Thoughts on using Double Elimination for the Top X cut

Tournaments and Events
1

After reading the tournament rules when the format for the Top X cut was changed, I was skeptical. Initially, it struck me as a less-than-stellar idea, but I was wary of an unjustified knee-jerk reaction - after all, I’ve been organizing Swiss-based tournaments (with the occasional top 2 duel every now and then) for a long time now, so there will presumably be at least some degree of bias. Therefore, I sort of shelved it in the back of my mind until I could gather some empirical data.

Fast forward to this saturday - I finally had the chance to observe a Top 8 double-elim in action. I’m less than impressed. From where I’m standing, there’s a bunch of downsides and I’m having a hard time coming up with upsides that would make it worth it.

  • It’s pretty non-transparent - I dare you, try to make sense of a top 8 double-elim just from watching the 8 finalists sit down and play, without seeing the chart. I couldn’t, after the first round was over. Making this appealing to spectators probably requires a data projector or some other method of making the whole structure available on-demand to them.
  • Intangible stakes of any particular match - Basically, without seeing the spider thingy, you have trouble telling how well your buddies are doing. This is mostly a trait of the whole “one of the finalists progressed into the final through the loser’s bracket” idea which messes up the clarity of single-elim. What that ends up doing is reducing spectator interest - hell, at saturday’s regional, a full half of top 8 was part of our expedition, and yet we ended up not watching them play. You can’t tell “he’s playing for top 4” or something similar - he will be top 4 if he wins, but he might very easily be top 4 if he loses, too.
  • Irregularities with the sides you play - This is admittedly an issue mainly in the first match, but seems to me like a pretty bad one. Since the higher seed gets to pick which side to play, presumably he’ll pick the stronger one. This is all fine in a reasonably balanced environment, but invites trouble if that’s not the case. Being eliminated simply because you didn’t get to play corp at all (for instance) due to a coin flip in the second game seems pretty * “meh” *to me, personally
  • Him again?! - playing the same opponent up to three times in the process of a single top 8 is probably something you’re not really keen on doing after 5-6 rounds of swiss, in which you’ve more than likely faced him once before already
  • Variable depth of the match-up tree - This is probably the worst offender. You need an equal number of victories total to get to the final, regardless of the bracket you came from, I get that. But that means not only that one of the people will have played one more match, it also means that a fair amount of waiting around “for previous games to finish” takes place, making it practically impossible to structure the finals into rounds in a transparent enough manner.
  • You don’t get screwed out by one bad draw (it takes two). As a frequent traveller to agenda-floodville, this one I can get behind. There’s naturally the issue that you’re only playing half-games now, but… whatever.
  • It’s faster…presumably. But… the speedup is the result of “you only play 1 game against an opponent”, rather than an actual property of the pairing/resolution method. If you really wanted to leverage that, you could change the swiss part of the tournament into “an even number of rounds, you play either runner or corp” - SW:CCG tournaments were using this format and it was totally fine. You could cut the time requirement for a tournament almost by half, and the results wouldn’t be too skewed, I feel.

Am I missing any upsides? If so, are there any resolution methods that actually provide them, without all the bad baggage double elim is bringing to the table?

Let’s hear some feedback. Thoughts? Experiences? Suggestions? Insults ( :stuck_out_tongue: )? Hit me!

4 Likes
2

What I like about double elimination is that it avoids the anticlimatic finish of so many tournaments under the old system. Under the old system, If I manage to win the first game while my opponent scores only 1-2 points, I’m basically a shoe in to win the match, with the final game pretty much a foregone conclusion. It’s kind of anti-climatic to shuffle up for a last game, only to have it end second turn when my opponent biotic labors a vitruvious, or snipes a project beale off R&D. Under the new system, each game starts from scratch, avoiding this kind of let-down endings to the tournament.

3

That’s pretty easy to fix, though - just say you’re playing win/loss only, with no concern for agenda points. In the case of a 1-1 split, either play a third game with sides determined randomly, or use a tie-breaker (like, previous match results from swiss or whatever).

This seems like an especially natural solution now that we’ve effectively gotten rid of the match structure in the swiss parts of the tournament, too.

2 Likes
4

I have played a few double elimination rounds already and I can say, that I really don’t like the format at all. I agree with pretty much all your points, with the exception of speed. I don’t think it is really faster with all the wait involved between some matches.

The huge question for me is anyway: Why have a cut to an elimination system?. Just play swiss again in the cut. For big tournaments it makes sense to split the event over two days anyway. That would allow for extra side events to be played the 2nd day.

I get it, that people might find it fun to watch a final. However, I have never seen anyone stick around to just watch. Other then those that pretty much had to stay because they shared a ride or were the TO, etc. Not to mention that I don’t really like anyone watching me in such an important game. Spectators tend to give away too much information to the opponent without even noticing. Which makes the idea even less appealing for me.

1 Like
5

For Netrunner I don’t like the elimination phase at all. I feel it can too easily skew the hard earned swiss phase results.

That said, in my opinion DE is much better than SE. The format is being used in fighting game communities for ages and delivers exceptional thrills for spectators (main contender got tossed to the losers bracket in the first round? watch his epic battle through the losers bracket along with two wins in the final against the same guy that tossed him down!). Also fighting games spectators don’t have any trouble recognizing what is going on in the elimination structure, this is complicated only during the first encounter with the system.

But Netrunner is not Street Fighter. It is slower and asymetric. DE is an improvement over SE. But both are unnecessary here.

3 Likes
6

I would prefer swiss without a cut, using either better tiebreakers or more rounds. Maybe a top 2 game if they are tied. However, if there has to be a cut, I like the current system better than the previous one.

For Swiss tiebreakers, weak side wins is hilariously bad. It just makes some wins count more than others, whether they were difficult or easy. Strength of Schedule is at the very least indicative of how tough a player’s tournament has been, and should be the first tiebreaker after prestige.

1 Like
7

Funny, I on the other hand feel like swiss is skewed :slight_smile:
Initially random match ups, byes and games going to time :frowning:

2 Likes
8

I used to think the same way. However, strength of schedule can be manipulated by others (dropping out of the tournament -> sos 0 -> psuh a friend in the top 8), while weak side wins doesn’t. I don’t like weak side wins at all, though.

I would really like to see an extra prestige though if you manage to win both games in a match. Reducing the ridiculous amounts of ties in the ranking of a tournament and punishes players that try to draw in the top 8 once they have a sufficient amount of points.

1 Like
9

I don’t understand why you are so concerned with making things spectator friendly / transparent. Seems like that shouldn’t be a priority, especially given how late these events tend to go.

I would add to the cons: the higher seed only needing to tie can be pretty dumb.

I like the new system because it’s better than the previous one, but it’s entirely possible that there are better systems out there (though I don’t think I’ve ever seen/heard anyone suggest one).

10

Not sure about that, doesn’t strength of schedule rely on random chance quite significantly? The SOS from your first couple of opponents is essentially random. It can also suck if anyone has to drop for whatever reason.

I agree that Swiss league is better than elimination. I like the idea of a final, but oftentimes the top table in the final round of Swiss can be that final. I think weak side wins is a pretty good tiebreaker. The nature of tiebreakers is that some wins are more significant than others. If someone wins more games then you don’t need tiebreakers. Swiss league and elimination aren’t too different in truth. You come top of the pile by winning a lot of games in a row. The most significant improvement of the double elim vs single ‘matchplay’ games is that agenda points aren’t worth anything unless you have a winning score. Games where one player only needs 4 points aren’t particularly interesting, and they make some decks horrendous.

2 Likes
12

Agreed.

Perhaps it would be better to play more rounds of swiss and then cut to a double-elimination top 4? Then only one person ever needs to wait for a loser’s bracket match one time.

13

I actually couldn’t disagree more with many of the sentiments here. I think there are two complaints many players are having here, so I’ll try to give my support to them individually.

  1. Elimination Rounds:
    I am a huge fan of elimination rounds! I believe the height of any competitive event should be an elimination round. This is as true in Sport as it is in Netrunner. The champion should always win their final game in my mind. There should always be the pressure of “win or go home” at some point in the tournament. Every player should know “if I win this last game, I’m the Champion.”

Swiss play just doesn’t give that pressure, since the final result will often be dependent on more than one game. Worse yet, if we have complaints about the imperfection of tiebreakers NOW, imagine if it decided the winner of an entire tournament!? Even worse, it would realistically remove a huge percentage of the players from championship contention halfway or three quarters through Swiss. There is something to be said for scraping through Swiss to sneak into Top Cut and then rampaging through the Elimination rounds to win the whole thing. It’s fun. It’s dramatic. It’s exciting to be a part of and exciting to watch.

Finally, and this will probably be a point of contention to many of you, the best player will not always win. This is a huge boon for me and the entire point of a tournament in the first place. We all play against players who will beat us 8 times out of 10 or decks that beat us 8 times out of 10. So why play a tournament and not just do a big league, clearly identifying the player who has the best win percentage? If you find this as displeasing a thought as I do, you should be a fan of Elimination Rounds, at least in large tournaments. Elimination rounds throw out the percentages. It doesn’t matter if you’re in your worst match-up or playing your greatest rival. You just need to win 1 game. Variance will ALWAYS be a part of more than just your decks. It’s a part of any competition and should be embraced.

2)Double Elimination:
There is no doubt that double elimination is brutal. Until I hear (or think of) a better option, I will support it though. In response to intangible stakes and transparency, these are mainly an issue of unfamiliarity with the structure. I’ve now played in two Regional Double Elimination brackets and clearly understood in every game what I was playing for. If I win my first game, I’m in the top 6, great. If I win the second I’m in top 4 and get an Andy, amazing! Agreed, it gets slightly convoluted once you hit the Loser’s Bracket. Still not to the point where a player intelligent enough to play this game at a high level cannot figure out.

As to irregularities on the sides you play, there is no doubt that playing losing on two consecutive Corp/Runner games is painful. That being said, at most 1/4 of Top Cut can face that problem. In addition, it promotes parity in your Runner/Corp strength (as does WSW). I’ll concede that it’s not ideal, but it’s an acceptable negative of the format.

Playing an opponent three times in Double Elimination is ONLY possible if you are both in the Championship game and it goes to a final “winner takes all game.” I don’t see a problem with two of the clearly top competitors of the day squaring off an extra time. In fact it also necessitates that the players split the first two games in Elimination play. A third game, creating an effective Best-of-3 sounds ideal rather than a negative.

The depth of the match-up is multiple issues. Firstly, if you lose your first game it will take MORE wins to reach the Final. You may like or dislike that, but it as an inherent part of Double Elimination. It actually puts higher stakes on the early games of DE, which I quite like. Winning early and often is clearly an advantage, and I think rightfully so.

The second issue about waiting is a problem, I’ll agree. I’ve also been on both sides of the Championship match, once having won the Winner’s Bracket and once having won the Loser’s Bracket. It’s tough for different reasons on each side. Once again, until presented with a better option, it’s acceptable.

Conclusion:
I am a huge fan of the current system. I think Weak Side Wins as primary tie-breakers are a good choice (promotes parity). I think Elimination rounds are absolutely necessary at large high level tournaments. They are simply more dramatic and decisive. I also find them much more fun. Double Elimination, despite it’s flaws, is the best solution I’ve played/heard of for Netrunner. It’s a brutal format, but it’s rewarding to those who slog it out. Winning the Finals from the Loser’s Bracket SHOULD be more difficult than from the Winner’s Bracket. It’s not impossible (already happened at a Regionals iirc), just properly difficult.

4 Likes
14

If two players of equal skill meet, shouldn’t there be two 50-50 games, in which case the disadvantaged player will sweep 25% of the time? This seems broadly accurate.

16

I have seen maybe 2 or 3 regionals reports, so I’m not sure. I do suspect that the winning player tends to be better than the 2nd place player (for obvious reasons), so I think the actual incidence of wins from the losing bracket will be less than 25%. Still, I’d be quite surprised if it were 0%.

18

The regionals in Toronto saw a win from the losers bracket, the winner has a tournament report here: 1st place Toronto Regionals @ Hairy Tarantula North (41 people) | Android: Netrunner. So yes, it does happen.

2 Likes
19

I don’t know what to do about the tournament structure of this game. Not having a cut to elimination in a high profile event makes no sense to me, so that is off the table as far as I am concerned. The old format sucked. I was an extreme beneficiary of it at Worlds last year, when I won my first 2 die rolls, and then won the runner games 7-0. All I had to do was mulligan into Breaking News, and that isn’t fair. The new system… I think it is better? It sure as hell isn’t faster, anyone who says it is hasn’t played under both formats enough. One issue with the double elim, is at a cut to top 16 in NC, I won out on the top bracket and had to sit out THREE ROUNDS waiting to find out who my opponent was for the finals. It also just takes a long time in general, as “half rounds” somehow don’t actually take half the time of a regular round. I also think something needs to be done about the scoring in general, as there are way too many people sitting at 12-14 points going into the last round of a 16 person tournament. It is odd that someone could be in 4th place, and then lose both and not make top 16 (I also think Intentional draws should be allowed, but that is another can of worms).

tl;dr The old tournament structure was really bad. The new one is better, but certainly not perfect and takes WAY too long.

For what it’s worth, both regionals I have won I did so from the top bracket, one of them I won all the games, the other I lost the first game of the “finals” and won the second.

3 Likes
20

At the Seattle Regional, the same two players played each other three consecutive times to end the tourney, and the same player played corp in all three games. GG FFG!

21

How is it that the same player played corp all three times? That doesn’t sound possible.

3 Likes
22

Yes, your early opponents are assigned randomly. If they are good, you are more likely to go 2-2 or even 0-4 against them, and you get compensated with a higher strength of schedule. If they are bad or mediocre, you are more likely to 4-0 them and you get a lower strength of schedule because your win was easier.

Given the choice, I would much prefer the “free” prestige and low strength of schedule that a bad first-round opponent would give me. Strength of Schedule is essentially a consolation prize for having to go up against a good player early.

The problem with Weak Side Wins is that it rewards you even more for playing against bad players, because going 4-0 is always going to increase your tiebreaker, no matter how bad your opponent was.

1 Like
23

It’s possible, albeit a very fringe case. For instance, take the ongoing OCTGN 3 double-elimination final (bracket here). Dawarmage has played Corp 3 times and Runner 1 time. I don’t have all his data, but I believe Nordrunner has played Corp 1 time and Runner 2 times. Dawarmage played Nordrunner in the second round. I’m not sure who played which side, but let’s say their singleton plays (Nordrunner’s Runner game and Dawarmage’s Corp game) were against each other. Based on what sides their opponents have played, it’s theoretically possible (if highly unlikely) that their sides continue to stack unevenly. They could match back up in the finals, and their sides are so distorted and uneven that both games (if necessary) would involve them playing the same two sides they played as in their first game against each other.