Are most best times cheated?

[kippesoep]

Or are there some severe glitches or paths I'm unaware of? There are a couple of names that come up quite often in these times. This of course could be because they just had some magical perfectly glitched run, but some times are just beyond ridiculous. I've got a feeling a lot of these could be explained by somehow being able to move diagonally.

Some of the early most eyecatching examples:
- Asida 1: 1. 0.083 2. 1.000 3. 1.550 4. 1.550 5. 1.550

From this it seems 1.550 is really possible, though that seems to be some frame glitch. I really can't get below 1.567 and consistently so. 0.083 and 1.000 really seem out of the realm of possibility.

- Asida 6: 1. 1.000 2. 1.000 3. 1.750 4. 1.767 5. 1.767

What's going on here? I've gotten 1.767 myself. I presume 1.750 is some frame-glitch? However, Are the 1.000 times cheated?

- Asida 9: 1. 7.200 2. 11.800 3. 11.800 4. 12.167 5. 12.183
Yea, 7s .. good luck with that.. I'd appreciate any tips to get to 11.8 or 12.16 as well though.

The current weekly challenge has a funky 1st time as well.

There are countless others though. Am I wrong in thinking these times are glitched or even cheated? And if they are cheated, why wouldn't they be banned on all times elsewhere?

I like this game, but it's really ruining it a bit for me.

[kilicool64]

I heard something about a bug that causes the timer to reset after a certain amount of time has passed. Perhaps it never got fixed?

[IHNN]

The game has underwent a LOT of logic updates that have rendered top times no longer attainable. Most notably, the current build has a slightly slower movement speed and pushing a block costs 1 frame now compared to not pushing a block. All of my times are legit (and mostly top 5s) but nearly all of them are no longer possible to reach.

Other changes that affect times:
Random elements are now pseudo-random with a fixed random seed (affects one Asida level)
A bug was fixed where pushing the direction of a force floor pointing against a block would push the block in a clockwise direction (saved around a second on an Eclair level)

Also: back when I played actively I had 1st place on every level of the game except for 3, which had cheated times. The leaderboards themselves are never moderated, even when there's obviously fake times on the boards. Asida 9 in particular, if that's the level that introduces golems, has had that time since the release on Steam. Since Asida 1 was replaced with a different level, I wouldn't be surprised if that time was legitimately obtained, and then the level was lengthened to render the existing time impossible.

[Anonycat]

In A1, 0.08 is really supposed to be 3600.08, but for some reason the game treats minutes:seconds as the primitive unit (notably forgetting anything about the EXISTENCE of hours), and only converts it to seconds when it comes to communicating with the server. This trivially allows any time whatsoever (well, as long as it's less than an hour) to be "possible" on any level, but to avoid rendering the current time snapshots even more worthless, I only demonstrated the proof of concept once, on the relatively meaningless first level.

More ruleset discrepancies:
-In some old versions of the game (before I picked it up, certainly), it was evidently possible to do Chip-style boosts off the end of ice/force floor slides, but no longer.
-Also in old versions, the yellow orb actually sped you up; nowadays it just slows down enemies and blocks, but doesn't have the courtesy of applying the same slowdown to the clock, so any stage that includes that object is obviously going to be hopeless to match the old route's time on
-Red and green squishy enemies (the smoke spewers) now make their first step of rotation at 0 seconds, rather than 3, which can affect the cycles it's possible to catch them at
-In a more recent change, if you push a block onto ice, you cannot follow it directly unless you insert a short wait first--otherwise it'll get to the other end, you will fail to push it since it hasn't completed its sliding animation yet, and you bounce all the way back
-If you attempt to push a block onto an ice path that is exactly one square long before it reaches a wall, you used to (up until the same update that caused the previous change) be able to safely dive out of the way to the side before the block got back. Now you are stuck in the "committed to pushing the block" state even up to the time when the block rebounds to crush you. This renders some custom levels completely impossible under current rules.

In a few custom levels, notably those that try to wall the exit behind a red fish socket rather than a blue one, there are leaderboard times that can't possibly be matched under the current ruleset, despite that being the ruleset under which those levels were introduced in the first place, except by using the time wraparound bug. As far as I can tell, this is a result of players still being able to connect on a copy of the game that's several versions out of date. When they attempt to download the level, it contains elements such as the red fish socket that weren't in the game's vocabulary at the time, and it responds to such cases by not only allowing the level to be playable, but by replacing all of those unknown elements with empty space. This often allows players who steadfastly remain on such old versions to completely trivialize the level in a way that later adopters can't.

[kippesoep]

Thanks for your quite extensive replies.

I think it'd be nice if the times would be kept track of by version or something similar. So, then the best time of a certain version (or just the latest) would give you platinum.

[blendenzo]

I'd also like to thank everyone for the very helpful explanations, as this is something I had also been wondering about.

As someone who frequently makes the top 10, I find it rather frustrating to know that it is literally impossible with the current version of the game to get a platinum metal on any of the standard level pack levels. I have attained multiple platinums on community levels, but that doesn't really mean as much to me.

I can't say there is a real solution to this, though, because it doesn't seem right to penalize players who legitimately got those times in older versions of the game.

[D£€€T]

I just found a bug that can give bogus completion times. Check out the level "work the system" (17395) -- the top score, "Player 284545", is me. That time is extremely bogus, and should be removed. I actually got that time when I finished a different level, "simple help" (17394).

I got that bogus score entirely by accident, due to a bug. I'm on a slow connection, so if I click on one of the user-created levels while Steam is downloading. it takes several seconds to load. If I click on several more levels before one loads, the first level I clicked isn't necessarily the one that gets loaded. Apparently, doing this may also result in assigning scores to the wrong level.

If possible, you may want to check all my other times -- if any look incorrect, they likely are, because I don't bother trying to get the best score, and I often do click multiple levels during loading.

[Allack]

OK here's a question then. There are some very old times in the database what should we do:

- Keep them
- Delete every-ones scores (we'll test that it doesn't wipe your progress)
- Change every-ones scores to something high like 99:99:99

[Allack]

Originally posted by 「£€€T」 Dana:

I just found a bug that can give bogus completion times. Check out the level "work the system" (17395) -- the top score, "Player 284545", is me. That time is extremely bogus, and should be removed. I actually got that time when I finished a different level, "simple help" (17394).

I got that bogus score entirely by accident, due to a bug. I'm on a slow connection, so if I click on one of the user-created levels while Steam is downloading. it takes several seconds to load. If I click on several more levels before one loads, the first level I clicked isn't necessarily the one that gets loaded. Apparently, doing this may also result in assigning scores to the wrong level.

If possible, you may want to check all my other times -- if any look incorrect, they likely are, because I don't bother trying to get the best score, and I often do click multiple levels during loading.

Interesting we'll see if we can reproduce this.

[Anonycat]

Originally posted by Allack:

OK here's a question then. There are some very old times in the database what should we do:

- Keep them
- Delete every-ones scores (we'll test that it doesn't wipe your progress)
- Change every-ones scores to something high like 99:99:99

If you can fix the main problems with the timer, and perhaps have the server verify times via replay rather than blindly trusting a text account of the number of seconds alone, then resetting the charts and having everyone resubmit as they get the chance wouldn't be such a bad idea. The main issues are:

-the hours places is completely truncated, so even if you tried to set a time to 99 minutes, it would only count as 39; and waiting around until the timer hits 1 hour leaves you with a time that counts as zero
-either the first step or last step isn't being counted for its full value, and the amount by which it differs isn't even consistent between plays. If there's a level where you can just hold one button and walk N squares in a straight line to the exit, the time should register as exactly 12 times N frames, every single time. Not 12N minus 1, not 12N minus 2, and certainly not the extremely elusive 12N minus 3. There is no reason why identical inputs on such a cakewalk of a level should ever be credited with different times.

[Allack]

Originally posted by Anonycat:

Originally posted by Allack:

OK here's a question then. There are some very old times in the database what should we do:

- Keep them
- Delete every-ones scores (we'll test that it doesn't wipe your progress)
- Change every-ones scores to something high like 99:99:99

If you can fix the main problems with the timer, and perhaps have the server verify times via replay rather than blindly trusting a text account of the number of seconds alone, then resetting the charts and having everyone resubmit as they get the chance wouldn't be such a bad idea. The main issues are:

-the hours places is completely truncated, so even if you tried to set a time to 99 minutes, it would only count as 39; and waiting around until the timer hits 1 hour leaves you with a time that counts as zero
-either the first step or last step isn't being counted for its full value, and the amount by which it differs isn't even consistent between plays. If there's a level where you can just hold one button and walk N squares in a straight line to the exit, the time should register as exactly 12 times N frames, every single time. Not 12N minus 1, not 12N minus 2, and certainly not the extremely elusive 12N minus 3. There is no reason why identical inputs on such a cakewalk of a level should ever be credited with different times.

Bugs are never ending, we’ll look at these and then there will be two more tomorrow. So back to the question:

- Keep them
- Delete every-ones scores (we'll test that it doesn't wipe your progress)
- Change every-ones scores to something high like 99:99:99

[kilicool64]

Originally posted by Allack:

Originally posted by Anonycat:

If you can fix the main problems with the timer, and perhaps have the server verify times via replay rather than blindly trusting a text account of the number of seconds alone, then resetting the charts and having everyone resubmit as they get the chance wouldn't be such a bad idea. The main issues are:

-the hours places is completely truncated, so even if you tried to set a time to 99 minutes, it would only count as 39; and waiting around until the timer hits 1 hour leaves you with a time that counts as zero
-either the first step or last step isn't being counted for its full value, and the amount by which it differs isn't even consistent between plays. If there's a level where you can just hold one button and walk N squares in a straight line to the exit, the time should register as exactly 12 times N frames, every single time. Not 12N minus 1, not 12N minus 2, and certainly not the extremely elusive 12N minus 3. There is no reason why identical inputs on such a cakewalk of a level should ever be credited with different times.

Bugs are never ending, we’ll look at these and then there will be two more tomorrow. So back to the question:

- Keep them
- Delete every-ones scores (we'll test that it doesn't wipe your progress)
- Change every-ones scores to something high like 99:99:99

I'd say delete them.

[Anonycat]

You may have glossed over it in the post, but I said I'd be for a wipe as well.

[kippesoep]

Originally posted by Allack:

OK here's a question then. There are some very old times in the database what should we do:

- Keep them
- Delete every-ones scores (we'll test that it doesn't wipe your progress)
- Change every-ones scores to something high like 99:99:99

Thanks for taking the time to look into this.

If these are the only option's my vote would go to deletion of them. I'm not sure what setting them to a max value would accomplish. If everything is on the table, I'd opt for a ranking based on version of the game, though I suppose the sent version number could be hacked in some way(?)