The Next Iteration of DSDA

[Zahid]

Very Brilliant effort...i got a suggestions,don't know if its possible though but some kind of link between id games wads pages to their demos page at dsda would be better imo...for example if someone is visiting id games page of Hell Revealed there should be a link to dsda page of all Hell Revealed demos...

[vdgg]

• The IWAD list features WADs that require IWADs to play, such as ht.wad needing tnt.wad. This list is very incomplete. I understand assigning a PWAD to an IWAD is done manually. Would you like me to point out a few WADs that I know that require plutonia or tnt and are not on the list?
• Caverns of Darkness (cc-cod) is odd in a sense that PrBoom+ demos are "incopatible demos" IMO. ZDoom and GZdoom can read cod-zd.pk3 patch which adds special monsters. Thus, marking two @Zahid's demos as "records" is very questionable.
• Apostrophes in notes do not seem to appear, at least when using Opera. In urania, MAP03, a demo by Spectre01 has the following note "UV Max attempt. Nine monsters unkillable on &apos-complevel 4&apos."

[GoneAway]

Yes, if you know some pwads that are in the wrong lists, that would be great!

 

That's an interesting case for cc-cod, I'll have to fix that. It's a dangerous road to go down too far (something like eviltech for instance that is broken in a lot of places in prboom+ but ostensibly compatible), but it seems clearer for cases like this.

 

Looks like that comment is broken for me as well. Most likely it was an error from parsing the original comments, and we can do a bulk find / replace and have it working.

 

Thanks for the notes.

[vdgg]

Plutonia: lethe, philplu, plut1, satangel, sp1

Tnt: 32hnpain, agony2, itwas, napalm, tnt-ren, waterw

Edited July 15, 2019 by vdgg

[GoneAway]

Thanks! I'll put this on the Todo list.

[Grazza]

I'm not convinced it's a good idea to include the millisecond times for demos as a matter of course. This might encourage people to start uploading "improvements" over demos that were never recorded with millisecond improvements in mind, and for which the author in many cases could have made millisecond improvements if they had seen any reason to. That is, ocelot might have been happy to grind a run down to 0:23.97 and seen no need to spend further time trying for a 0:23.94 or whatever, even when he certainly could have done so, as he saw no hope of taking off a full second.

 

Personally, I would tend to upload the better-looking demo rather than the one with the best decimals. Since, you know, they are the same time really.

 

It might make more sense only to include millisecond times specifically for categories for which millisecond improvements are generally accepted. (Whichever they are - is it just short nomos? I'm actually unsure.)

[GoneAway]

Yes, it's not decided yet if we'll hide tics for some categories or not. We've discussed this at length on the discord but some notes:

 

1) Some maxes are very short. So filtering by category directly is probably not ideal. Doing so by time would make more sense.

2) Some runs that are "long" are actually short in terms of skill-based content. I.e., runs with lockins and such. This means filtering by time blindly isn't completely accurate either.

3) A lot of runs end up being rather poor because there's no incentive to get a faster or better looking run if the next second seems impossible. Of course, the difference between 9.97 and 9.00 is massive. But this is still pretty true at 30.97 versus 30.00. It's not obvious to me where a line would be drawn.

4) Allowing tic improvements supports gradual drilling down of a run that is basically abandoned in the community right now. A lot of people look at a run and just give up because they don't think it can be improved. But if you see someone pushing things down closer and closer, the initiative to make it to the next second becomes stronger. So, I think that sorting in this way will actually make full second improvements more likely.

5) Not counting tics has made doom runners lazy in many contexts. Just rush to get the easy second on a run and submit it instead of getting a great run.

6) Faster is faster. If we're sorting by time we may as well sort by time. We could sort second ties by date, but I think this is arbitrary anyway.

7) Competition is dead on a ton of runs that could be improved, and noticeably. This is one of my main motivations. This is kind of a different way to state the other points but it's the central topic in my mind. I want to see more and better runs. Will this result in a bunch of runs that are barely better? Probably it will, but I think we will also see a lot of demos that are much better. That's a favorable exchange to me.

 

Some of the main objections I've seen:

1) As you mentioned, it's a bit unfair to compete with people that didn't care about tics or who potentially didn't even have access to that timing at all. This is a fair point, and it's why I have the note for the first person to reach a time. In many cases it isn't super relevant but I wanted to acknowledge the runs in some way, for the reason you mentioned. However, I think the community evolves constantly and this is just one thing that causes competition with the past to "break down" in some way. Another obvious example is advanced HUD used in max. There will never be truly even competition with the past so I think it isn't a rule that such things should be maintained.

2) There might be "trolls" that could beat a run by a lot but instead beat it by one tic. This is something I'm inclined to wait and see on, because I hope no one is so inclined. But, people can already do this in other ways, like waiting at the exit until the last second. So I don't think it's a new issue, and it doesn't seem prevalent now, so I'm inclined not to expect it.

3) Isn't 2:59.97 vs 2:59.63 probably an unnoticeable difference? I think people are capable of realizing this themselves and won't read into the sorting much in such cases anyway. If it's obvious to us it would be obvious to a viewer as well, so I don't think anyone is hurt by this (and as I mentioned above, the time on a run could be deceptive). Either one might look much better than the other (luck vs skill) but we can't know that objectively so we can't sort by that.

 

As I said, where and how to display them isn't finalized, but these are my feelings on the matter at this time.

[Da Werecat]

3 hours ago, kraflab said:

There might be "trolls" that could beat a run by a lot but instead beat it by one tic.

 

Unless we're talking about TAS, this would require so much skill and precision that it would be a worthy accomplishment in its own right.

 

[almostmatt1]

Hi @kraflab, just wanted to bring something to your attention. I love the table view function, it makes looking for some specific records much more user-friendly, but what it displays as the fastest time is a bit inconsistent. For example, the E1M1 UV-Max table view shows the fastest time as Adam Hegyi's 0:26 TAS (fastest 'human' time is shock's 0:28), but the E1M1 UV-Speed table view shows the fastest time as being shock's 8.97, despite there being a 0:07 TAS run. 

I know that in the settings section you can turn on or off TAS demos being shown, but it'd make sense that the default fastest time showed in the table view is either always TAS demos or always 'human' demos. Presumably, the vast majority of the time people would be looking for the fastest human demos.

I imagine this is an unintentional inconsistency so I just thought I'd mention it. :)

 

EDIT: Upon further inspection it looks like the actual problem is that Adam Hegyi's 0:26 TAS simply hasn't been labelled as a TAS on dsdarchive. Looks like it IS labelled as a TAS on doomedsda, so there was probably just some issue with transferring the demos over. My mistake. 

[GoneAway]

Thanks for the report! We have a long list of these kinds of missed notes and such, so we'll get to them soon hopefully 😅

[Bob9001]

I really like how the site has developed but just one little thing as a suggestion, like how you have the chess piece for the record. You could do something to let people know that a demo is solo-net? I see the 2P which is nice!

[GoneAway]

It's an error that solo-net is missing as a label. This will be fixed soon I hope.

[j4rio]

On 7/17/2019 at 3:20 AM, Grazza said:

I'm not convinced it's a good idea to include the millisecond times for demos as a matter of course. This might encourage people to start uploading "improvements" over demos that were never recorded with millisecond improvements in mind, and for which the author in many cases could have made millisecond improvements if they had seen any reason to. That is, ocelot might have been happy to grind a run down to 0:23.97 and seen no need to spend further time trying for a 0:23.94 or whatever, even when he certainly could have done so, as he saw no hope of taking off a full second.

 

I don't see the harm though, other than preserving ocelot's volatile feelings. You're instigating more incentive to run shorter demos that would not see any competition otherwise. We could, for example, create an arbitrary boundary that any run shorter than a minute or half a minute shows tics, regardless of category. The first person to reach the fastest second would be marked, but so would the actually fastest run when counting tics, maybe using some other icon. 

[olio]

Fantastic effort. For me the biggest issue with getting into Doom running was the lack of centralisation. For most games you go to speedrun.com and you have pretty much the whole history of runs in one spot. Despite the fact that you've stated some runs are currently missing, this is the most comprehensive archive I've seen.

 

The list of rules outlining run eligibility is neat, because the sheer number of different ways you can run Doom can be overwhelming and I had to go back and rerun some demos that I produced with GZDoom because I didn't do my homework right and wanted to submit them.

 

Doom (and any speed game really) needs an 'official' leaderboard so records and competition is readily visible to prospective speedrunners. I understand that with a game as old as Doom and with the sheer variety of ways it can be played, that's much easier said than done, but this site seems like a solid step, moving forward.

 

Newb $0.02, pls no flam

[Linguica]

So the argument against moving to tic precision is that we might see too many people breaking existing speedrun records? And this is bad?

[j4rio]

No, the argument is that old retired farts that may already be dead will be annoyed.

[Linguica]

Whatever, kill your darlings.

[olio]

I think timing to the tic is a good idea. There is no downside to more precision, as long as old runs are given credit for being the first to break a certain second threshold, as I think that is significant. Doom is now old enough that it is multigenerational and like anything competitive, comparisons with past runners breaks down, as more time passes. Circumstances change and mentalities with them, so I think it's futile trying to resist the phenomenon.

[Andromeda]

1 hour ago, olio said:

Fantastic effort. For me the biggest issue with getting into Doom running was the lack of centralisation.

Uhm, http://doomedsda.us has been the definitive speedrun archive pretty much since COMPET-N faded back in 2008 (I think?), so centralization seems a weird issue to have.

[Da Werecat]

Old records are still gonna be broken, so I'm not sure what the problem here is. Unless the possibility of beating a record by a few tics and uploading this very minor improvement is concerning somehow.

[olio]

Once you're in, it feels pretty centralised, but it's not as clear cut for newcomers. I'll use myself as an example. I'm not the smartest guy, but most people aren't the smartest guy so that should be decently representative. 

When I wanted to learn about Doom speedrunning, I went to speedrun.com. I now understand why that wasn't the best place to start, but a lot of people are going to make this mistake. There is a stickied thread in the forum that links to http://doomedsda.us, but not everybody is going to check the forum. I don't know if it's possible to make DSDA more visible on the main Doom speedrun.com page, but it would be helpful if it was. For me, the main problem was that neither these forums or DSDA appeared in my google results for 'doom speedrun' or 'doom speedrunning'. I found old DSDA in that stickied thread and this new DSDA in the speedrun.com discord. That was after finding out that speedrun.com was basically deserted and that things like Compete-N and Competition Doom existed, only to then learn they were abandoned projects. All this was before learning about complevels and how they are required for legit runs. And that was after I had already recorded some apparently non-legit demos. :P

 

And all of that was before I found this place, which is the place you actually post demos. I'm sure a bunch of stuff will bring these forums up as results from google, but when you are looking for doom speedrunning and 'doom speedrunning' doesn't bring up the mains pages where doom speedrunning takes place; it isn't as straight forward or as streamlined as things appear from within the community, once you know where everything is and that it's as simple as making a forum post and your times magically appear on another site.

[DoomedDude]

Does anyone know when the original DSDA site will stop being updated or if it will stop being updated at all?

[GoneAway]

Andy plans on maintaining his site for the time being at least. Even if the site goes down in the future, he will still keep a backup of all the demos that are posted as well.

[Dimon12321]

I noticed there are demos that have YT-videos links included next to their timings.

Are they placed manually or there is a way of automatic submission?

[GoneAway]

There's no easy way to set these yet. Most of the ones there were either added manually or coming from zm's youtube list. We'll figure out a better procedure for this later down the line.

[Poncho1]

Actually, scratch my previous suggestion. Time-consuming, stupid, pretentious idea.

 

I'm sorta with 0xf on this; I do think an updated DSDA is a neat idea, but an icon is a bit silly, to put it gently. Just having the name at the top of a particular category should be enough to show who's got the record. And I do agree that the DSDA is meant to be an archive. It CAN be used as a leaderboard, but I just don't think it's the primary purpose. As usual, just my two cents.

[GoneAway]

Wouldn't singling out a name and moving it to the top be even more intrusive? 😅

 

If the archive didn't store tas or coop or solonet etc runs then it wouldn't matter. I like having everything mixed into the table though, whereas your proposal would require filtering them out or putting them in a separate area. *That* would be sending more of a "these demos are less important" signal than the icons, imo.

 

The primary purpose of the archive is to store demos. This could also be accomplished by having a raw list of download links. I think you'll agree that there's more we can achieve beyond that.

Edited August 4, 2019 by kraflab

[Ancalagon]

3 hours ago, Poncho1 said:

Just having the name at the top of a particular category should be enough to show who's got the record.

 

Spoiler

 

The record here is the 18th demo in the list.

[Poncho1]

2 minutes ago, Ancalagon said:

 

  Reveal hidden contents

 

The record here is the 18th demo in the list.

... Well, you got me there.

[DoomedDude]

How often is the new DSDA site updated with new runs? Andy’s site was updated on the 11th, and the new site hasn’t had any new runs added since before the 11th.