Post by dwangoAC

Posted: 1/30/2021 9:02 PM

Quote

Memory

She/Her

Site Admin, Skilled player (1551)

Joined: 3/20/2014
Posts: 1765
Location: Dumpster

Giving up isn't just negative feedback, it's bad advice. Good advice would be either to submit with the current strategy or what have you.

[16:36:31] <Mothrayas> I have to say this argument about robot drug usage is a lot more fun than whatever else we have been doing in the past two+ hours

[16:08:10] <BenLubar> a TAS is just the limit of a segmented speedrun as the segment length approaches zero

Posted: 1/30/2021 11:04 PM

Quote

dwangoAC

Moderator, Senior Ambassador, Experienced player (906)

Joined: 9/14/2008
Posts: 1014

Regarding previous comments in general, I will say that the April 1st submission was obviously rushed and was in part an amusing way to mess with ais523 who took it in stride (I actually gave him the ability to nuke the entire submission off the face of the earth if he chose to). Clearly, there's no room for taking a really optimized run and then finishing it off in an extremely sloppy way and that's definitely not what I'm advocating for here. Let's be honest about what we are - we're perfectionists. The whole lot of us. We want things to be absolutely in order, as optimized as possible. I very much want to see us pull off a tight run, and I think we can do a 2003 turn plan that's both tight and optimized based on the rules we know about. There was an assertion of procrastination but that's not what's happened here per se - it's more that we've been stuck trying to find workarounds to some of the most absurd problems any TAS has ever faced. NetHack has an incalculable number of possible paths unlike a simple platformer where there are only so many routes you can try. I fully believe we can both make a quality run that lives up to expectations and make progress around the blocking issues by working with routes we know actually work. Even the 2003 turn options are utterly preposterous after all. We'll worry about presentation of the run later - this is the kind of run that would require a very large amount of explanation no matter what form that takes, whether it's the copious notes referred to before or a possible video breaking down what's going on at each stage. I believe the final form can in fact be a positive result. We're half a million rerecords in on this TAS between automated and manual tests and we might be at twice that when we're done but I have confidence we *can* finish this and do it with excellence.

I was laid off in May 2023 and became too ill to work this year and could use support via Patreon or onetime donations as work on TASBot Re: and TASBot HD is stalled. I'm dwangoAC, TASVideos Senior Ambassador and BDFL of the TASBot community; when healthy, I post TAS content on YouTube.com/dwangoAC based on livestreams from Twitch.tv/dwangoAC.

Posted: 1/30/2021 11:07 PM

Quote

Patashu

He/Him

Joined: 10/2/2005
Posts: 4042

Throwing my hat in as well, I'd much rather see a good-as-we-know-how-to-make-it Nethack TAS with a possible improvements section than to never see a Nethack TAS. Whatever you choose, good luck!

My Chiptune music, made in Famitracker: http://soundcloud.com/patashu My twitch. I stream mostly shmups & rhythm games http://twitch.tv/patashu My youtube, again shmups and rhythm games and misc stuff: http://youtube.com/user/patashu

Posted: 2/1/2021 1:56 PM

Quote

ais523

Editor, Player (44)

Joined: 7/11/2010
Posts: 1029

From my point of view, finding an optimal plan is actually not what the TAS has been stuck on for a while. I definitely wanted to hit the 2003 turn mark (which seems to be possible with room to spare at this point), but 2002 is probably too much. The basic reason behind this is that some strategies, while it's great to know that they're possible, simply aren't going to be any fun to TAS or any fun to watch. For example, say you hypercharge an engulfing monster and get it to carry you across Air, randomly stumbling into the portal. The minimum possible travel distance for that is 33 squares, and hypercharged monster movement is effectively a random walk; and the average time a walk that randomizes between left and right takes to reach any particular square is proportional to the square of the distance. 33² is 1089. Take into account the possibility of moving up or down (this happens ¼ of the time), and we're going to need an average of around 1452 hypercharges to reach the horizontal position a minimum-distance portal (somewhat more to stumble into the correct vertical position, too, but that's minor by comparison). Luck manipulation can bring this number down a bit, but the more you try to save on hypercharges, the more luck you need, and there are limits to how much it's reasonably possible to luck-manipulate. Now, let's consider that hypercharging a monster requires spending turn boundaries; this isn't something you can increase the speed of by using a faster polyform. Let's say we get insanely lucky and find an RNG state that crosses Air in just 1000 hypercharges (there may not be one of these in NetHack's seed space, or it may require many decades of wallwalking, etc.). The fastest we can hypercharge a monster is 4 hypercharges per turn boundary, so the final run will involve 250 turns – ⅛ of the run – spent just repeating the same glitch over and over again. There might be a way to make that interesting to watch by doing other things in between the hypercharges using spare actions, but there are limits, and it would be incredibly tedious to TAS regardless. And this is just one glitch! My basic problem with the 2002-turn plans is that they require strategies with this level of tedium all over the place. Shaving off actions is great, but at some point the cost of entertainment – and the cost to the TASers' sanity – is going to make the time saving not worth it, at least if you want any chance of the run to be finished. At this point, I think 2002 turns is something that isn't desirable to aim for unless some new glitches are discovered that make such a run actually tolerable. (2003-turn strategies were in this area for a while, but have since been improved, which is what started to bring the 2002-turn strategies within range.) Instead, I think that what would lead to the best and most entertaining run would be to hit 2003 with a lot of strategy variety. I think it'd be both more fun to watch, and more fun to TAS, if we can come up with an endgame strategy which isn't just enexto'ing over and over and over again. There are lots of potential strategies we've discovered, like overflow hurtles and zero-timed descension runs, and I think that using those to help hit 2003 would make a better run even if they aren't quite necessary. Anyway, the reason I'm writing this is partly because I forgot to write it earlier, but partly because I think people are assuming that I'm waiting for a 2002-turn plan before resuming the run. That's not the case – I'm happy to run on a 2003-turn plan. However, I've had various problems in my personal life, and TASing hasn't been a huge priority for me as a consequence (and a lot can change in terms of personal circumstances in 11 years); I've had to put a lot of hobbies on hold, and TASing is one of them. That's not to say I've abandoned the TAS; it's more like I've put it on hold until I can have the opportunity to work on it again. (I think dwangoAC has also been busy with other things, quite understandably! So there hasn't been much progress in terms of actually grinding out rerecords.)

Posted: 2/1/2021 9:52 PM

Quote

Patashu

He/Him

Joined: 10/2/2005
Posts: 4042

I'm fine with a TAS that has boring parts to watch (I can just skim over them, and honestly I'll be primarily enjoying the TAS through the text description I imagine), but obviously it should be exciting to TAS or it'll never get made. So, whatever you're most enthusiastic to make is the best choice!

My Chiptune music, made in Famitracker: http://soundcloud.com/patashu My twitch. I stream mostly shmups & rhythm games http://twitch.tv/patashu My youtube, again shmups and rhythm games and misc stuff: http://youtube.com/user/patashu

Posted: 8/5/2021 5:33 PM

Post subject: NetHack TAS progress resumed, join Libera ##nethack-tas

Quote

dwangoAC

Moderator, Senior Ambassador, Experienced player (906)

Joined: 9/14/2008
Posts: 1014

A recent run of NetHack any% by luxidream at ESA sparked renewed interest in completing the NetHack 3.4.3 "Lowest in-game turncount" TAS and I'm happy to announce we now have a home for discussions over at ##nethack-tas on Libera IRC in a bridge. There are a number of interested parties who are now versed in the basics of using the rerecording framework I designed which I managed to get stood up on a new system. There's been a bit of derusting (we forgot how to get into Santa mode until we did some source diving and rediscovered that it's X as in shift+x) but there's a lot of new energy with some very talented people involved. So far, we have direct help from luxidream, Antigulp (Kes), Stella, and machinespray along with additional support from spicycat, tone, and Kougyoku among others. We're rediscovering how overcharging works again and trying to work out what things still need tested and documented. I'm excited to see progress again. For ais523, I hope you're able to pop in to IRC to answer a few questions but we'll do everything we can to respect the limited time you have. I'm also busy most of August between presenting at DEF CON and building a flamethrower for the Skorpios BattleBots team / going to the recording sessions but I'll be here to support as best I can. I want to see a 2,003 turn plan succeed with good entertainment and I think we can do it. Thanks to everyone who has stepped up!

I was laid off in May 2023 and became too ill to work this year and could use support via Patreon or onetime donations as work on TASBot Re: and TASBot HD is stalled. I'm dwangoAC, TASVideos Senior Ambassador and BDFL of the TASBot community; when healthy, I post TAS content on YouTube.com/dwangoAC based on livestreams from Twitch.tv/dwangoAC.

Posted: 5/22/2022 1:13 AM

Post subject: Wiki page created, progress being made

Quote

dwangoAC

Moderator, Senior Ambassador, Experienced player (906)

Joined: 9/14/2008
Posts: 1014

We're making renewed progress and thanks to help from Kes there's now a wiki page up at at https://nethackwiki.com/wiki/User:Agulp/Tool_Assisted_Speedrun to document the project a bit better. We're focused on producing a 2,003 turn run that doesn't make major sacrifices to entertainment. It's a monumental effort that continues to consume hours beyond count for those involved but the payoff is starting to materialize. We're going back to "hex in" some changes to adapt what we've done with a new plan and we're working on better documenting what that plan actually is so folks can make meaningful progress. All that to say, the dream isn't dead - please feel free to get involved if this project interests you!

I was laid off in May 2023 and became too ill to work this year and could use support via Patreon or onetime donations as work on TASBot Re: and TASBot HD is stalled. I'm dwangoAC, TASVideos Senior Ambassador and BDFL of the TASBot community; when healthy, I post TAS content on YouTube.com/dwangoAC based on livestreams from Twitch.tv/dwangoAC.

Posted: 5/22/2022 1:18 AM

Quote

Samsara

She/They

Senior Judge, Site Admin, Expert player (2108)

Joined: 11/13/2006
Posts: 2821
Location: Northern California

LET'S HECKIN' GOOOOOOOOOOOOOOOOO

TASvideos Admin and acting Senior Judge 💙 Currently unable to dedicate a lot of time to the site, taking care of family. Now infrequently posting on Bluesky

warmCabin wrote:

You shouldn't need a degree in computer science to get into this hobby.

Posted: 5/22/2022 1:50 AM

Quote

Patashu

He/Him

Joined: 10/2/2005
Posts: 4042

2003!

My Chiptune music, made in Famitracker: http://soundcloud.com/patashu My twitch. I stream mostly shmups & rhythm games http://twitch.tv/patashu My youtube, again shmups and rhythm games and misc stuff: http://youtube.com/user/patashu

Posted: 5/22/2022 2:18 AM

Quote

Animand

Joined: 8/22/2006
Posts: 24
Location: Georgia

Super excited!

I'm not really here. Pressing Up+Down too many times glitched your graphics. And audio. Yeah.

Posted: 8/9/2022 3:17 PM

Quote

timvisher

Joined: 8/9/2022
Posts: 4

Still _super_ excited at the prospect of seeing this run! :D

Posted: 2/14/2023 3:05 PM

Quote

timvisher

Joined: 8/9/2022
Posts: 4

Just coming here to say that this is still the TAS I'm most excited to see. :)

Posted: 8/8/2023 2:06 PM

Quote

timvisher

Joined: 8/9/2022
Posts: 4

Would still be over the moon to see this TAS completed. Hope everyone's doing well who's a part of it.

Posted: 11/16/2023 2:54 PM

Quote

timvisher

Joined: 8/9/2022
Posts: 4

Sending good vibes to the NetHack TAS team. :)

Posted: 2/9/2024 7:52 PM

Quote

ais523

Editor, Player (44)

Joined: 7/11/2010
Posts: 1029

I just thought of a simple way to make the events on the quest goal level much easier to manipulate, without costing any time: use two quantum mechanics. The events on the second monster action of turn 2000 would be "Archon ORKOs the Dark One, quantum mechanic teleports us onto the Dark One's corpse and we autopickup the Bell, quantum mechanic teleports us next to a level teleport trap, steed throws us off into the trap". For both teleports, we have access to teleport control, so the only events that need manipulating are a) the Archon landing next to the Dark One, b) there being a level teleport trap somewhere on the level, c) the steed throwing us off in the right direction, d) arriving on the correct square after the teleport. There is some scope to use monster actions in between these in order to help to manipulate the results. The worst of these to manipulate will be d), which may be solvable via the "fire the arrow at the tree, then paint the target around it" principle (i.e. we first find a seed that works and check to see where we would end up, and then do a setup on the Quest home level that works starting from that specific square) – I estimate that the odds of there being a setup for the Quest home level that works from some specific random square are around 1 in 4, although lining up the RNG for that could be difficult because lots of random numbers will be consumed during the setup itself.

Posted: 2/10/2024 6:21 AM

Quote

ais523

Editor, Player (44)

Joined: 7/11/2010
Posts: 1029

Some progress: I've confirmed that monster hypercharging is a real glitch (and have documented it at Wiki: GameResources/DOS/Nethack#MonsterHypercharging). The setup that I used for the tests was to be in the a corner of a room, with a hole next to me, and water on all the other adjacent squares and below me: a dragon is flying over the water, with my character in homunculus polyform (which is speed 12, able to fly, and bad defensively). Waiting the turn would cause the dragon to attack the character, rehumanising it from HP damage and causing it to climb out onto the square with the hole, immediately falling through to a lower level. The monsters on the level that I fell onto would gain a turn's worth of actions, and if I immediately level-teleported out and repeated the process, the monsters would gain another turn's worth of actions, and so on. (I also confirmed that, if I waited on the new level rather than level-teleporting out immediately, the monsters would immediately take all their stored actions.)

Posted: 6/30/2024 8:36 PM

Quote

feos

Site Admin, Skilled player (1250)

Joined: 4/17/2010
Posts: 11473
Location: Lake Chargoggagoggmanchauggagoggchaubunagungamaugg

Dumb question: why does the game run so fast in the tas? Is it because DOS input polling is that quick, and there's no other limitation?

Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.

Posted: 6/30/2024 9:43 PM

Quote

ais523

Editor, Player (44)

Joined: 7/11/2010
Posts: 1029

The game's limited only by a) lag and b) reading input from the keyboard. Lag does have some effect, but the speed at which input can be read is often the limiting factor, especially during luck manipulation (because the luck manipulation techniques are designed to minimize the amount of lag produced). There are very few moments where the game intentionally waits for the next frame, or performs a timed delay (and timed delays can be disabled in the options). So in most cases, you can play the game as fast as you can press the buttons. Although I'm not 100% sure, I think that input on DOS uses a maskable interrupt rather than polling, i.e. it actually interrupts the processor in much the same way that, say, vblank would on a NES. That means that if you press a key, it's handled as soon as the keyboard manages to communicate that fact to the processor – there's no "poll once a frame" or equivalent. The limiting factor on luck manipulation speed is therefore the rate at which the keyboard is able to send the "a key was pressed" message.

Posted: 7/1/2024 2:29 PM

Quote

feos

Site Admin, Skilled player (1250)

Joined: 4/17/2010
Posts: 11473
Location: Lake Chargoggagoggmanchauggagoggchaubunagungamaugg

ais523 wrote:

Although I'm not 100% sure, I think that input on DOS uses a maskable interrupt rather than polling, i.e. it actually interrupts the processor in much the same way that, say, vblank would on a NES. That means that if you press a key, it's handled as soon as the keyboard manages to communicate that fact to the processor – there's no "poll once a frame" or equivalent. The limiting factor on luck manipulation speed is therefore the rate at which the keyboard is able to send the "a key was pressed" message.

Is there any info on how often this happens?

Warning: When making decisions, I try to collect as much data as possible before actually deciding. I try to abstract away and see the principles behind real world events and people's opinions. I try to generalize them and turn into something clear and reusable. I hate depending on unpredictable and having to make lottery guesses. Any problem can be solved by systems thinking and acting.

Posted: 7/1/2024 7:17 PM
(Edited: 7/1/2024 7:19 PM)

Quote

Warepire

He/Him

Editor

Joined: 3/2/2010
Posts: 2178
Location: A little to the left of nowhere (Sweden)

ais523 wrote:

Although I'm not 100% sure, I think that input on DOS uses a maskable interrupt rather than polling, i.e. it actually interrupts the processor in much the same way that, say, vblank would on a NES. That means that if you press a key, it's handled as soon as the keyboard manages to communicate that fact to the processor – there's no "poll once a frame" or equivalent. The limiting factor on luck manipulation speed is therefore the rate at which the keyboard is able to send the "a key was pressed" message.

That sounds correct, input was usually handled by a PS/2 chip (e.g. Intel 8042, but this is newer). Which was usually routed directly to the I/O port interface of the CPU. See https://wiki.osdev.org/I/O_Ports and https://www.dosdays.co.uk/media/intel/Intel%208042.pdf This was standard input processing on x86 before USB took over keyboards and mice/mouses.

feos wrote:

Is there any info on how often this happens?

This would depend entirely on the speed of the processor and how much higher-priority load it's under.

Forum MS-DOS Games NetHack Game: NetHack

Forum

MS-DOS Games

NetHack Game: NetHack