiDG). Because of the two 3-way ties, I had to institute tie-breakers. They were as follows:

(a) Head-to-head against teams with same number of wins
(b) Wins against higher seeds
(c) Win-loss record
[Overall win-loss record gets a lower priority because of the aforementioned variance in teams from one Ladle to the next. Also, if I needed another step, I might've gone to margin of victory. Like I said, this is just a one-off.]

Final seeding, including record and tie-breakers:

Code: Select all

[1] Speeders (3-4)
[2] Twi¦×¦ted ¦×¦ats (2-1)       (a) +1 uNk
[3] \\MeetYourMaker (2-2)        (a) +0     (b) +2 SP
[4] Team.uNk (2-2)               (a) +0     (b) +0
[5] Crazy Tronners Alpha (1-0)   (a) +0     (b) +1 SP     (c) +1
[6] Rogue Tronners (1-2)         (a) +0     (b) +1 SP     (c) -1
[7] you in danger, girl (1-0)    (a) +0     (b) +0.5 uNk

[The 0.5 there is meant to indicate the lower value of uNk's 2 wins to SP's 3—an actual point value corresponding to number of wins might be more appropriate, but again, just a one-off.]

The result is a bracket like so:

Oh, by the way, handy site: http://www.printyourbrackets.com/

[sinewav]

March Madness hangover becomes our profit, haha.

[PokeMaster]

IMPORTANT NOTE: I looked only at the team names. I did not take into account the rosters of those teams. Indeed, I think the prime challenge to the "legitimacy" of a super-tournament based on prior Ladles is the inconsistency of teams—in terms of participation, team names, and rosters; even the number of teams per Ladle. That's why I think some sort of simple league/season would have more competitive value and meaning. Doesn't have to replace the Ladle; they could be separate entities. But before I digress, on with it...

I totally agree about the conflict between the degree to which teams can be in flux and how much the tournament itself only names the best team the champion (aka, its "legitimacy"). I think though that we should, overall, favor that flux factor in teams rather than trying to setup a perfect equivalent between winning the Bowl and being "the best." Hence, this is good reason for a season to only be 6-9 months long, so that we can allow the months in between to give time for that change to happen. That's not to say that teams shouldn't be in flux while in season. For that reason I think we should also have few yet firm rules for qualifying a team as "being the same" during the season (probably similar to seeding rules).

Remember, one of the things that we're hoping to achieve in this tournament is to encourage higher levels of play from all teams, especially the top ones. Changing teams is a great way to keep producing different (and hopefully better) results, and it's a great contributor to improvement, on both individual and group levels. Taking that tool away may very well backfire on us and work against us really achieving a higher level of play.

</goals rant>

[sinewav]

Don't make this crazy complicated. Let's just use the stats page to take the top six teams and drop them into a bracket. The requirement to play on that team is dependent on your playing for them previously in the season. The talent in this game is pretty diffuse. I'm not worried about "legitimacy," whatever that means.

[Lucifer]

I wouldn't worry too much about rosters either. There have been some historical pro league events where the teams were so in flux it was hard to tell who was who anymore, iirc.

[Titanoboa]

Mid-season switching of teams shouldn't be restricted. Let's imagine Durka throws a fit and everybody leaves Tx, but he manages to recruit 3 solid players and keeps playing ladles with 2 random guests on his team, but they do well and qualify for bowl. Of course they should be allowed to participate?

However, for the actual bowl, like sine said, you'd have to have been on the team during the season. Makes sense, I think.

[sinewav]

Right, more freedom = more fun. Keep it as simple as possible with few rules.

I was thinking it might be good to prepare a warm-up bowl to work out any bugs. I suggest this coming June, to mark the anniversary of Bowl 1. If it works out, maybe we can restart the season in September/October for the following June.

I'm going to guess the small volume of dominantly positive feedback is an indicator of strong community approval, as is usually the case. But if there is a sudden uproar nearing June, we can always hold the Bowl alongside Ladle so everyone gets their Fort-fix that month. Seems the easiest way to fill the 6-team bracket would be Ladle-wins/Final-visits, then Match-wins/Match-totals to break ties.

Good? Good.

[Concord]

I'm not crazy about primarily counting Ladle wins or finals appearances, because if the season is only 6 months long, the door is open for a team that only played one Ladle qualifying, and this should really be about performance over a longer period of time, not just a collection of good teams. Furthermore, it's likely that a team that only plays one Ladle and wins has members that usually play on other teams

Furthermore, teams that lose 1-2 to the winner in the semifinals or quarterfinals shouldn't necessarily be penalized against a team that loses 0-2 in the finals.

Here's my qualifier

Code: Select all

( game wins + finals match wins ) * match winning percentage

This rewards total wins, finals appearances and performance in the finals. With the winning percentage multiplier there is a balance between total accomplishment over the whole season and high levels of performance in just a portion of it. Winning percentage both rewards losing 1-2 and punishes winning 2-1, both of which are significant.
Here's what the numbers would look like for the past 4 months

Code: Select all

1 - Revolver             ( 7 + 4 ) * 1.00 = 11.0
2 - Meet Your Maker      ( 7 + 3 ) * 0.71 = 7.1
3 - Crazy Tronners A     ( 8 + 2 ) * 0.65 = 6.5
4 - Team Unknown         ( 6 + 1 ) * 0.61 = 4.3
5 - Rogue Tronners       ( 5 + 0 ) * 0.68 = 3.4
6 - Twixted Xats         ( 4 + 0 ) * 0.55 = 2.2
  - Pure Luck            ( 4 + 0 ) * 0.55 = 2.2
xxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
8 - Cool Trainers ( 3 + 0 ) * 0.67 = 2.0
9 - Speeders ( 3 + 0 ) * 0.53 = 1.6

Teams that tie should play a best of three play in match.

[sinewav]

Here's my qualifier

Code: Select all

( game wins + finals match wins ) * match winning percentage

Great, even better. Nice and easy. But I guess we need a new field for stats. Any way to modify this with what we have currently and get a similar result?

I imagine the Bowl as a natural extension of Ladle and I've made a little spot for it on the operations page and cross-referenced here. As time goes by, we can flush out ideas and tighten up that section, trying to keep it as simple as possible of course.

http://wiki.armagetronad.org/index.php/ ... s#The_Bowl

[Concord]

Code: Select all

( match wins * 0.5 + ladle wins * 2 ) * winning percentage

this has a substantially different meaning, slightly different numbers, but the same teams albeit in a different order. It punishes 2nd place teams that win a match in the finals when compared to the other qualifier. The number to the left of each team is the ranking with the other method, and the number in parentheses is the score from the other method.

Code: Select all

1 - Revolver	         11.0  (11.0)
3 ↑ Crazy Tronners A	 6.9   (6.5)
2 ↓ Meet Your Maker 1	6.8   (7.1)
5 ↑ Rogue Tronners	   4.4   (3.4)
4 ↓ Team.uNk	         4.3   (4.3)
6 - Pure Luck	        3.0   (2.2)
  - Twi¦×¦ted ¦×¦ats	 3.0   (2.2)
xxxxxxxxxxxxxxxxxxxxxxxxxxxx
8 - Cool Trainers	    2.0
9 - Speeders	         1.9

I'm against just counting raw match wins, as in this second model, because the Ladle is based on a best of three format, and the logic behind counting raw match wins is the same as counting point differential within matches, and I'm against that too. A win is a win, regardless of the margin of victory. Including match wins in the case of the finals I am for, because winning a match in the finals means that your team was one hundred points away from winning a Ladle, and this is significant.

Alternatively, we could simply sum Ladle scores or quality scores. Ladle score is just rounds played * winning percentage. Ladle win quality score is a flawed statistic [1] but it would also work well for our purposes. Summing the simpler score, Ladle score, and adding twice Ladle wins would be a very close approximate to my original qualifier. This is probably the best method, as dlh's stat's are semi-automatic and Ladle wins are already counted.

Code: Select all

ladle score + ladle wins * 2

[1] And I should know, I invented it. It is essentially a measure of parity, and rewards winning in Ladles with high parity. It's based on the assumption that every Ladle victor is roughly even in skill, so therefore teams that win Ladles against tougher competition score higher. A team that wins every game 2-1 scores higher in quality score than a dominant team that wins every game 2-0, because it is assumed that the dominant team simply faced easier competition. It is a flawed assumption, in my opinion. However, for the purposes of qualifying for the Bowl, absolute scores are not as important as relative scores. The dominant team, while having a lower absolute score, has a higher differential between their scores and the other teams from that particular Ladle. That relative advantage is preserved as long as that team also played in the Ladles that had high parity. The statistic also has some more serious flaws, like the occurrence of a team losing 1-2 in the opening round having a higher score than the team that beat it, in instances where the victor loses 0-2 in the next round.

[Phytotron]

I'm not worried about "legitimacy," whatever that means.

Using the example I gave (or even using stats) of using a year's worth of Ladles, what meaning would any tournament have if the SP team had a different roster of players in each Ladle? Or if, say, CT and RT had the same players? That's why I'm suggesting a season format that isn't based on and doesn't incorporate the Ladle itself at all. It doesn't have to replace the Ladle (though eventually could, depending on popularity), but would at least run independently side-by-side. Of course, players or whole teams could participate in both.

I wouldn't worry too much about rosters either. There have been some historical pro league events where the teams were so in flux it was hard to tell who was who anymore, iirc.

From one season to the next, sure, but not within a season. Sure, some roster changes occur within the season, but the majority of the team remains intact, and there are incentives and penalties in place to retain that stability. Again, otherwise, what meaning does it have to even say "this is a team?" What meaning does a championship have?

Also, a championship is for that single season, not multiple seasons. In this respect, the Ladle is basically a season unto itself. I'm not aware of any team sport or game where multiple seasons or tournaments are used to base another tournament on, and it's for this reason of team continuity.

To all of you: Do you allow teams to completely change up their rosters in the middle of the Ladle? Or rename themselves? Or for new teams to join? Would you consider it legit if the two teams in the final had completely different rosters than they had in the first round?

I'm not crazy about primarily counting Ladle wins or finals appearances, because if the season is only 6 months long, the door is open for a team that only played one Ladle qualifying, and this should really be about performance over a longer period of time, not just a collection of good teams. Furthermore, it's likely that a team that only plays one Ladle and wins has members that usually play on other teams

Totally agree.


Listen, I don't think a simple season-based format is complicated. (I think the stats proposal based on the Ladle and a bunch of scattered, amorphous teams is way more complicated, and a lot less meaningful.) Do this independently of the Ladle:

Season Length: Three months.

Number of Teams: However many sign up for a given season, I reckon. Of course, the teams need to be decided before the season begins; no new teams afterwards. How many do you think could be constituted and remain more-or-less stable? Say, six? In a round robin format that would be 15 games. Think you all could fit those in three months?

Rosters: I don't know how many players are on a typical team, but whatever that is plus 2-3 spots allowed (not required) for subs/alternates.

Roster Rules: Simple rules that allow for reasonable flexibility within rosters (things come up in people's lives, etc.) while keeping them mostly stable and preventing abuse. Again, what's the point of even doing this if all the teams have different players at the end of the season?

So, I think a good, simple base rule to start with would be to require that, say, 2/3 or 3/4 of the original roster must still be in place at the conclusion of the season. Those players would not have to be designated at the outset, just so long as that percentage is there at the end. If not, DQ.

You could stop there, allowing any player movement beyond that, but possible rules beyond that could be, for examples: Once a player leaves a team, they cannot rejoin. Or, they can. Or, they can, but only the team they left. Or, they can join any other team. Or, institute a trade rule where players can only join another team if they swap players between the respective teams. Or whatever. Develop it as you go along, just as you have with the Ladle. Just as have other sports and gaming leagues.

Scheduling: Like I suggested earlier, don't you think you all can schedule at each team's convenience? I've noticed that clan forums usually have a "challenge us" section, so it seems they've been doing that. And then there's pickup.

[Speaking of pickup, talk about complicated rules. If you all can organize that spontaneously, plus all the Ladle rules, I really don't see why a format like this should be considered complicated.]

So, open scheduling provided all matches are played by the conclusion of the season period. But, if you all think there's a better reason to schedule them all, then by all means.

Any matches not played simply go down as a forfeit in the loss column of those teams. No DQ for missed matches.

Championship Tournament: Could take place a week after the official end of the season. Seeding determined by win-loss record, and any necessary tie-breakers. Tournament format determined by number of teams (see website I linked to earlier). With the example of 6 teams, the top two seeds would get a first round bye. Tournament is played, champion crowned.

New season begins. All new teams—number, names, and rosters. Rinse, repeat. Over time you could expand it.

Honestly, what's so complicated about that?