Berserked punching randomness

Been wondering this for a long time, what's the deal with the berserk? Why sometimes a pinkie/spectre can take more powered-punches than a hell knight to kill? to mention an example, my first time punching a baron, he died after 6 punches. Usually pinkies/spectres take in between 1 to 5 punches, there's a lot of HP of difference in between the two. It all has to be because of RNG, but I don't understand so much about this topic, and the berserk seems to give the most random RNG. Seems...  

The berserked fist has a range of 20-200. Yeah it's dumb. 

Is there an existing GZDoom mod that averages out all the damage values to make them always consistent? Seems like something interesting to try out and see how it feels.

1 hour ago, rdwpa said:

The berserked fist has a range of 20-200. Yeah it's dumb. 

That's really interesting, actually. 

That's because the Berserk powerup just multiplies the 2-20 damage by 10, right? So 2*10=20 and 20*10=200.

They didn't think about that and the damage range ended up being really stupid lol.

2 hours ago, Spectre01 said:

Is there an existing GZDoom mod that averages out all the damage values to make them always consistent? Seems like something interesting to try out and see how it feels.

I think that there's no such mod to fix the value of the berserk...

BTW, if a Pinky tooks 5 fists to take down, you have the "best" rolls of all time.

There are a lot of RNGs in Doom. Occasionally, you can kill a Hell Knight or an Arachnotron with 2 rockets or something like that. That's why I always say Doom speedruns need not only skill, but also luck.

40 minutes ago, GarrettChan said:

There are a lot of RNGs in Doom. Occasionally, you can kill a Hell Knight or an Arachnotron with 2 rockets or something like that. That's why I always say Doom speedruns need not only skill, but also luck.

How positive are you that this actually happens with Doom's PRNG table? And not just with a source port that changes the table? Remember that Doom has a very uniform table with a lot of connected occurrences, this results in things like the SSG having a very balanced damage and spread despite the successive PRNG calls, because its resulting hitscans, in order, have an effect on each following hitscan when chewing through nearly the whole table. This is opposed to ZDoom which completely changes the PRNG algorithm and gives every kind of action a unique table, allowing for more results of higher/lower successive outcomes, and in some situations allowing for damage outcomes completely impossible in vanilla.

18 minutes ago, Edward850 said:

 impossible in vanilla.

I don't know what made you think that 2 Rockets MUST be 2 successive calls from the PRNG table. Even they are two numbers next to each other, that 7 and 255 seems OK to me.

9 hours ago, Spectre01 said:

Is there an existing GZDoom mod that averages out all the damage values to make them always consistent?

All you need is this DECORATE: (based on the last example on this page, I set Berserk damage to 150 so that Pinkies die in one hit)

actor CustomPlayer : DoomPlayer {
  Player.StartItem "Pistol"
  Player.StartItem "CustomFist"
  Player.StartItem "Clip", 50
  Player.WeaponSlot 1, CustomFist, Chainsaw
}
ACTOR CustomFist : Fist
{
  States
  {
  Fire:
    PUNG B 4
    TNT1 A 0 A_JumpIfInventory("PowerStrength", 1, "Berserked")
  Normal:
    PUNG C 4 A_CustomPunch(2 * random(1, 10), TRUE)
    Goto FireEnd
  Berserked:
    PUNG C 4 A_CustomPunch(150, TRUE)
  FireEnd:
    PUNG D 5
    PUNG C 4
    PUNG B 5 A_ReFire
    Goto Ready
  }
}

And this MAPINFO:

GameInfo {
   PlayerClasses = "CustomPlayer"
}

2 hours ago, GarrettChan said:

I don't know what made you think that 2 Rockets MUST be 2 successive calls from the PRNG table. Even they are two numbers next to each other, that 7 and 255 seems OK to me.

You miss quoted the wrong part, as I was saying some actions were impossible (for example, ZDoom's SSG & BFG potential damage output is higer), not all of them. I was questioning if your experience was valid to doom speedrunning in regards to Arachnotrons and rockets.

Also your PRNG sequence is wrong. A rocket impact will need to do at least 3 PRNG calls on a direct hit per rocket (remember a rocket has two sets of damage, not one, so two sets of painchance). Start with the initial projectile damage, there is a monster painchance call inbetween the projectile damage and explosion (assuming a direct hit), it isn't guaranteed to do perfect damage depending on distance from center (and explosion damage doesn't apply the PRNG), and then another painchance for that damage.

I really like the RNG of berserk sometimes. It's funny when you hit an imp twice and it only hurts him a little, then the final blow is a 200 damage power hit that sends him flying off the catwalk and into the nukage ocean.

1 hour ago, 40oz said:

I really like the RNG of berserk sometimes. It's funny when you hit an imp twice and it only hurts him a little, then the final blow is a 200 damage power hit that sends him flying off the catwalk and into the nukage ocean.

*paf*

*paf*

"FALCON PUN-"

*imp goes flying*

I still wish there wasn't RNG, period.

9 hours ago, Edward850 said:

You miss quoted the wrong part

I only said 2 Rockets for Hell Knight and Arachnotron, nothing else.

I don't want to quote the whole thing, and yes, I admit that I didn't read the code, but the way you automatically assume I'm using a sourceport, and I'm using a sourceport that modify how PRNG works doesn't make me want to deal with your question. I wrote the second sentence because I don't have time to read the code at the moment, and I don't want to waste my time to test it out for you.

I like some RNG in the Doom games, but it does seem a little strange to have made such a high-risk move even more precarious in a way the player can't control. Especially with the iwads doling out Berserk Packs pretty conservatively.

20-200 is just too big a range. But oh well.

It makes more sense when you factor in that the fists normally do 2-20 ( or, more specifically, 1-10 * 2 damage ), and Berserk just slaps a x10 multiplier on it and that's all.

Probably didn't have much thought put into it, really, likely just slapped a huge damage increase onto it, tested it out, and didn't find it overpowered or too weak and just went with it.

I may be dumb, but I never use Berserk if I dont have to. 

I think berserk is generally more fun than chainsaw.

Turn fast weapons on and use your fist, hold the attack button down for comedy!

I never liked berserk. I always change back to the previous weapon I was using.

On 9/10/2017 at 3:23 PM, GarrettChan said:

I only said 2 Rockets for Hell Knight and Arachnotron, nothing else.

 

I don't want to quote the whole thing, and yes, I admit that I didn't read the code, but the way you automatically assume I'm using a sourceport, and I'm using a sourceport that modify how PRNG works doesn't make me want to deal with your question. I wrote the second sentence because I don't have time to read the code at the moment, and I don't want to waste my time to test it out for you.

As far as I can tell, Edward850 never made such an assumption. Seems to me that he simply stated some curious facts, in a pretty considerate way. I would think that you got questioned because what you said is simply not true unless you use a sourceport. There is only 1 PRNG in vanilla Doom that is used for AI.

So the fact that you got asked, and were given an easy way out, was a rather kind result, hardly deserving the "wasting my time" jazz, don't you think? Especially when someone "wasted their time" trying to delicately inform you. Not delicate enough, apparently...

1 hour ago, kb1 said:

Not delicate enough, apparently...

Sorry, my bad. I've taken my time to try in vanilla and I'm 100% sure that 2 Rockets can kill an Arachnotron.

I've taken some more time to try in vanilla on the question about whether 2 Rockets can kill a Hell Knight, even though it has 500 HP like an Arachnotron. The answer is yes, you can.

2_rocket_tron_kill.lmp

2_rocket_tron_kill.lmp (fixed link)

prboom-plus.exe -iwad DOOM2.WAD -warp 10 -complevel 2 -record 2_rocket_tron_kill.lmp

I have no idea what this argument is about.  seems like a classic case of two people arguing about completely different things.  but here's a demo of an arachnotron dieing in 2 rockets.

2 hours ago, NoisyVelvet said:

I have no idea what this argument is about.  seems like a classic case of two people arguing about completely different things.

Who's arguing? All I asked for was evidence/conformation (extraordinary claims requires extraordinary evidence, etc etc). That demo seems to be it.

7 hours ago, NoisyVelvet said:

2_rocket_tron_kill.lmp (fixed link)

Out of interest I printed the damage values.

First rocket did 140 direct + 116 splash = 256, leaving 244 (from initial 500).

Second did 140 direct (again) + 107 splash = 247, so 503 total, leaving -3 (dead).

You got two consecutive missile damage rolls of 7, out of a possible 8, multiplied by 20 for rockets.

On 10/09/2017 at 0:24 AM, Spectre01 said:

Is there an existing GZDoom mod that averages out all the damage values to make them always consistent? Seems like something interesting to try out and see how it feels.

Some time ago I tried doing this in my source port as an experiment. In particular I changed the single die roll with 20x multiplier to a sum of 20 rolls, to even things out. My conclusion was "revert this immediately, it's awful". I quickly realised the fun of the berserk is the chance that you can turn an imp inside out with your fists, or floor a demon in one hit. Making the damage consistent tending towards a normal distribution ensures this will almost never happen, every punch predictably lands 80-120 damage or so. It feels a lot weaker, and more boring. Of course you can increase the base value, but that just makes it boring and overpowered, not boring and underpowered.

It turns out the massive variation is the fun of the berserk fist. It is a "special" weapon you choose to use in specific favourable situations, when you feel like taking a bit of a gamble, giving the possibility, but not the guarantee, of spectacular results. A consistent predictable medium damage output rate seemed to take all the fun out of it, leaving it as just another workhorse weapon, like a shotgun with a very short range.