Card Game DB
Sign In
Create Account
Register now to gain access to all of our features. Once registered and logged in, you will be able to create topics, post replies to existing threads, give reputation to your fellow members, get your own private messenger, post status updates, manage your profile and so much more. If you already have an account, login here - otherwise create an account for free today!
0
Checking into Rehab (Tournament Rules Changes, part II)
By now you've probably seen my (joined by wwdrakey & JCwamma) rant about the new tournament structure. I do want to make it clear that I don't think the changes mean the sky is falling or will single-handedly ruin tournaments. I just think it's a universally bad decision, and I struggle to see *any* upside (aside from the "upside" of shorter tournaments, which wasn't a stated goal) - AND the whole thing is at odds with itself - contradicting its stated intentions.I can't help but be frustrated when I think that people discussed and implemented a system that is, as far as I can see, marginally worse than the status quo in every way. So, if yesterday's snark was questioning what FFG were smoking, today, I thought I'd offer something more constructive. Some Rehab, if you will. I'm told the first step is admiting you have a problem.
(Some) Solutions
Option 1: Reverse the changes, go to the old system
This one is simple enough - just rewind. There are plenty of things I dislike about the old system (true swiss to cut <25%), but it was better than the proposed system in every way. This is the easiest change to make, and will probably upset nobody.
Option 2: Make the Tiebreaker skill/play based (e.g. Margin of Victory)
If you're going to use the proposed system, including some (sometimes as few as a third of) X-1 records in the cut and not others has to be done in such a way that it does not feel random. While the first tie-breaker (if you've played everyone with your score range) is not, it's a very rare case, and the second to fourth tie-breakers (Strength of Schedule, Extended Strength of Schedule, and Random) all feel very arbitrary. The way to make this decision feel less fickle is to make a player decisions meaningfully impact their score.
The most common example, in sports, is using Goal Differential, typically as a second tiebreaker. In Thrones, the closest equivalent would be ending power differential; a 15-0 win would be considered more 'convincing' than a 15-14. This would favour different play than the existing model, but it's something you can build and play around and gives you at least a perception of agency - a feeling that your decisions impacted your tiebreaker. Other examples include a bonus for each unused plot in your deck, provided it hasn't recycled (take that, plot cyclers - And Martell!). Again, this would benefit certain decks, but it still gives you a feeling of control. You could even have a more complex system where a player "picks" one of several tiebreakers for their deck going into the tournament.
I quite like the idea of someone losing a game because they're trying to improve their tiebreaker score and get cocky... (Oberyn, anyone?).
Option 3: Scrap the cut
Play pure swiss. I know this isn't a popular one, because people like the cut (as spectators, as a goal to shoot for) and don't like the perception of single-elimination early on in a tournament. However, unlike their proposed system, it's logically and internally consistent. It creates a winner for the day when everyone is playing by the same rules throughout the tournament. There's no tie-breakers, and you can still reward the top X with additional prizes. It also creates a fixed tournament length for all players, which can potentially make travel easier.
I don't personally like this choice much, because I feel it's ill-suited to a high variance game, but I list it because unlike the proposal, it has some actual advantages over the current system (even if I don't feel those advantages offset what is lost).
Option 4: Expand the cut
Surely if shrinking the cut created such an uproar, the solution is to expand it! More games, more chance of making the cut, more forgiving to variance... it's everything we want!
Maybe you can tell from my tone that I actually quite oppose this option, for a number of reasons. The larger the cut, the less of a 'goal' it becomes to shoot for. It also means that tournament lengths vary widely for different players, making planning/ride-sharing even more difficult. It tends to make the latter rounds of swiss less meaningful (or those are omitted, so we don't actually get anything longer), and it also increasingly punishes players who are performing well, who are forgiven fewer matches than those who squeak into the cut. Do you really want your tournament champion to have lost thrice as many games as another player did?). Eventually, by being more forgiving to losses you make them far less meaningful, and lessen the impact of skill noticeably.
Option 5: Extend the swiss
Rather than shortening the tournament by cutting the number of swiss rounds, do the opposite - increase the number of swiss rounds beyond the number required to crown a champion (especially if you want a smaller cut). While it's impossible to dictate a number of swiss rounds that gives us the ideal "all X-1s make the cut" without capping registrations at specific numbers, or resulting in non power-of-two cuts (e.g. a cut to top 7), more swiss rounds increases the skill-based determination of ranking. It's the reason the proposed tournament changes aren't quite as damaging for Netrunner or Star Wars; they functionally play twice as many games going into the cut, so they are more often going into the cut because they have a stronger record than those missing the cut - the 'arbitrary' tiebreaker (SoS) is used less frequently.
Option 6: Alternate tournament formats
Group Stages, Round Robin (excellent for small tournaments), etc. They're too numerous to list, so I won't bother - we did go over many of them in podcast format here (see the comments too). They'd require a major rethink, and unfortunately many of them don't scale ideally from small to large tournaments - or tournaments with variable entry numbers. I'd love to hear your favourite formats and proposals, but I can't elaborate them all here!
WWID (The Romans didn't have a J): What would Istaril do?
Well, Istaril wouldn't do anything without talking to Thrones players first, and making clear the goals of any change - what are we hoping to solve here? Still, if I could have my pick of systems, I'd favor the following:
Swiss-to-Finals. Swiss rounds are played until only two players with no more than 1 loss remain. The swiss rounds end there, awarding the top "X" (where X is equal to or less than 25% of the field) prizes for making podium. The two remaining players with no more than 1 loss face off in the finals. If one of the players is undefeated, he or she enters the finals with a game in hand (i.e. if he or she loses the game, the players rematch to determine the winner).
This is functionally a pseudo "double-elimination", which is pretty close to the current system (no X-2 made the cut at Gencon this year), but more consistent in several ways.
Advantages:
-The tournament format remains identical throughout, meaning no elimination games and non-elimination games (and no swiss champion playing single elim against opponents who have been forgiven a loss)
-The tournament length, for the majority of players (except 2 finalists) is predictable to within one round.
-We keep a distinct final table for spectators/televising, and all attendees except the finalists are done in time to watch it.
-More games are played (by more players) than in the current format.
-While a second loss elimnates you from the win, it does NOT eliminate you from chasing a high spot at the podium (3rd-nth will be a ranking of X-2 players).
-It scales up easily, regardless of tournament size. It can also be broken down into multiple days without scouting concerns, since pairings are not as predictable.
-It can actually make certain very large tournaments shorter
Disadvantages:
- The exact number of swiss rounds can't always be known at the tournament start (never off by more than one round, though).
- It is longer, for most players, than any of the current formats by an average of two swiss rounds. The length of the tournament itself (including all finals) is not drastically longer.
(A 20 person tournament goes from 5 swiss rounds in the old system, 4 in the new, to 6-7), plus 1-2 final matches, for a total of 7-9 rounds vs 7 rounds (incl top 4) in the old system
(A 40 person goes from 6 old/5 new, to 8), for a total of 9-10 rounds vs 9 rounds (incl top 8) (old)
(A 60 person goes from 6, to 8-9) for a total of 9-11 rounds vs 9 rounds (incl top 8) (old)
(A 120 person goes from 7 old/6 new, to 9-10) for a total of 10-12 rounds vs 11 rounds (incl top 16) (old)
- It requires a reworked pairing/bye system over the existing software to ensure the end condition is reached AND you're not playing the same opponents repeatedly.
An aside: I'd also be sorely tempted to award 0 points for a mod-loss, as this would not incentivize dragging out games unless you think you can win it. This is only suitable if SoS becomes less important (either by changing the system, or by adding a Margin of Victory), though, as we want people to fight out games and not make a concession *more* rewarding than a timed loss.
- imrahil327, scantrell24, Gingerben and 2 others like this
30 Comments
I personnel would go with a double elimination Swiss style event for the tournament. People could get the chance to play for rank and everyone would be playing the same game. #AGoTWith2lives
Which is exactly what I proposed?
I like the system you are proposing, istaril! :-)
since its swiss you just keep playing like you do in swiss know when you take two lose's and know you wont make the cut.
Note, the version I'm proposing works sort of like a double elim (in that 2 losses puts you out of contention for winning the tournament), but you keep playing until the end and as prizes are awarded based on final placement, and everyone from 3rd down will have at least 2 losses, you still stand a chance of making a prestigous placement at 0-2.
Alex, does this system actually produce a different final table than would occur in the final (or penultimate) round of Swiss? Seems like the final would frequently be a rematch of one that occurred in either of those rounds (but I have not done any math on this yet, as I am hoping you have done so already and can respond to these concerns).
Actually, I think I've worked it out a bit and it'll be okay, but since you may have some details or an actual response ready for this kind of criticism I am going to post it anyway.
The system can handle it. There is a question, as I mentioned, of whether you assign byes to the highest standing player that hasn't recieved a bye yet or the lowest, and how strongly you weigh against re-matches within the swiss, and whether you use the bye within the top bracket (as if the other players were eliminated) but while the final being a rematch is possible in some of those systems (if there are exactly 2 X-1, 1 X-0, and the X-0 loses to an X-1 and the X-1 loses to an X-2, then the final is a rematch), it's not forced by the system.
The tricky bit isn't avoiding the rematch, but avoiding the situation in which both of your X-1s get eliminated in the same round (one against the undefeated and one against an X-2), crowning you a champion without a final. This can be avoided by judicious use of a top-bracket bye.
Yeah, you're right - I'm thinking Gencon (which had over 25 more players). Edited to fix.
Very nice system. +1
In addition, you could also have the "podium" players face off to play for 3rd (and 5th and 7th...) place at the same time as the finals to still keep the "made the cut" goal somewhat intact.
When it comes to tiebreakers. If you have such a larger number of swiss rounds, wouldn't the SoS be a much less random number than it used to be? Maybe a system where SoS from early (random) pairings, have lower weight than those from later matches?
I don't know if comparing power totals is a good tie-breaker. Maybe if it is dialed down a bit. A 15-10 win could often be as convincing as a 15-0 win in my experience.
Doesn't double eliminations discourage traveling.
Much less so than single elimination.
Again - the only version of double elimination I'm proposing isn't a true double elimination, it's a continued swiss set of pairings until you have a final table (no more than 2 people with no more than 1 loss). No matter what your record is, you still get to play (on average 1 swiss round more than you used to) to the end, then watch the finals. Note that even the current system (not the proposed one) is functionally double Elim (to win) from 1-31 players, ~55-63. The proposed system is even worse, as it's single-elim to 15, double elim 17-40 (and can be to 44), becomes double elim again at 68...
As for tiebreakers, I understand you entirely. It is worth noting that we already do this a great deal with, say, tournament round time limit (favour certain decks over others). The goal isn't to bias certain decks, but to create a functional tournament environment - people will build the decks for that environment. I'd much rather have a tiebreaker I have some control over than the current SoS IF the cut has to be designed to increase the emphasis on the tiebreaker (as it has in this change).
I think that they should just remove the timed game all together.
What did they do with the math in the new version of the game?
I consider Strength of Schedule to determine who goes to Single Elimination to be broken, since is just random.
I also consider a player with 2 losses winning the tournament while a player with only 1 loss got eliminated broken.
You think both those things are fine? For what reasons?
One reasoning I've heard a few times before was that it adds suspense to the tournament. I'ld prefer a fair tournament system over suspense.
I also think that having a fair tournament system is important for a competitive experience. I do understand that most people are going to a tournament for more casual reasons (I often do so myself), but a tournament system should cater to the people who are serious about tournaments. And going with what istaril suggests doesn't come with any downside for the less competitive group of players (Playing 1 additional game, which you just as well can forfeit, is an advantage in my book).
Also I want to stress that what istaril is proposing is NOT bracketed double elimination.
It is Swiss with (at most) an additional 1-2 rounds and pairings/byes done a little differently. Swiss-to-finals as he calls it. And it gets rid of the two problems I mentioned above.
SoS works fine when you have enough rounds of Swiss, which the current FFG suggestions do not provide.
It's basically saying that someone who played "good players" (for this purpose, "good players" is defined by a player's match record within this tournament) and managed an X-2 record deserves to make the cut, while someone who went X-2 without playing the "good players" does not. In practice, this means that losing early means you get paired against players who also have losses, and your SoS for going LLWWWWWW is much worse than someone who went WWWWWWLL.
One way to mitigate this is to base your tiebreaker only on the SoS of your losses. So if you lose to someone who goes undefeated in the swiss, it doesn't matter if you lose to them in round 1 or round 7.