This topic is for the purpose of discussing #4291: morningpee's A2600 Seaquest "maximum score" in 00:06.02

That was fast finding that glitch!

Voting No. Nobody wants to sit there listening to buzzing noises for 22 minutes.

Voting No. Nobody wants to sit there listening to buzzing noises for 22 minutes.

After 9 minutes and 15 seconds, it starts changing colors. Entertaining colors.

I did enjoy the length of the movie file, so I'm voting Yes. I only watched part of the rest of the encode, but I have to hand it to scahfy: the changing colors are pretty mesmerizing.

6 seconds and it ended up being one of the more boring Atari TASes I've ever seen in my life. No vote.

This is a rather difficult, albeit interesting, case. Should entertainment be judged based on the gameplay that happens during the input (ie. the first 6 seconds), or should it be based on the entirety of the time that's spent gaining the maximum score (ie. 22 minutes)? Does the run "end" (and thus be evaluated in terms of entertainment) at the 6-second mark or at the 22-minute mark? Yes? No? Maybe? Screw this, I'm going home?

I think this shows that the idea of ending input early is stupid and arbitrary

Should it be noted that "maximum score" is not a vaultable goal?

Should it be noted that "maximum score" is not a vaultable goal?

Does this game even have some kind of ending? Is it at all eligible for vault?

Should it be noted that "maximum score" is not a vaultable goal?

Does this game even have some kind of ending? Is it at all eligible for vault?

I'm not sure. I was intending to test out how the game works in that respect. Either way, if it has a way of ending/losing/resetting, then this run also fails to properly end the game, much like how Virtual Pinball was rejected. Oh, and also, the link to the glitch finder leads to a video that reaches maximum score over 15 minutes faster than this run does. (7:02 vs 22:53 is quite a big difference.) How to end movie input is up to the author of course, but either way, this run is way behind realtime records. EDIT: The same run also dies and the game resets shortly after reaching maximum score.

As the submission describes the game does have sort of a kill-screen which is achieved by getting 999999 points. That could be considered the game end, which then makes the game vaultable. While setting up the glitch in 6 seconds is impressive, I am voting against publication as this game shall be timed like we time Super Mario Bros. imo. Thus rendering this TAS MUCH slower than an RTA.

I found the glitch funny (and thus entertaining), though I certainly didn't watch more than about 30 seconds after input end. You can reach max score 15 minutes faster though? But this at least shows off a neat trick to come up with a clever early input. So there's some value to doing it this way.

I wonder if it would reach max score faster than the RTS if you somehow manage to make it come in contact with the fish non stop.

The TAS's length is judged by the end input. Thus if it is possible to make the score reach 999999 in less than 22 min 53 sec, but it requires more than 6.02 seconds of input, then sorry, that counts as a slower run. However, this is boring regardless of whether you judge it on the six seconds of input or the full movie, so a no vote from me. And therefore reject, due to the unvaultable goal choice.

The TAS's length is judged by the end input.

This is false. The length of a TAS, for comparison purposes, is judged by the time required to get the game in a "completed" state. For example, Super Mario Bros. is judged by the time it takes to hit the axe. If a TAS takes longer to hit the axe, even if it has a shorter input time, a judge will reject it.

It works both ways: http://tasvideos.org/JudgeGuidelines.html#InputLength

It works both ways: http://tasvideos.org/JudgeGuidelines.html#InputLength

That doesn't address how to compare with different runs (in particular, real-time runs), though.

It can get tricky how to compare two runs that end input differently. Let's say someone makes a smb1 warps where the only "improvement" is ending input earlier. Well, that's not an improvement, that movie would be rejected. But let's say a movie ends input early, and does so in a clever manner (say gradius and pretend speed is the only factor). And someone else figures out an even more clever manner to end input. Is that an improvement? I might consider that as a legit improvement. But, we've never really had it be an issue. Because when is this the only significant difference between two movies? In general an improvement is an improvement because it is a bigger technical achievement, regardless of ending choice.

It's an interesting trick, but there's something different on VELHO's run; his score seems to go up faster than yours. Is it because he's more inside the fish? I may be wrong about it, so I'll wait for an answer before voting. edit; Yeah, VELHO's run, even playing in real time, reaches the maximum score a lot faster than this submission, so my vote is No.

There have been a few discussions on this, specifically I remember in the past TMNT and SMB1. But I think the general agreement was that input length should be as short as one can possibly make it, but the ultimate goal be the fastest completion of a title. I just think making the input file length the game completion time doesn't really make sense. Even the idea of ending input early seems to just skew what the actual length of the run is, but I agree that ending input at the earliest point makes sense. I feel both timings should be noted, but the run itself should be based on actual completion time.

Okay, it looks like people are put off by the length of the movie, especially because an RTA exists that reaches 999999 much sooner. I was aiming for ending input as soon as possible, with the secondary goal of max score, like Virtual Pinball. But, if the goal is to reach the max score as soon as possible, it becomes like NES Tetris, where it is important to minimize completion time. I am making a new movie now that reaches 999999 as soon as possible, and it will presumably beat the RTA. However, since that would be a different goal from this submission, I will cancel this one if it turns out that the new one is better.

Very good TAS,yes vote,hope it goes to the vault at the very least!

To address my above post that some people disagreed with, let me clarify: this particular TAS is measured by input length rather than completion time.

Very good TAS,yes vote,hope it goes to the vault at the very least!

That's the one place it won't end up, since it's not an any% or 100% run.