PSX NHL Blades of Steel 2000. Fastest win. Gameplay input lasts 90 frames.
- Emulator used: BizHawk 2.9.1
- Genre: sports
- Uses hardest difficulty
- Goal: Win the game (exhibition mode)
- Aims for fastest time (end input early)
- Manipulates luck
Changes made in the options:
Difficulty: All-Star (maximum)
Goalie difficulty: All-Star (maximum)
Period length: 2 minutes (minimum)
It takes 4 frames to change the period length from 10 to 2 minutes. You could theoretically save 4 or 2 frames by leaving the period length at 10 or 5 minutes. Big improvement is improbable because almost all games will be lost, so you would need to waste some frames to adjust the RNG.
It takes 3 frames to change the opponent team twice. These frames could be reduced if you can find a winning solution with another team composition. I did not test many combinations of teams, because a satisfactory winning solution was found, and I don't want to play the game anymore.
If you like to do botted runs, then this run can be used as a template. The bot needs to record frames 4113 to 4203, so there are 90 frames of theoretical gameplay improvements to be made. The goalie seems pretty good with the maximum setting, so I don't know what kind of timesave to expect from a truly optimal run. There are 4 different shots that can be made, and most frame-by-frame movements affect the shot outcome, so the number of possible combinations increases staggeringly. One difficulty with brute-forcing is that we don't know how long we should wait after the shot to confirm that it was not a goal. Is this problem similar to the halting problem in computer science?
Before starting the game you can leave the controller icon to the middle, so you select no team for the human player, but I don't consider this legit because then both teams are controlled by the computer and there is no indication of which team is ours.
A fun aspect of the run: It looks like the computer was programmed so that the approach is offset to the front of the human player because the assumption is that usually the player is moving forward. The result against the idle player is that often the computer is just skating in front of the player not achieving anything useful in the process.
eien86: Yet another entertaining use of god-like hockey goalies. This one is well executed and, although there might be room for improvements by means of botting, looks pretty much at the limits of manual TASing.
Accepting to Standard
