Submission #1201: Neofix's Genesis Side Pocket in 03:17.55

Sega Genesis
baseline
GENS
11853
60
513
Unknown
Side Pocket (JU) [!].zip
Submitted by Neofix on 9/15/2006 7:52:00 PM
Submission Comments
Side Pocket From Wikipedia, the free encyclopedia
Side Pocket is a Billiards ("Pool") video game made by Data East in 1987.
During play, the player may put various spins on the ball, such as "English", "Massé", or perform a "jump shot".
The single player mode, called "Pocket", requires that the player clear all the balls off the table in a limited number of shots, as well as achieve a predetermined score. The player earns points by sinking balls, sinking balls on consecutive shots, and sinking balls in numerical order. On occasion, a pocket will light up, and if the player sinks a ball into that pocket on that shot, a bonus will be earned in the form of points, extra shots, or a chance to earn either by doing a "trick" shot. In the 2-player "pocket" mode, the game plays similarly, except the two players take turns; if one player does not pocket a ball or scratches, control of the ball changes to the other player.
There is also a Nine Ball mode, but it may only be played with two players. A variety of "trick shot" challenges are also available to the player, requiring the player sink all balls into select pockets using a single shot.
The game had two sequels, Side Pocket 2 for the Sega Saturn, and Side Pocket 3 for the Sega Saturn and PlayStation.
Trivia If you win all the "Trick Games" it is a puzzle of a red-hair woman.
  • Autor notes: Trick Game complete in 3:17min, as fastest possible
See this movie in Youtube: http://www.youtube.com/watch?v=gkH2qfGOJO4

adelikat Rejecting, mainly because there is nothing surprising or unexpected. All the solutions are predefined by the developers and so nothing is exploited. This makes it not so interesting for a TAS. On the upside, it was optimized and not boring. But ultimately not TAS worthy.

adelikat: Unrejecting in consideration for the vault.
FractalFusion: Judging.
FractalFusion: Rejecting because it is not deemed entertaining, and it is a sports game, so it doesn't fit in any tier.
Noxxa: Unrejecting again for consideration of publication to the Vault following a rule change regarding sports games.
Samsara: I'll look into this one.
Samsara: This is an oddball (lmao) of a submission. As adelikat mentioned ten years ago jesus christ, the solutions are predefined and thus nothing can really be exploited, but after some testing it goes deeper than that: I've managed to get multiple different solutions (as in angles/power levels) that all end on the exact same frame. Granted, these solutions aren't drastically different from each other, in fact barely noticeable during playback, but it more or less proves that you can't really deviate from the intended solutions at all, and that in theory it is actually possible to get a run that cannot be improved at all with relative ease (note only 500 rerecords with fair optimality). The solutions are so precise that in some cases it's actually frame perfect (I mean in the game, as in if your positioning is a pixel off then it doesn't work), so ultimately this run wouldn't look any different from an RTA, it would just get to the exact same solutions faster. Looking up a guide confirmed my suspicions about how linear/trivial the game is: The guide gives exact solutions through telling the player how many times to tap a direction to get the angle right, and since all the angles are set and the physics are consistent it will always do the same thing every time. That, combined with the right power and spin setting, will get you the same solution 100% of the time, and since the guide gives you everything you need, that's your lot.
There's a problem, though. Under the new rules for Vault:
  • The run must not be seen as trivial.
    • If a run consists of doing a trivial strategy, made only nontrivial by having to do it over the course of several rounds, it will still be judged as trivial. For example, bowling games where the player gets a strike with ten pins every time.
I could argue that this game falls under that trivial nature, but there's the counterargument that this is less of a sports game and more of a puzzle game, which always have one intended solution and a variety of ways to get there. We have a published Sokoban run, for example, which is fairly set as far as maze solutions go, in that knowing the solution as an RTA player means the run could be matched almost perfectly. This is the same way.
What pushes it further away from being trivial is the fact that the developers gave intended "routes" but basically left free reign on how to accomplish them. For example: The intended solution for a level would read "Aim the cue at the left ball, give it some spin so it curves around and hits the right ball", but you can adjust the power and the amount of spin and even precisely where you aim and it would still work. The first stage in this run could be improved by 200 frames just by removing the spin from the ball, and stage 11 could be improved by about 2 seconds with a slightly modified strategy. The actual stages would look nearly identical between this run and an improvement run, but the stages would end quicker due to the cue ball stopping its movement earlier. The low re-record count leads me to believe that there wasn't any testing done on other strategies, and the two strategies I found as improvements over this run were in the guide I found, so I feel like those should have been checked at the very least. On a more pedantic note for improvements, a number of frames can be saved on the opening menus by not auto-firing Start and pressing Up instead of Down to select the Trick mode.
I'm rejecting this run (for good this time) due to the lack of testing and research. I'm still completely on the fence about the game's triviality, so I don't know the fate of an optimized version as of yet. I feel like an improvement could break the triviality barrier as long as it's well-researched, but I'd have to see one to be completely sure of that.
Last Edited by adelikat on 9/10/2023 1:55 PM
Page History Latest diff List referrers