United StatesI've been poking around with the debug menus on MAME to try to see what's different. Info on them here for folks that don't know about them: https://tcrf.net/Darkstalkers:_The_Night_Warriors_(Arcade)/Debug
Kinda seems like there might be an inconsistent number of delay frames on T3 for chains. I wonder if it's merely dropping frames of startup and end animation to achieve that speed.
If so, there might be more viable high damage strats on T2 than T3 in general. Interesting.
Damn, that's a good one! That's much more doable than forcing him into a teleport loop. Kinda wonder if you can occasionally put a screw shot between them for more damage since the distance is close.
Ah after playing with it a bit, I get it! This exploits the fact that Rikuo's Crouching Heavy Punch can only be blocked high in the early Japanese ROM. The later Japanese ROM does not have this. I've tested this out before and it didn't seem to work for me, but I've never tried it on Turbo 2.
Makes me want to compare the animation frames... really wonder what's so different on T3.
This would work on other characters, but there's a lot where this would be hard to execute on as their responses to this aren't favorable.
It's fine to play on that version, however since the 1.50 patch is not officially sanctioned, not included in any official release like 1.40b23, and adds significant features / makes significant changes, we'd likely have to give it it's own variable and separate runs out.
That shouldn't stop you if you want to do runs on the modded version though! Please feel free to post a completed run on 1.50 here in this thread and I'll add a place for the modded versions of each of the runs to the board!
Hi all,
I just made a few tweaks to the board and added a few categories for likely the most common versions. If there's a version you'd like to run that isn't here yet, please feel free to ping me or post here! I know there's a million versions of this game and some that aren't even documented in the Wikipedia article yet.
There's likely some versions (I suspect DOS and Amiga for example) that may be similar enough to be grouped together, but we'll have to see. There's already some surprising differences in the versions I've been testing.
You might notice the 30k population goal is now a variable and is now named State Capital. This is because I discovered that the Playstation version actually gives you State Capital at 10k! Weird, I know! If we do eventually want to expand the board, it will be easier to add another population goal for say, the Military Base, or Arcos now.
I've also grouped "Glitchless | Defualt Map RNG" and "Glitched | Any Map". It's very inelegant (sorry!) and if anyone has any suggestions on how to shorten that up and simplify things without losing the meaning, I'm all ears. We could just make it Default Glitchless and Any% I guess. I figure these are likely the only things folks would be interested in running based on feedback but we could separate things more if wanted.
Please feel free to voice concerns or things you'd like to see here!
From a Quality of Life standpoint yeah, just makes more sense.
Just as a follow up, I had a chance to take a look at all the current runs and it's clear we can see a demarcation when the button is released at many different cycle speeds. Either the button itself is visible or more commonly the first frame where the window begins to disappear on higher cycles makes it easy to tell. So I think it's pretty obvious that the timing should start there as @Arran suggested.
This, timing ending on the first frame where the paper is fully visible, having the map generation within the submission video, and allowing the use of MAKE and having at least 1 River/Coast seems to be all we need to define in terms of rules!
I should have time to spruce things up and add to the board next weekend. Let me know if there's anything you'd like me to add or clarify! I'll be doing up a How to on changing the cycles for speedrunning on the DOSBOX config then as well.
So here's what I mean:
(Wife joking about 2020 being terrible included!)
And yeah, I guess the point is we probably want folks to show the map generation values one way or another from now on, as that is the quickest way to prove nothing's been cheated. Make button would have to be pressed as just using the map you just did a run on leaves it in an edited state.
By the Terrian Generator popup window that flashes up whenever a map is generated, the correct values for a default map are:
Hills 25% Water 10% Trees 31%
I also highlight the scenario loads here and how they throw a wrench into things by allowing you to move the river to the coast, but not ever eliminate either. Another point to be made is that the river/coast mechanic is clearly something the developers intended the RNG engine to make use of. It is rather balanced and so long as MAKE is used for more variable seeds, there might be little benefit to coast over river in reality.
Only problem is, you can load a scenario and move the river to the coast and change seed RNG as well in this. We could outlaw the edit mode, but it would still be possible to do what I mentioned through game mechanics, and it'd be difficult to definitively rule that someone had or hadn't loaded a scenario at some point.
Just banning edit mode won't solve the need for at least some complicated rules.
The end goal is partly to make it so that we don't have to try to determine all of this retroactively, which would probably be impossible. Luckily, we can see the map generation values every time the New City loads, the Make button is pressed, and loading a scenario does not change the default Water, Hills, and Trees values for RNG, only the map seed.
I'll do a video demonstration of this later today to show what I mean. The main take away is, the base land defaults aren't changed via either MAKE or Scenario loads.
The start frame sounds reasonable enough. I'll have to see if we can get some sort of common visual cue on that in the current runs.
I agree on the map reroll personally. I've recently discovered that the MAKE button on the map edit screen resets the seed every time which tends to give you slightly better chances than just pressing New City. It probably should be legal to use so long as you show the full generation per attempt to confirm no cheating since it's the easiest way to reset the seed beyond constantly quitting and reloading the game.
How do you feel (and others feel free to chime) in on the river/coast switch? Since it doesn't change the fundamental RNG and it's a part of the game's generation algorithm that you can access, it seems like it's fair game to me. Though we probably should specify that you must have at least 1 river or coast.
If anything we should probably be promoting things that make this ~1 minute or less run less tedious so long as we can ensure a fair playing field!
Though it is interesting to note that some of the versions, notably SNES, has no map generator and one of the choices is entirely flat. If we ever add more game versions that'd be something to account for.
I like the idea of making glitched into an anything goes category. Sounds fun to me!
And yeah, I plan on making the rules (and a guide to them) a bit more clear soon. Just want to make sure everyone is on the same page about what they should be. =P
If there's any interest in running one of the many versions of this game on a console not specified yet, please also feel free to mention here! I have a semi competitive run of this on SNES done but I feel like it doesn't need a board if it's only me. I know for some seeing an empty Category is a motivator though.
Okay, I think we have a consensus among the runners now.
We're going with option 2: Change the Arcade perfect rule set to allow up to Turbo 3 and keep everything as is. Since we're the only runners of this version, now would be the time.
Arcade category runs will allow up to Turbo 3 to fit more with the tournament scene. You're welcome to use lower than Turbo 3 in your runs as there certainly is merit for being the fastest on lower speeds, and I'll hopefully be adding a variable soon to account for this.
If you have any questions or objections, please feel free to voice them here or DM me!
Yep. The Ressurction bundle standardized things. Rest assured that the extra stars are above what the arcade had to offer. Setting things as if the max was 8 is valid for runs. The Turbo max was probably a mistake and I'll correct it shortly.
You could technically get those extra "9" settings in the ROMs via ram edits, they just aren't exposed in the setup menus.
Yes, the settings would need to be:
Difficulty 4/8 stars 99 seconds Turbo speed 2/8 stars Damage gauge 2/4 stars
FYI, there does seem to be a consensus that we raise the Turbo to accept up to Turbo 3 in the arcade perfect category though, I just need to have one more runner weigh in to change it officially. Please feel free to submit at Turbo 3 if you'd like, we'll either add a category for it or change the main board officially soon.
It's the one category where the defaults on the arcade machine don't make sense as it adds duplicate animation frames and the tournament scene plays on Turbo 3.
It's likely needed for CD32, bare minimum. Aside from the hilarious loads, the game's AI is... actually pretty good? It's at the very least vastly different in controls and gameplay. I can definitely see how including this as a pack-in game KILLED the CD32. It's poison for the eyes and ears, even with CD quality audio!
Amiga is actually fairly similar to DOS so we might be able to get away with it, but something seems a bit off on the control scheme. I gotta see if it's a bad disk image or if the manuals are just plain wrong or something.
Geez this game.
Soooo... All the versions of this game actually are vastly different. I'm thinking we need to break this up to DOS Amiga and CD32. CD32 might actually be a decent yet still awful challenge!
Okay, so after some testing on an actual DOS laptop and a DOS Virtual machine, it does look like higher processor speeds to make Date transitions move faster. You can emulate this in DOSBox by setting the configuration to run the game at around 200k to 300k cycles. Basically the equivalent of an 833Mhz Pentium 3 is the cutoff for performance gains - much beyond that and the game accumulates a bunch of interface lag in too many places, and the Date transitions go slower.
I've known about this behavior in DOSBox for a while but haven't had a chance to test on real DOS until now.
Given that we can get this to happen on real hardware, and it's easy to configure, it's hard to argue against it's use in runs, but please feel free to voice concerns!
@camero56 @AquaBlake @ShinkenX
Hey All,
Just wanted to quickly ping all the runners here as I've just noticed something in a submitted run and a run we already have up on the board.
Camero56's two most recent accepted runs are actually using Turbo 3 in the settings, but the current rule set lists Turbo 2 as the accepted turbo setting. I've manually frame counted a few common moves from all of the runners and determined that Cam's 5:18 run, AquaBlake's latest, and ShikenX's current run do appear to be on Turbo 2.
I do apologize for not catching this sooner, but I thought it might be a good time to talk about the rules for arcade. Turbo 2 is also confirmed 20% slower by my frame count than Turbo 3 which does make it an uneven playing field. In both a tournament setting and online play, Turbo 3 is the commonly accepted Turbo speed as it has no cut or duplicated frames. There might be some merit in changing things to accept Turbo 3 for Arcade perfect, as it does play much more like a tournament setting.
So, we could either do one of two things:
-
Keep the Turbo settings on Turbo 2 for Arcade and remove Camero56's times that use Turbo 3. No leaderboard positions would be affected.
-
Change the Arcade perfect rule set to allow up to Turbo 3 and keep everything as is. Since we're the only runners of this version, now would be the time.
What does everyone think?
I'm personally in favor of #2, changing things to accept up to Turbo 3, but it doesn't matter much to me either way.
Once again, I do apologize I didn't catch the settings screen sooner Cam, entirely my fault.
