Death in demos

[vita]

What should the dead player do not to invalidate a demo? Should he keep moving or just exit? I'm not talking about -solo-net demos with death abuse.

Edit: not only co-op, but also SP.

Edit 2: currently deaths are allowed only if you exit the map.

[Azuruish]

Well... I died in my non-Tas demo and this demo is still valid so question of death in demo is just aspect of quality and nothing more.. The purpose of speedrunning is time i guess? So non-exit death really slow down the demo and you have to replay if you care about speed too much.

[4shockblast]

To clarify, this is from discussion in the Discord channel on co-op rules for deaths.

The current rule in co-op demos to my understanding is that at the end of a demo, if a player dies then the corpse has to be moving in order for it to still be a valid run. This, from my perspective, is a rule that is both not logical and inconsistent as to single-player, where it doesn't matter whether the corpse is moving or not, all that matters is the exit. This creates a whole lot of potential confusion around co-op runs.

For instance, in certain maps, at a certain point, a player can exit after dying, for example IoS maps. Whether the player is moving or not is irrelevant to the exit trigger. If you were to make an exception specifically for IoS maps, you'd run into a myriad other examples of this. For example, the ending of Kama Sutra map 24 is a telefrag of a barrel that triggers a chain of barrels that pushes a voodoo doll into an exit. In another example, thissuxx map 23 (which doesn't support co-op) has a crusher across the entire map that crushes the player and a Romero's head without any spawner. Still further, it would be possible to come up with contrived examples where all players on a map are crushed early on but the map exits much later via voodoo dolls or something similar to that. Alternatively, only a few players get crushed without escape and only one can only access the exit much later, and potentially the only player crushed or unavoidably killed at some point could be player one, which would mean co-op and single-player both just do not exist. In all these examples, not moving in co-op after death would make for random exceptions to the rule that have to be instituted for co-op to make sense, which complicates the category and is inconsistent with respect to single-player.

Beyond that, the rule is exploitable. Scrolling floors can potentially provide infinite movement for a corpse, so if a player ends up on a scrolling floor after death, they won't ever stop moving and the run would technically be valid. Furthermore, in vanilla, a corpse can continue sliding forever on an edge of a platform, which would also be a form of abuse of the rule.

Ultimately, my point is that the rule just doesn't seem to make sense. It appears to have been made to make death exits possible without having deaths allowed in general, but it seems to have been made without paying any attention to various caveats and Doom bugs as well as features of more advanced ports. I would propose two alternate variations on the rule that would make more sense. First would be some sort of stipulation that a death is allowed if an exit was guaranteeably triggered beforehand. This is very dangerous because an exit trigger could potentially be randomly triggered (for instance, an off-map voodoo doll exit that depends on a randomly moving monster opening up the way for the voodoo doll). Plus it wouldn't solve the situation of one player being forced to die by map design somehow. Alternatively, all deaths could just be allowed, which would make sense to me as it is consistent with how single-player deaths are treated. It would also make routing in co-op more interesting, as a player could tactically death slide to open up a trigger earlier that isn't the exit but makes the exit faster to get to.

[rehelekretep]

is it not simpler to say deaths are allowed, but most would never actually die in a demo because it conveys no advantage and will almost always be a slower demo than one in which the recorder survives? or am i missing something obvious here?

[Aqfaq]

38_ViTa_38 said:

currently deaths are allowed only if you exit the map.

Seems like an arbitrary rule. Death is not essentially different from any other game states. The goal is not to avoid death. The goal is to finish the map/game. Death is irrelevant.

Besides, the player is never truly dead. 100% of the two senses (vision and hearing) remain active after health goes to zero. The player merely stops moving. It is as if he just loses all limbs and gets paralyzed. That is not death.

[Looper]

4shockblast said:

Alternatively, all deaths could just be allowed, which would make sense to me as it is consistent with how single-player deaths are treated. It would also make routing in co-op more interesting, as a player could tactically death slide to open up a trigger earlier that isn't the exit but makes the exit faster to get to.

This gets my vote.

The moving corpse rule is outdated, and I think there's not a single such demo, where the rule is used. I might be wrong though!

Aqfaq said:

The player merely stops moving. It is as if he just loses all limbs and gets paralyzed. That is not death.

Yes, the corpse with ribs clearly showing after taking a rocket to the face... NOT DEAD! :D I swear the puddle was still moving!

[Ancalagon]

Azuruish said:

Well... I died in my non-Tas demo and this demo is still valid

Well... In the "other" category anything is valid.

[Aqfaq]

It is true that I have sent six bullets through the head of the imp, and yet I hope to show by this statement that I am not his murderer. The scariest thing is not that I am unable to kill the imp that paralyzed my comrade. It is the eye. The imp has the eye of a vulture, a pitch black eye that always stares at me with malicious intent. It lies dead, yet the eye keeps staring at me. The same is true for all the dead things in this mad realm! It is utter madness! Even the unearthly ammo box in the corner of the room follows my every move, like an owl whose neck never breaks. The ammo box is always watching me, the twitching imp corpse is always watching me, and for some bizarre reason my dead comrade keeps staring at the imp with glassy eyes that never blink. Nothing dies here. Their bodies rot, but their eyes are filled with living horror and hopeless rage. The owls are not what they seem. I hear metallic footsteps coming from sector 3. I feel that soon I must join the legions of these accursed watchers...



That is not dead which can eternal lie.

[cybermind]

Ok, I have a question here. Let's suppose that you're recording a movie and on some level you don't have enough health from start to perform an arch-vile jump. Is it valid to commit suicide to get 100% of health and successfully perform the AVJ?

[4shockblast]

cybermind said:

Ok, I have a question here. Let's suppose that you're recording a movie and on some level you don't have enough health from start to perform an arch-vile jump. Is it valid to commit suicide to get 100% of health and successfully perform the AVJ?

In single-player, any sort of death exit is fine, AVJ or death slide or whatever - the only rule, obviously, is that you exit. For instance, Plutonia map 17 rocket death slide, for instance, is allowed in movies. Only co-op rules have this weird caveat to them.

[j4rio]

The obvious intention of c-n rules is simply put "don't die, but death slides triggering exit are ok".

Don't get your panties in a wad.

Azuruish said:

Well... I died in my non-Tas demo and this demo is still valid

Nobody bothered to watch it so nobody could disqualify it.

[cybermind]

4shockblast said:

In single-player, any sort of death exit is fine, AVJ or death slide or whatever - the only rule, obviously, is that you exit. For instance, Plutonia map 17 rocket death slide, for instance, is allowed in movies. Only co-op rules have this weird caveat to them.

You got me wrong. I know that these are possible. I mean, imagine a situation where you record a movie and while recording you start one of the maps with HP < 90. There is an arch-vile that could help you to skip half of the map (not to exit directly by means of death slide) but there are no stimpacks/medikits, so the only way to use that arch-vile is to commit suicide and restart the map with your health raised to 100. Is that considered valid?

[ZeroMaster010]

cybermind said:

You got me wrong. I know that these are possible. I mean, imagine a situation where you record a movie and while recording you start one of the maps with HP < 90. There is an arch-vile that could help you to skip half of the map (not to exit directly by means of death slide) but there are no stimpacks/medikits, so the only way to use that arch-vile is to commit suicide and restart the map with your health raised to 100. Is that considered valid?

It is not considered valid, and for the same reason I agree with the current co-op rule. Allowing deaths opens up for low effort demos in my opinion, in these very rare cases where it could be useful for someone to slide under to trigger something for the other player it would be enough to make an exception to the rule, instead of allowing players to die in co-op unintentionally and still continue.

[4shockblast]

There's a difference between respawning on the same map and dying and staying dead until the next map. I'm not suggesting that respawns should be allowed, but I am saying that any death where a player is dead until the next level is fine because it matches the current single-player rule. The current rule may have the intention of only deaths that trigger exits being fine, but it doesn't achieve that purpose as I showed with a bunch of possible examples where it wouldn't work. I'm also against having to rule in every possible example where it could be useful for a trick or where the death is fine instead of just allowing all of them.

[dew]

ZeroMaster010 said:

It is not considered valid, and for the same reason I agree with the current co-op rule. Allowing deaths opens up for low effort demos in my opinion, in these very rare cases where it could be useful for someone to slide under to trigger something for the other player it would be enough to make an exception to the rule, instead of allowing players to die in co-op unintentionally and still continue.

Exactly. The "corpses must move" wording sounds obscure and awkward, but it works very well for its intended purpose: running coop on CN wads. People here immediately jumped to special cases like non-exiting death slides, delayed exits, Boom scrollers or crushed Romero heads, but none of those are present in the CN wads. The rule is establishing one simple concept: if you screw up and die, the run isn't valid even if your teammates manage to exit. The sliding corpse exception was inserted later as an answer to death slides.

I'm not against rewording the rule for our general usage (outside of CN) to reflect the discovery of new tricks, emergence of post-death "cutscenes", so on and so forth, but none of those should undermine the core concept of, um, not screwing up and dying in a demo you're presenting as a record submission. If a team of players wants to play with "sacrifice abuse", like e.g. having a dedicated aggro-drone that dies in every map after distracting the most dangerous monsters, that's fine, just mention it in the category description and maybe Andy can add it into the comment while still keeping it in the coop section. However a massive retroactive redefinition of rules to the existing category isn't too appealing to me, especially when we're discussing it over a TAS while very little actual coop demos get made.

[Looper]

dew said:

The rule is establishing one simple concept: if you screw up and die, the run isn't valid even if your teammates manage to exit. The sliding corpse exception was inserted later as an answer to death slides.

Makes sense.

[Memfis]

I remember this caused a slight "controversy" when the French guys made a coop e3m8 demo for C-N where one of them died, I think after the spiderdemon was killed. So it was kinda like Doom 2 Map30 where you're allowed to die after you destroy Icon of Sin, except in e3m8 you need to be alive in order for the level end trigger to work... But in coop it still works even if some players are dead. So it got a bit confusing. In the end they just recorded a faster demo instead of trying to figure out the exact rules.

[4shockblast]

dew said:

The rule is establishing one simple concept: if you screw up and die, the run isn't valid even if your teammates manage to exit.

You're only protecting yourself against one kind of screw up, though; it's still possible for anyone to make a low-effort co-op or SP demo with a myriad of mistakes that won't invalidate the demo if it exits. It would make sense for the same to be the case in co-op; if someone screws up but a different player can still exit, it's an exit. I don't expect this to change officially in Compet-N, especially given that that competition is practically unchanging at this point (both in submissions and rules), but in the general case, I do not agree with the rule or guarding against one arbitrary sort of error that can happen in runs.

[TimeOfDeath]

I don't like the rule: "deaths aren't allowed, oh except for death exits".

Coop and -solonet are unique in that you can die and respawn without the map restarting. Therefore, I think there should be two categories, "no deaths" and "deaths allowed". That's what I've done for some solo-net demos. Some coop wads are actually designed for deaths and respawning. Calling demos with deaths in them "low-effort" isn't fair. Obviously you can optimize a deaths-allowed demo just like a no-deaths demo, and a deaths-allowed demo should have strategic deaths to save time.

[4shockblast]

I would say that there are three separate categories for deaths in demos. No deaths allowed at all, which is something that doesn't exist in Doom speedrunning at all currently. Deaths allowed but no respawns, which matches the current SP rules exactly and is closest aligned to the co-op/-solo-net rules which is why I am suggesting it as a general rule for most runs. Finally, there could be runs with respawns which would only make sense in co-op or -solo-net. The above idea of dying on a map in order to start over with enough health for a trick doesn't make sense at all to me because Doom uses in-game time to measure its runs, and a strategy like that would make sense only with RTA (otherwise the time up to the death on the map has no effect on the final time).

Some runs with respawns already exist in -solo-net, and I don't think there is necessarily anything wrong with any of the three rulesets because all three are just separate arbitrarily chosen ways of determining validity of a demo and all three result in their own tactical changes and strategies. But yes, the problem is when there is some arbitrary exceptions for death exits that increase with complexity as people find out more and more things about the engine (corpse moving, IoS, etc.) which is the current situation. CN will obviously not change, but everything else would make a lot more sense if it followed one of the above concrete rulesets.

[dew]

TimeOfDeath said:

I don't like the rule: "deaths aren't allowed, oh except for death exits".

Given forced death exits, there's technically no way around that. And I wasn't belittling coop/solonet demos with player respawn, I made one myself a few years ago, heh. The "low-effort" part stems mostly from opening the category to unintended deaths, which I personally consider a deal breaker.

[GoneAway]

I don't understand the "low-effort" argument. One player dying in a coop run may still end up with a really great time, while another run where neither player dies could be horrible because they failed a trick 20 times. But the horrible run is more effort somehow because no one died? Death is just another failure. Time is time, why try to decide what errors are more important than others? Doom is already open infinitely to shitty exits, this does nothing to prevent them, yet discounts potentially good runs. So what exactly is the point?

Speedruns are supposed to be about getting a good time and limiting errors, but this rule is just an arbitrary "challenge" aspect about getting a run without dying, which is not at all the same thing.

[j4rio]

At this point both sides have their main argumentation point boiling down to "I don't like it this way".