TST revised

Designed a level that creates new, exciting, and challenging Armagetron game play? Or maybe just something funky. Put it right here.
Post Reply
Match Winner
Posts: 498
Joined: Wed Mar 26, 2008 8:11 pm
Location: Amsterdam

TST revised

Post by Olive »


Do you start sweating when you see Nanu spawn in front of you?
Ever felt like ragequitting as your's truly digs his way into your space at the start of the round?
Have you ever been stuck at 1980 points for 4 rounds by courtesy of Andrei?
Has Force ever holed your wall in favour of a third team?
Does doov randomly 4.0 dig your wall in order to cuck you?
Do you pity wolf as he's stuck in a box for the majority of the round?
How do you think apple feels when he hits three consecutive pt's in one round only to be rewarded with a measly 50 points?
Do you feel envy when a team munches 180 points for free?
Does it frustrate you when 6 players beeline towards you from the start?


Then I might have the perfect remedy for you.

The TST map has been a community favourite for competitive sumo games for as long as I can remember. It is considered the holy grail of sumo maps. Every minor proposed tweak has consequently been answered with religious zeal, as shown by this seven-page discussion about hole points. While there have always been complaints about the highly repetitive starts and the high degree of randomness at the later stages of a round, recently there has been a different kind of complaint that echoes over the grid. The 2v2v2v2 setup promotes opportunism to such an extent that creating temporary alliances aimed at the leading team has been an evident part of the meta, often to the frustration of the team in lead. Spamming pickups for over a year has lead to a significant improvement in the general understanding of the map and the various ways to approach the start. In turn, players have become more apt at gangcucking. Blatant unfavouritism can ruin a player's experience, especially in situations where it's not in the best interest of the sender to do so.

So how can this point be tackled while keeping the spirit and feel of the game intact? A simple yet dramatic solution: reduce the number of teams to two. In doing so, the entire aspect of unfavouritism disappears. No more third party enjoying free space and points after a succesful kamikaze-attack. Instead, any decisions to deliberately cuck a specific player will be viewed from the risk-reward paradigm of the team. Increasing the number of players on a team also bolsters the teamplay aspect, as it allows deviation from the current two rigidly defined roles.

While there are numerous drawbacks to this proposal (the first and foremost being that it's blasphemous), I hope some of you are as curious as I am. In case you are, please drop a reply. Hopefully it provides some leverage to test the idea out in pickup and replicate the success of Twister.
Olive a.k.a ZeMu, MoonFlower & chicken.
User avatar
Core Dumper
Posts: 124
Joined: Sat Aug 16, 2008 11:16 pm
Location: Desert, Arizona

Re: TST revised

Post by Shock »

Jesus, yes. Please remove all multi-team (> 2) game-modes. It feels like it's been decades since any of it made sense to me.

I have had a long internal quarrel with the way TST has been played out since its inception. Granted, it has been very successful at generating an environment where we have an intense sumo zone, the best players are all fighting it out in a very cramped space, and there's even some team coordination involved. Not only can you just be insanely good at sumo, but maybe you can also develop some tactics that help you win in this mode. Sure, and we saw tournament after tournament of the most stacked finals lobbies. That should be considered a success, as the format allowed the best to go up against the best. But consider this: five minutes into the finals was it ever really a surprise who was going to win?

There's a huge flaw to how those matches always end up playing out. Inevitably, one team takes the lead and becomes a frontrunner. What should happen next? What does happen is one of two things:
1) every team continues to play exactly the same as they were before. This is the highly sought after "braindead comeback". You are convinced that the way you play will eventually result in a big fat W for you and your teammate.
2) a subset of teams, let's say the team in second place and team in last place, tries to do something about being behind. They get aggressive (maybe on each other, hopefully on the team in first) but fail to close out rounds. The lead grows because the team in first made no mistake - their neighbors all died and they got to envelop half the zone. (Or maybe they did make a mistake and another team clawed back - is this all that interesting though?)

Let's think about what went wrong in case 2. Did the team in 2nd try to pressure the team in 1st? If so, what did every other team do? Did they apply pressure in turn (4th pressures 3rd, 3rd pressures 2nd, 2nd pressures 1st) to form some sort of shit centipede? If there is a cucklord in your lobby the answer is probably YES. "See zone, take zone" is the mantra that all sumoers have long lived by, regardless of whether you are in first or last. This is exactly why this game-mode does not work. Taking space away from a team that is trying to eliminate the team in 1st.... it happens every single pickup, every single tournament, and it serves no other purpose than further removing the potential for any real comeback to occur. Any team that tries to make a play to get back in the lead is almost always instantly punished for attempting to do so. Like waves crashing against a cliff, one wave on top of the next.

Olive, what you are describing as a reason to stop playing TST (Does it frustrate you when 6 players beeline towards you from the start?) is closer to how this game-mode should have been played all these years. Here's why:

The entire span of this game-mode's life is plagued by the instantaneous incentive to get 30 points and share zone points with the remaining teams. It's high time average joe sumoers like me (people who don't make finals, maybe win a tst pickup once a month) realize that this does not result in big fat Ws when the lobby contains 4 skilled teams. Say you kill the team in 2nd, leaving a big void in the middle of the zone as the walls start to fade. Who is going to get it? You, Capone? You suck and are probably dead already. The team in 1st? That's a better guess.

Let's say teams are endowed with a per-team match win probability in a "void" - as in, if each team has half the zone just sumoing it out. The team of Mr and Apple has p probability to beat the team of Capone and Destiny, who have a probability q of beating Mr and Apple. If p > q then what I am saying is that a comeback in tst is likely if Capone and Destiny (C&D) are early frontrunners and a comeback is unlikely if Mr and Apple (M&A) are frontrunners. The latter would only happen by chance, all else equal.

Now, say all 4 teams employ a default strategy - each team gets a box equal to a fourth of the zone, perfect split. The probability each team has of winning is an order statistic depending on the pairwise win probabilities of each team. The result is that the team with the highest order statistic is likely to win. C&D, with its lower pairwise win probabilities, ride the curtails of opponents with better head-to-head probabilities against M&A in order to get some points but not enough to actually contest.

What can happen intramatch to change this result? What happens in other sports where one team is expected to beat the other? A big aspect of a comeback, or even just parity at all, is luck no doubt about that. But sometimes.. a team is outplayed or outstrategized. Does that actually ever happen in tst? No, because any counter strat is incredibly risky due to the existence of two other teams in there with you clawing for 1st. No one is willing to counterstrat.

And because no one is countering in any way, the ex ante best team becomes the ex post winning team.

What needs to change? Lower the risk of applying pressure, or sumo sweepers need to get one million times better. Theoretically, teams should be able to communicate moments of pressure (hey I am greeding for space near M&A for the next 3 seconds so cover my tail). The game just moves really fast so unclear how viable that is.

Back to the earlier point: What should happen next?
The three teams not in 1st place should apply pressure on the team in 1st place, with the teams spawned as neighbors to that team applying the most. The non-neighbor team holds space and guards until 1st place is eliminated. The reason we never do this? The strategy employed by the non-neighbor team doesn't feel like a winning strategy. Because NO ONE has ever won that way. I think we are sometimes too quick to dismiss things that don't work out. Imagine you are that non-neighbor team and then 2nd or 3rd place team decides to cuck you. It literally ruins everything, as there is now enough space for M&A to survive in the zone no question. You are better off siphoning space from the aggressive neighbors.

Countering a dominant team takes the implicit coordination of all three other teams. That's the entire rub. That in a noncooperative equilibrium every team fends for itself, but in a cooperative equilibrium we see 1st place targeted. It is almost impossible to enforce that kind of cooperation.

TST is a mess, pls fix.
User avatar
Match Winner
Posts: 733
Joined: Sat Jul 07, 2012 3:07 am

Re: TST revised

Post by delinquent »

2v2 feels somewhat barren. However, I wonder what 4v4 would bring to the table. We could even adjust spawn points somwhat to cater for such a game - and in doing so create a new sumo format that might be a little more fun to play.

I know it sounds like WST with extra steps, but hear me out.

in a 4v4 scenario, it is functionally almost impossible to occupy a zone with all eight players. This implies that some strategy will need to be created on-the-fly in order to produce a result. Ideally, this would be in the form of one or two players occupying some zone space, with the rest of the players attacking the opposing defenders (and each other, presumably). The difference between this and WST is the intensity and speed at which this takes place.

Assume, for a moment, that we start with each team split into two groups, with both groups facing their opponents, just outside of the zone. The logical course of action is for the near-side players to secure zone space, perhaps even attempting to centre the opposing players, with the off-side players securing the rear of the potential space. At the same time, the near-side players must co-ordinate with each other in order to not only secure space quickly, but safely. It would be easy at this point to kill a team-mate.

Once space has been secured, the offside players are free to make their attacks. In fact, they aren't even obligated to stay within the boundary of the zone - much like the current TST format. Their best course of action is to get inside the opposing offside payer, preventing them from encroaching on the space that the nearside players have secured and are protecting. This is where the fighting occurs - offside players must war for supremacy, and that outcome is a deciding factor in the round scoring. However, the nearside players can also attempt to push their way into not only the zone centre, but also the space of the opposing team. Again, very similar to TST, but with some extra steps and more cooperation.
Post Reply