Category system is a lot better now, but could we please clean up the category names? Maybe something like this?

Also category rules were never updated after this change. Should probably mention which players setting method allowed in each category to make it more clear for new runners.

Thank you so much, BokserKabaty! I made these changes:

Normal P1 SC Assassin Added/removed the runs you mentioned. I couldn't find any videos for first sub 1:30 Teo assassin, but I did find this: https://www.reddit.com/r/speedrun/comments/2v9aok/wr_diablo_2_lod_any_12543_by_teo/ and I'm referencing that in the history. It also mentions the Nightfall run before it, but I feel like that might have been a HC run. Sadly I think video of the Nightfall run is also missing. Does anyone know the date of Nightfall 1:34:39?

Normal P1 HC Assassin Added missing runs, fixed the dates. Removed Bender's 1:17:21 run, because Meow had the WR.

Normal P1 HC Druid Added 29 Feb 2020 1:26:25 by 327.

Normal P1 SC Paladin Added the missing runs you mentioned.

Normal P1 HC Paladin Added 6 May 2019 1:44:20 by Bender.

Normal P1 SC Amazon, Normal P1 HC Amazon Added all the Bender runs.

I think we have a lot of missing data for P1 Druid, both HC and SC. If someone is a good speed historian then let me know and let's fill it with data.

Anyone got a good idea how to handle early SC/HC separation? Early on SC/HC runs competed with each other, but putting together the history now and filling the Nightfall HC runs into SC seems odd. Yet when I remove them from the history then suddenly it seems like there's a lot less competition and overall makes less sense.

Edit: Nevermind, found the Nightfall 1:34:39 run and it's from Jan 2 2016

Thanks, it's another category with broken data, because of the Teo incident. There's a lot of them. I think all of normal categories. I thought only p1 initially, but I guess it's all of them.

The runs with broken data have their "date played" set to 5 Apr 2020 which makes them kind of easy to spot. It is difficult to figure out the correct dates for them though. Any idea when you ran this? I added the run to the history document already though.

I've been just keeping "5 Apr 2020" as the date if I'm unable to figure out what the actual date was. It's a date placeholder.

Our WR history through the statistics page is useless so couple months ago I started a project and started writing down the history. Some automatically, a lot manually.

There are categories where data is incorrect from the Teo incident: https://www.speedrun.com/d2lod/thread/mankr I'm not exactly sure which categories are impacted by this.

It would be nice if we could come together as community to fix our history. Let me know if you want to help contribute to the history document. I'll use the data from the document to put together fancier history page on diablo.run as well.

Report any mistakes you see in this thread or message me on Discord.

D2 WR History: https://docs.google.com/document/d/12jo9KozhOo7nURb21DqJsxOyw3a6pg7Huus3qY94uvw/edit?usp=sharing

The biggest issue I see is that if normal sorc can abuse 6 minutes of timesave with IGT then it's going to spread easily into longer categories. Especially hell. There will be some dumb IGT abuse eventually in hell runs.

The example 327 pointed out is a good one: 1-02-45 RTA->1-02-00 IGT 1-02-39 RTA -> 56-35 IGT

In hell runs it's going to be even more dramatic. Instead of improvising a leveling route you must use leveling that's optimal for IGT or you're out of competition.

Fishing for the 5 Jail setups is 1 in 50. Fishing for the best IGT setup is 1 in 250. Anyway the entire point of converting to IGT at the time was to make it more fair, but the more runs we see in P1 sorc category the more I realize that it does the opposite. RTA much more fair compared to the IGT category. IGT failed to do what it was meant to do.

I know TCP/IP loading isn't included in IGT for map reroll abuse, but when it comes to actual timesave then it doesn't make as big of a difference (for p1 sorc at least). On these two example runs if TCP/IP loading would pause the timer then it would've only saved 30 seconds.

I don't mind the IGT category, but RTA should be the default for D2. Also IGT pausing during TCP/IP load would make the category much more accessible. IGT at its current state doesn't solve anything.

I'm not expecting anything to change, just wanted to get it out of me.

Every time I run P1 sorc I get reminded of the broken timing method system. Recently I did some runs and that feeling is still there.

IGT favors one particular setup way too much. "The Best" Jail setup gets even better in IGT. Here are two runs:

Both runs get the leveling setup at level 5 (9:30 RTA). • First run gets level 18 at 31:45 (RTA) 27:48 (IGT) • "The Best" setup level 18 at 31:45 (RTA) 26:00 (IGT)

57:33 (IGT) is faster than 56:35 (IGT) run, but the time doesn't reflect it well. I just got an extra 2 minutes of time save in the Jail for doing nothing.

"The Best" setup gets almost a 2 minute head start in IGT and there's nothing other setups can do to compete with that in IGT. RTA ends up much more balanced because the best setup gets longer load times which makes it more fair compared to other setups.

Basically in RTA all 5 setups are competitive, but in IGT only one works.

I put together a demo run to find oddities. One thing I noticed it includes that number 5 font fix so there may be some more hidden changes. I also got a game crash during leveling in Jail when I smashed enter and mouse click fast during save and exit lag.

Overall it ran well and was a very enjoyable experience. Much better than downscaling my 4k monitor to 1080p to run the old glide. Much more comfortable on the eyes and the game doesn't have weird scaling glitches or mouse bugs.

Settings I used: https://pastebin.com/raw/tGJmUEQC Demo run:

I've been using these settings currently: https://pastebin.com/raw/tGJmUEQC

The high FPS can be disabled as well with nofpsfix=true. I think it doesn't change gameplay. Gameplay always locked to the 25fps, but it gives you the smooth high fps mouse like in Bnet which is nice.

Best part for me is that this enables me to run the game on my 4k monitor, I have to go through annoying hoops otherwise to get the game to run. Another benefit is that I haven't encountered any black screen issues like there are with old glide through game capture in OBS.

It is open source, we can definitely request more settings if something needs to be disabled for d2 speedrunning.

https://github.com/bolrog/d2dx There's an updated glide wrapper. It has some additional features that I can see are not allowed like the wide screen resolution, but they can be configured easily from the config file.

Anyway my question is if this is allowed for speedrun.com d2 submissions? One of the main reasons I'd like to use it is that they fixed the mouse bug that original wrapper has where it jumps up each time you open a menu with a scaled window.

D2DX allows you to scale the window a little bigger and not get that mouse bug.

I support 17 april, I must leave tomorrow, surprise vaccination tomorrow

I'm in, what are the rules? Is the winner based on RTA or IGT?

Current WR in HC P1 Sorc is 1:00:41 (IGT). Yesterday I got a 1:02:00 (IGT). The run in IGT is 1 minute slower, but when comparing RTA the run is almost over 3 minutes faster.

1:00:41 (IGT) - 1:05:12 (RTA) 4min 41s difference 1:02:00 (IGT) - 1:02:45 (RTA) 45s difference

The run yesterday got a leveling setup where I didn't need to do as many save and exits as usual so my IGT/RTA time difference ended up very small. This is the reason I couldn't make act 2 leveling work anymore as well, because IGT hugely favors strategies that save & exit more.

The currently available diablo.run extension was not made by us and it's closed source. It stopped working as well some time ago and we built our own extension, but it's stuck at Twitch review process. Can take up to 2 weeks. It's been a week since we submitted it so should be available soon.

D2ID stopped getting updates couple years ago and no longer works. We built diablo.run as an alternative. https://diablo.run/setup

Could someone refer me to a runner who is known to have bad load times? I looked at top sorc runs and didn't see anything out of the ordinary there when it came to load times.

@BokserKabaty I think you also didn't have -seed active. Your spawns seem inconsistent. Better to keep -seed for now so we have some consistency.

Edit: Actually it's probably best to test both anyway. I'll do some runs without -seed as well and see if I get similar results.

I did some runs without -seed and also got ~99 second difference in RTA/IGT. I'll post some videos and start a Google sheet to keep track of all submissions.

I started a topic for such an experiment https://www.speedrun.com/d2lod/thread/ci3bt

I also tested out the time difference between RTA, IGT and -ns. IGT is op compared to -ns. I'm kind of surprised how little time -ns saves. -ns in p1 normal sorc run saves maybe like ~1 minute. IGT saves ~5 minutes.

If we converted to IGT so we can have a fair competition with the -ns run then that definitely didn't work out. If there could be a -ns penalty and the run retimed into IGT it would still be WR. Edit: or maybe not, I forgot my run exists.

It has created an odd situation now. I don't want to scam 327 of the WR as well with the rules that were put in place. No good solution anymore.

Trying to figure out how inconsistent the load times really are between systems. Currently we assume they're inconsistent, but it would be nice to actually test and prove it.

Take part of the experiment:

1. Download the save file. Ask for the save file from my Discord if it's not working. Indrek#9137 https://www.dropbox.com/s/asqjw5mx92c6wfc/Helene.d2s?dl=0
2. Add -seed1400148573 to your target line to keep Jail spawn consistent.

Start the timer. Use Telekinesis on waypoint, walk towards the champions, cast one nova, save & exit. Repeat until level 16 and stop the timer.

My submission: My first slowest attempt was RTA: 5:53 / IGT: 4:20 Time difference: ~93 seconds

After this I did a lot more attempts and time difference was always the same ~93 seconds. Here's my fastest attempt: RTA: 5:42 / IGT: 4:07 Time difference ~93 seconds

Bonus experiment: -ns comparison RTA: 5:13. While IGT saved 90 seconds, then using -ns only saved 30 seconds.

Probably won't be the last experiment. For example should also test shrine checking load time consistency.

@Bender Any% normal would be for p1/px/p8 sorc? I guess even in these sorc categories it's just p1 sorc where the timing breaks down due to load times.

Actually one question I have is that I've heard people say "load times are inconsistent", but I haven't had that experience myself. They seem very consistent to me based on the way you play. Even TCP/IP load times are consistent and you can work around it. It's the same in-game where certain actions just make your load times longer and you can strategize it. That's what Slimo did with -ns and it's what we lost when we swapped to IGT. We lost a big portion of strategizing the sorc route which is what made it interesting in the first place.

For example I'm confident that p1 sorc can be faster with act 2 leveling with RTA (it's something I was working on), but when we swapped to IGT the Jail setup was not beatable with anything in Act 2. We made rules based on one specific strategy and inadvertently made it super optimized. This also carries over to other classes. If IGT is maintained then all classes best route in p1 is Jail leveling (I'm generalizing, but only tested assassin and necromancer), where otherwise in RTA you would be behind too much.

@Wafu It's not really ignored, it's just something we have no control over. We could wait another year for speedrun.com to provide these features, but we've waited long enough and the community is mostly fed up with the current situation.

Also when we voted on these issues I was under the impression that this is how it was going to be setup anyway. I'm sure there were others who voted on this premise, but later turned out that it's not possible and was done differently. Well, I would have voted differently and so would have other people. That's a big reason we saw quite a few runners give up and leave.

Retiming promise was another one that probably made people favor IGT, but was later decided that runs won't get retimed.

Timing method setup and retime promise were the only reasons I favored IGT to begin with, but none of that happened.