Just tested it and Livesplit running faster than IGT is probably normal; my game runs about 59.5 seconds IGT per RTA minute (so 4:00 IGT corresponds to 4:02 RTA). My games UPS was stable at 30 for this test, did not drop to 29 at all; I suspect the UPS is rounded and the actual UPS was stable around 29.75. This would also explain why you see different numbers between devices.

CHLEFF's Ruins Repurposed run has the same effect and around the same numbers as me (5.50 IGT ~ 7.71 RTA -> 36.70 IGT ~ 39.12 RTA) would correspond to about 0.4 second difference per minute of gameplay.

Are your numbers relatively consistent with the above?

I have been too busy to get practiced runs submitted unfortunately, but I have been paying attention to what is happening here. If you want help clearing up the categories I can get ok runs on several levels, also to give players a base strategy for most of the levels.

To my best knowledge Archon can be held to 0.00 creeper for about a minute without anything too crazy, and indefinitely with insane micro (achievable with pause). I can get a video submitted to the leaderboards both to confirm getting 0.00 creeper is possible and to give players an RTA viable method of achieving it.

I think I should also ask a question about starting when closing the ADA log; what is the policy on moving the camera while the ADA log is up? Right now there is no rule against moving the camera so it is in an ideal position for the level, but it feels to me like performing an in-level action so I thought I would at least bring it up for discussion.

As far as categories are concerned; we could remove Mverse to reduce category bloat (I don't think there are any runs for it), and bring it back if we get future requests for it. Thoughts?

[quote]Likewise, I would not have 110% on Archon and any other map where incoming Creeper cannot be stopped[/quote]

Did some testing, Archon 110% is possible (Creeper graph reads 0.00) but it is not easy to perform. I set up disarmed shields and armed them all at once to stop the orbs for ~5 seconds to get it. !() edit: copy paste url, I don't understand speedrun.com image embedding.

[quote]especially with @Fireswamp finding new cheese strats[/quote]

I'm not really finding new strats as much as getting RTA runs for known IGT strats; if you are curious you can filter for the 'specialevent' group in the play logs to see some of the submitted times we have. The only thing is the Founders 'End the Beginning' objective was patched and is no longer possible in 1:06.2.

There are 2 more runs I have cheesy strategies fully planned for, one I was just being dumb, found a consistent method and will try 4x later, trying the other one now; those I should get submitted today/tomorrow but irl stuff is keeping me busy a lot.

edit: both runs are submitted.

[quote]IGT's subcategory would be pausing. Personally, I'm leaning slightly towards RTA, but I'm fine with either timing method as default.[/quote]

I agree for RTA as the timing method; no pausing is not a great option because the story missions have several autopauses.

[quote]some levels are loooooooong without time warp[/quote]

Longest IL category should be Founders 110%, I just tested it and got <32 minutes, would have been close to 30 minutes if I didn't have trouble tracking down some creeper on random tiles. Good runs for all IL no time warp categories should be sub 30 minutes I think.

Also I made a mistake on something earlier, the .rec files are actually .cw4rec files, sorry about that.

[quote]Copying every input seems like it would be ridiculously hard - has anyone tried this?[/quote] I have done this with my own runs before, the categories that are easy to recreate the top any% runs that I know of are Ruins Repurposed and Wallis, with 09 Leo, Farsite, Sequence and Serious not being too difficult. A proper recreation for most other levels I would expect to be relatively difficult and take a few hours for the shorter levels (ex. Not My Mars, 100% Ruins). Longer levels (ex. Founders, Tower of Darkness) I would classify as extremely difficult and time consuming to do. With how small this community is I think only Wallis any% and Ruins Repurposed any% would be something that we should consider.

edit: I should probably be more clear about difficulty to recreate; these are my opinions on time to recreate. If you know what you are doing you can recreate Wallis or Ruins any% in about 5 minutes given a video. I would say 09Leo < Farsite < Serious < Sequence < 1 hour of time. Most other runs would take over an hour to recreate, but I think even 100% Founders could be fully recreated in a day by someone dedicated and knowledgeable at the game. I have not tested fully recreating a run that long and extensive though.

[quote]Do .rec files hang around on a local PC after reset?[/quote] They do, they are just saved as files on computer; but there are some things I am not sure about in regards to the file being generated. To my best knowledge, it generates the file for any level where you Rift Jump but there is an option (disabled by default) to limit the number of recordings to a certain number; I suspect this deletes old recordings, but I do not know.

The main problem is that this file won't find splicing using saves or replaying runs back faster to improve RTA times. So it probably won't be as effective against main cheating methods, only against map/game modification, which I suspect won't be very common. Thinking about it I am not sure I still think the rec files would be effective enough to justify having people need to worry about them.

[quote]So if I'm understanding this correctly, that means physics and IGT are tied, right? In this case, when set to 4x speed, the physics engine is only running at 2.5x speed and the IGT would also only be 2.5x speed?[/quote] Yes, you are understanding correctly.

Does the game running slow/lagging impact IGT? I don't have time to check right now but if the timer is tied to performance, that would be an easy solution.

Game physics and IGT should completely separate to lag, but there have been lag based bugs before. To my best understanding the mechanic is that for each frame, the game calculates the next frame and increments the IGT by 1/30 of a second. If this calculation takes less than 1/30,1/60, or 1/120 of a second depending on the speed set by the player, the game will wait until that time has passed. What this means is that if your computer takes 1/75 of a second to process each frame, the game will run at full speed if set to 1x or 2x, but will run at 2.5x speed if you set the game to 4x speed. You need the game to process all frames in less than 1/120 of a second to run at 4x speed. The bottom left menu will show your UPS (Updates Per Second) count in tiny text.

do you think there are any chances for a dedicated speedrunning channel/category?

We used to use the competition channel for any lengthy speedrun/tas discussion, but with the amount of stuff we want to discuss it is easier to jump into voice chats than write massive posts for us. The competition chat is still there and is something we should be able to use for speedrunning chatter.

Also for the .rec files, they are automatically generated for all completed levels, hit Recordings in the main menu to link you to the file location. Giving it more thought they probably won't be that good at catching cheating though, so probably still not worth it.

I have gathered thoughts from myself, Cornucanis and Hypnotic22; for context we are the main 3 TASers/Speedrunners active on the Knuckle Cracker Discord. As we usually do, I think we went a bit overboard, but figured we should just list all the things we are thinking about.

Firstly two issues in regards to clarity in the rules: The rules do not ban the use of the mission editor. What about saves? Obviously no loading saves not made during the run, but could I made a save during the run, then load that save if a trick doesn’t work first try? This becomes a very major point for in game time runs, as save loading difficult sections is one of the most powerful tools we have, but opens up the potential for save file manipulation based cheating and can make runs exceedingly long recordings as large portions of the run could be failed attempts that don’t contribute to IGT.

Also some minor cleanup on the categories: IGT runs are using the RTA time when you view all level leaderboards. For the first 3 levels, 100% and any% are the exact same. Most levels do not have easter eggs, this will be several duplicate categories. How are you planning to verify no creeper on the map for 110%? The bottom left menu will work, but for something like Archon it might be difficult to verify that there is no creeper because of the rain. I would be sure to test Archon beforehand to see if any issues arise and to get a verification method clearly defined in the rules. The game by default tracks runs not by any% or 100% completion times, but by times for each of the individual objectives. Those are potential categories, but there is some consideration for how many categories a game should have.

In regard to preventing cheating: The information in the bottom left UI is probably the most powerful tool for preventing cheating and it is not something that can be manipulated easily. The bottom right is also useful, but significantly less so. It is very easy to clone a map into the mission creator, playing that mission and using the restart mission would make it undetectable that someone is not loading the story map. The easy ways to prevent these are to require bottom left to be visible and ideally legible, but that would add a minimum spec of ‘can record at 720p’. Not insignificant, but becomes a much bigger problem due to another concern that will be mentioned later. The cloned mission can be fixed by having a requirement that recordings start or end on the mission select screen or show the play log. Lastly, consider having players upload the .rec files associated with their runs, it is a file type that we do not know how to manipulate, and faking it well is very difficult. It gives a top down pixelated view of the entire map and contains details that could be matched against a video submission.

How do you plan on dealing with different versions or builds? There is no way that I know of to go back to previous builds, so if a record is set on one build, but an update makes it impossible to match that record, then what? There are 2 campaign levels that have legitimate completion times that have been shown to be impossible to match in the current build. In fact, there is currently a bug that makes one of the campaign levels a lot easier and faster to beat; though no times have yet been submitted that beat the record set without the bug. Minor aside to the above; demo and full game are different builds, I do not know the differences in detail though.

Probably the biggest issues that came up was how to keep the competition fair. One was in regard to IGT runs, the other was in regards to RTA runs. For IGT runs, is there a way to prevent copying runs? From the recordings you can easily copy every input made by a player. The deterministic nature of the game means that if the same inputs are made, the same time will be achieved excluding some RNG elements. These RNG elements are rarely significant though and I believe the campaign levels have been set to give the same RNG seed to everyone. So, with an hour or so free (depending on the length of the run), there is nothing stopping someone from going in and copying the 1st place run and matching their time. This might not be something that can be solved while maintaining high standards of verification, but if anyone has a solution that would be awesome.

For RTA runs, if a level gets well optimised, the leaderboads can easily become competitions based on hardware. Obviously different levels will lag to different degrees, but the game lags even on simpler levels if the speed is set to 4x and your hardware is not up to par. On more complex levels, even strong computers can lag, reducing speed significantly; this lag is made even worse if you are recording using software instead of a capture card. For an example of this already on the boards, you can notice in Cornucanis’ Farsite run that the game was just running slower. Farsite is not a map that lags heavily as well, because of this there are a lot of levels that could become impossible for even medium hardware to compete. And this is before we even mention MVerse; to run 4 player MVerse at 4x speed without slowdown requires good connections of 5MB up/down with 60 ping max between players on top of the hardware limitations. One idea that we did think about was the idea of a No Time Warp category, where speed cannot be changed from 1x. While this won’t completely remove the factor of lag, it will make it so that it affects only the lowest end hardware.

Sorry for the mountain of text, if you have any questions or want to discuss stuff, the 3 of us are active on Discord and I will be paying attention to this forum thread.