Live data from Hacker News

The Miracle of Pakistani Tekken (2022) [video]

youtube.com

1–10 of 49 posts

Re: The Miracle of Pakistani Tekken (2022) [video]

#3
Wow this was really interesting. As a semi-competitive player during the Tekken 3/TTT1 days one of the under appreciated parts of getting good at an international level is diversity of arcade "cultures".

What this means is that your arcade generally provides the same players again and again, and performing well against those players provides little guarantee against players in another arcade who've developed entirely different techniques and playing styles.

So you have to play, a lot, against a huge range of players. You can't really just play against your arcade's local players and get to a level that's competitive on even a regional level. I remember the first time I emerged, after a few months of heavy playing, into an arcade in a neighboring city, and getting crushed the first few matches.

I've since played TTT1 in arcades around the world and I've always found somebody who does something I've neve seen before.

That these guys can pull off winning, against ultracompetitive leaders in ultracompetitive hypermarkets like Japan or Korea is absolutely incredible.

It'd almost be like Pakistan fielding a Superbowl winning American Football team after only getting together enough guys for the offensive and defensive lines and only playing those lines against each other.

Re: The Miracle of Pakistani Tekken (2022) [video]

#6
I was a (semi-)competitive player two decades ago (Tekken 2, 3, TT) during the golden age of Tekken; for the uninitiated, the challenges are multiple layers:

1. The vocabulary of the game: there are two dozen or so characters, each with dozens of moves. Each move comes with its own attributes (e.g. hit points, latency). If you are anywhere near competitive, you need to memorize all of the moves and attributes, in the hundreds. Some of the detailed analyses count the latency in frames (60 game frames in a game second), and a few frames difference does make difference in match results.

2. The grammar of the game: the game boils down to predicting the opponents next move and the best to counter it. There is a path dependency in the game but it's weak i.e. what your opponent did ten moves ago may not affect your current decision aside from the current state of the game (hitpoints left, location on the stage). Weak path dependency means that you can model the game as a Markov process

P(opponent's next move| your character attributes, opponent's character attributes, opponent's style, game state)

If you are competent, the only unknown on the right side of the pipe is "opponent's style". Based on your observed probability of the opponents moves thus far in the game, his regional background (i.e. Japanese), etc, you are supposed to estimate opponent's style. The more accurate the estimate is, the better you can counter or parry his moves.

3. The meta game: Once you know the opponent's style and the opponent knows you (given everything else in the probability function above), what should you do? You need to choose a different character, so that the opponent has to learn your style again. Any learning comes with its sampling cost, and your goal is to deal maximum damage while the opponent is learning your style. In order to do so, you need to prevent any kind of shortcut in opponent's learning, e.g. transfer learning. Transfer learning is possible if your new character is similar to your first character. Thus, you need to master completely different characters. Another is to adopt a completely different fighting style for a given character, but this is easier said than done -- people are surprisingly predictable unless something (i.e. different character) forces them too. Another aspect of meta game is the psychology. Players get stressed, nervous, shocked as humans do, and make sub-optimal decisions.

Overall, the game is fun but surprisingly complex and competitive. Of course, I wrote above can be applied to other games and 'games' (e.g. starcraft or trading). Knowing how difficult it is to compete on the international level, I am truly impressed by the Pakistani team.

Re: The Miracle of Pakistani Tekken (2022) [video]

#8
post #5
post #4

The name of the province - it's pronounced pun-jab, not poon-jab. (pun intended)

Also Mos-co not Mos-cow :)

Is that really true? In Russian it's Moskwa afaik so neither of those would sound correct anyway. In German it's Moskau, which sounds like Mos-cow.

Re: The Miracle of Pakistani Tekken (2022) [video]

#9
For anyone not in the Tekken scene (the video explains better than I will) Pakistan came out of nowhere and took the world by storm. Usually, it was Korean and Japanese players that dominated, with a few Americans sprinkled here and there.

But little did we know, Pakistan was brewing up a new type of player that we had never seen before, and were completely defenseless against (regarding the meta at the time). Arslan was and is a force of nature. The funny part is that when he first began to get fame, he made it clear that he was in no way the best player in Pakistan.

Top level players in Japan and Korea now regularly fly out to Pakistan to practice.

Re: The Miracle of Pakistani Tekken (2022) [video]

#10
post #5
post #4

The name of the province - it's pronounced pun-jab, not poon-jab. (pun intended)

Also Mos-co not Mos-cow :)

The Anglicization is mos-cow.

Using Cyrillic it's pronounced in neither of those two ways that you've written.

Post reply on HN