The start of the timing should be before the first cutscene because of potential abuse. Right now i'd say #2, the first black frame after the splash screen. Start of the ingame timer might be also a good point for an autosplitter (if that ever happens).

yes i do. read the wrong line BibleThump

Note that the ingame timer continues to run during loading screens. #2 could be a loading screen. I'm going to do some timings and report back. If it's not a loading screen #2 and #3 would be pretty much the same, otherwise #3 is the better choice.

ok done my timing

Note that the youtube video might have slight deviations since my timings are based on my dxtory recording. Frames could have been eaten by my compression before uploading or youtube transcodes.

The first pass is done from a freshly launched game, since there is absolutly nothing in the memory, the loadings are longer than 2 and 3.

The time between first black Frame after splash screen and the cousins bellic is 3.6 sec, 3.267 sec and 3.2 sec. In your case it's 2.09 sec.

Yes, the start needs a new rule. No, it shouldn't be right after you press new game.

The first loading screen is extremly inconsistent it can range from 2-3 sec to over 30 sec. Your Video has relativly fast loading times but i had attempts with 5+ different loading pictures sliding through that initial load.

I remember you resetting your game and even stream a couple of times because your first loading screen became inconsistent, so you should know about that.

I'd say timer start on the frame "the cousins bellic" is visible in the bottom right corner. That's when the loading ends.

alt tab to do things = evil

Or at least tool assisted. Since you used a function of a third party software for that. Might not work on different operating system.

There was also another thing that gets skipped by tabbing out in the jp version, but i forgot what it was.

Since we fixed VC timing start to a proper point we should do that with sa too. Even if there is no abuse yet.

So there are 2 reasons to enforce rules in speedruns.

1. To make a run more entertaining.

That is basically the idea of classic%. Remove the more boring stuff like warps all over the that can be consistently repeated. But i also see why menuing makes a run less entertaining.

GTA is mostly about driving. If you allow to change traffic density mid run, your driving skill over the full run becomes less relevant, but your menuing skill is more important.

You say that the speed difference is marginal to non existent. That might be true. Still with this rule change future classic runs would be on 1% for the most part of the run. Because warps are not allowed and you drive for the most part.

An again driving on empty streets is less entertaining than driving on somewhat full streets.

I know you will say a speedrun doesn't need to be entertaining for the viewers and that it's about the fastest times, but the classic rule set was made with entertainment in mind for the exchange of speed.

(1.5 i would flip less on traffic 1%. Less jolzi flips = worse Kappa )

2. Make runs and runners more comparable to each other.

Speedrunning is sortoff a competition with the leaderboards. If you can't compare runs properly then what's the point of leaderboards?

If traffic is set to 1. You can't really say person x is better at driving, because he doesn't hit cars that didn't spawn. The most time spend in a GTA run is driving after all. But on the other hand menuing skill becomes a thing.

And since you mentioned SA. It's also about Hardware dependencies. While the frame rate fucks up old GTA way more than in GTA4, it's still important in a run. If you have a consistent FPS the inputlag is lower and tricks are more consistent. Some missions depend on your framerate. Loading times are also affected by FPS.

The traffic density doesn't affect the FPS too much, but you can notice it in crowded places like south Algonquin or center Alderney.

Not only that. For me it feels like pausing the game often also causes worse FPS. Another thing that would come from menuing.

Having a set rule for this would reduce those limitations and make the runs on the leaderboards more comparable.

I'd say 33% traffic density should stay part of the classic rules. With my 2 arguments + the fact that you yourself say that it wouldn't make the run faster if it was removed.

I would be ok with a different value than 33% aslong as everyone is using that and it can't be changed mid run. But 33% is very consistent in terms of everything, that i don't see a better permanent density setting.

Also the traffic density discussion started because of GTA5. I'd say classic rules for the PC version of GTA5 shouldn't be set before the game is even out...

the reason people were blowing up in speedruns was because of the old liberator strats without replays. Your cartel cruiser would be almost burning by the time you get to kenji (with old speedrun strats)

Or maybe do waka gashira the intended way? The main reason for using the rockets is that your car is very likely to get blown up, but has anyone tried the same replay strats as in liberator to stop them shooting you?

And then leave the car par with the usj ramp . Since the cartel cruiser is very fast this might save time over a sports car on the way back to donald.

As for the gta 4 glitch showcase my own video finished uploading faster than i expected. So replace it with the one in my submission.

A good layout is where the game touches at least 3 edges of the video.

Then it's still a bit different. With the dupe it's "get the 0,53% for finishing quarry" vs no dupe "beat every single one of the quarry missions". So yes it is about getting the % higher and no so much about beating everything. Also think about gta vc. 100% and all missions. Where duping is allowed, but not to skip any mission or asset with it.

Well i don't run 100% so i don't care enough. People who actually run it should decide.

Splits are more meant for your viewers to show how well you are doing in the current run. If you don't want to give your viewers information prepare for more "is this wr pace" or "what's your pb" questions in chat.

If you submit a PB to the leaderboard you should always time your recording since manual splitting is inacurate and timers can glitch out. Unless there is an autosplitter for your game or a different timing method that uses demos.

I have two issues that come to my mind with the quarry dupe in 100%.

1. This would make it a "get the stats menu to show 100%" run and not "complete all game defined objectives that lead to 100% completion". I don't want to start the same discussion that sda had on eidgods run, since we had that already. Just wanted to mention that it would be a different kind of 100%.

2. Can the game be finished after the dupe?

Quarry 1 spawns 7 barrels + 1 bulldozer (and maybe more). Duping it 7 times would spawn at least 56 objects on the map and into memory. We learned from duping in other gtas, that after some point you don't get new markers on the minimap. I see this as a big problem during paramedic (yeah i know the route but maybe El Quebrados paramedic will be a thing Kappa ), firefighter, vigilante and gang wars. Think about losing a lot of time when doing gang wars, but can't find some dudes because they are not shown on the minimap.

Missions with kinda free moving AI that has to be killed would be a problem without markers too. Like House Party, Catalyst or Los Sepulcros. T-Bone Mendez and Intensive Care would be a real pain in the ass since they move over the whole town. I imagine hitting the trigger in the air during Freefall and Saint Mark's Bistro would be very difficult when you don't know which direction to go.

Another downside we know from duping the same mission over and over are ghost towns. This can be a problem for car collecting and again gang wars. Not being able to start a gang war because of ghost town could be a bad thing.

Since strimming is bad and getting secret WRs is the new big thing right now, some might wonder what is the best lossless codec. I compared some lossless vfw codecs which can be used with most recording software like Dxtory or MSI Afterburner.

https://docs.google.com/spreadsheets/d/18cszqLt88rtc-sPq9_I_4obbj_RpGYcPPB-6PWx5MSw/

You could get the oyster on boat school.

Because everyone says stuff feels faster than other stuff but nobody timed it.

https://docs.google.com/spreadsheets/d/1LwERc2JPm9Lj1O-5sXDPR4rrOkBSpZxmY7l1WZvbE/

notes:

• I downloaded a fresh vc installation from steam before timing

• I used the slowest HDD i have available right now (Samsung f2, 5400 rpm, 1.5tb if you care)

• The HDD was 95% full because full HDDs run slower

• Game was lossless recorded to the same HDD that i played from the whole time, since lossless recording is one of the biggest strains on an HDD

That way i tried to simulate the worst possible scenario for somebody who has only one old and full hdd, is not streaming and has a too slow computer to live encode for smaller file sizes.

Just as a reminder, in case you really want a glitchless run. There are 3 glitches in the muscle memory of every runner: Infinite sprint, keep the door open when entering car and reload by switching weapons. Have fun to get it out of your muscle memory.

You can't leave a fullscreen without pausing the game. You need to play it in windowed mode if you want to leave the game and chat.

Most modern games have an Option called borderless window or fullscreen windowed. Which looks like fullscreen but behaves like windowed mode. That way you can also leave the game and chat.

If the game doesn't have that Option, you can still run it windowed and use a tool called shift window to make it fullscreen windowed.

But most people who chat while playing usually use a 2nd pc or Laptop.

Also another argument for not adding los santos to the leaderboards: some guy would need to verify the runs BrainSlug

S whens 20% race?