Heat Drain bionic power badness.

[Griffinhart]

Version 0.7-6-gab998b1.

The current description for CBM: Heat Drain doesn't mention that it consumes bionic power on every unarmed attack, regardless of success/failure.

That, or line 1198 in melee.cpp needs to be removed, so that Heat Drain doesn't consume bionic power every time the player makes an unarmed attack.

Or, as a third alternative, make Heat Drain an activated bionic, instead of a passive.

A fourth alternative would be to move line 1198 to somewhere between 1216 and 1226 (non-inclusive), so that the power consumption only occurs when Heat Drain succeeds.

As it currently is, Heat Drain is... very much not a good bionic. Every unarmed attack costs 1 power, but only has a 50% chance of restoring somewhere between 0 and 4 power (if my reading of the rng(val1, val2) function is correct), for a net gain of -1 to 3 power. That means the player has a 60% chance (50% from the drain not occurring at all; and 20% of the remaining 50% has a net gain of -1) of losing power with every unarmed attack.

Further the description of Heat Drain seems to indicate that it only deals extra damage on a successful drain, and according to melee.cpp, it doesn't even do that - it just slows the successfully-drained enemy down slightly. (I may be missing Heat Drain's effects elsewhere, though - I'm judging based on lines 1189 through 1227.)

[ejseto]

The expectation value for net energy change is 0 so I wouldn't call it a bad bionic. If there were drawbacks to installing bionics, then you'd be quite right. As is it, it's basically just 50% chance of sapping 80-120 moves from the target, which isn't really that bad considering how some players stack tailored tank tops and underarmor to get their melee cost down to like 25. Assuming you do that and can attack at least twice per turn, you can basically stunlock your target, bad luck notwithstanding.

While it seems like the energy cost shouldn't be assessed unless the bionic actually procs, I get the feeling it was actually intended this way (expectation value of 0 is kinda convenient don't you think?) as it's not supposed to be a power source.

Compare this with the shock bionic which does 2-15 damage and saps 360-900 moves but costs 2 energy. Obviously this is quite superior, but has a non-zero cost, and it can't be deactivated. Shocking a zombie that has speed 80 pretty much guarantees its death. Shocking it twice is overkill.

[Griffinhart]

"While it seems like the energy cost shouldn't be assessed unless the bionic actually procs, I get the feeling it was actually intended this way (expectation value of 0 is kinda convenient don't you think?) as it's not supposed to be a power source."

Then why not remove both the cost and the drain? Congrats, you've still got an expectation value of 0, and the bionic still retains its slowdown effect (although, again, the description would need to change, since that's not any sort of "extra damage").

"Compare this with the shock bionic which does 2-15 damage and saps 360-900 moves but costs 2 energy."

It's also up-front about its cost (and also why I don't install it; I dislike bionics that make basic actions, such as unarmed attacks, have a permanent cost with no way of deactivating the bionic).

[i2amroy]

I think it might be a good idea to set it up as a "toggle" type bionic rather then a passive one. Soemthing you can turn on (to drain with every melee attack) or turn off at will.

[ejseto]

Who knows, it's a roguelike, people love their rng. How is the heat drain bionic not upfront about its cost? It has, statistically speaking, zero cost.

[Griffinhart]

No; the wording in its description makes it sound like it acts like a power source (i.e. that you will in fact have net positive power gain).

Statistically speaking it does have net zero cost, but on a per-attack case, the player stands to lose more often than they gain (especially if they're able to reliably one-attack kill).

[ejseto]

The description should probably just be changed. I don't think, based on the fact that it's not grouped with power sources, that it was intended to have net positive power gain. It's still, almost strictly speaking, a purely beneficial bionic. Yeah, you might get unlucky, but this is a roguelike. You might also get really lucky and gain 3 energy over and over due to dumb luck. Statistically there are no drawbacks, only benefits.

And of course the player stands to lose more often than they gain, because the gains are often significantly more than the losses when they occur. That doesn't change anything.

[Griffinhart]

It's actively bad, in my play experience. With a character that's experienced in Unarmed (21 Unarmed skill) and a decent martial art (e.g. Judo), I had a net loss of 40 power (170, down from 210) after trashing a school full of children.

Also, "this is a roguelike" is hardly a justified excuse. "This is a bad roguelike", sure. But I'm assuming the point of continued development is to make CDDA less bad, not more.

[ejseto]

Well you must have bad luck since the expectation value of energy change is zero. That doesn't make it a bad bionic. Outside of luck, it's strictly beneficial. It's possible the pseudo rng is skewed, but I find it hard to believe losing 40 power from something that should have zero average effect is a normal thing. I really don't see your problem. You got unlucky. It happens. It's hardly representative of the whole. This isn't something where you play competitively and randomness voids skill. Roguelikes are meant to be unfair at times. That's hardly an indicator of being a "bad" game.

Mathematically, it has zero average cost and has the potential to stunlock anything of average speed. I'd call that an amazing bionic. Your one bad experience is hardly proof of the contrary.

[KA101]

1) Characters who are reasonably dressed (say: belt rig, army pants, trenchcoat, sneakers, headgear, everything fits/is reinforced if that matters) seem like a good baseline for attack timing. Balancing CBMs with the expectation that their user is expected to know about and use this negative encumbrance-stacking stuff (which I believe @i2amroy will be fixing soon) seems like a bad idea.

2) High-level Unarmed folks don't get many attacks in, because they don't need them. If installing this gives me a 60% chance per Unarmed attack of losing bionic charge, and I only need one attack per zed (Long Talons, anyone?) I'd be hard-pressed not to classify it as malware and seek to remove it from the character.

[ejseto]

It's not really a matter of balancing, that was just an ideal-case scenario. I was just trying to point out how good it can be, and how, statistically, it has no cost for this benefit.

I don't see how it matters whether it takes 1 attack or many to kill a monster. You still have the same probability per attack of gaining or losing energy, and the expectation value isn't changed. It's not like that's the last attack you'll make either, there's always more zombies. #2127 doesn't change that btw, the expectation value of energy change is still zero since the payoff has been tightened as well. If being able to toggle something that is strictly beneficial within the bounds of probability will satisfy people though, so be it.

[Griffinhart]

"I don't see how it matters whether it takes 1 attack or many to kill a monster."

Because if it takes only one hit to kill something, then the speed penalty that a successful Heat Drain proc would inflict is meaningless; so now it's a strictly useless bionic, since it has effective cost 0, and I still posit that it's a bad bionic because currently bionics are permanent installations and there's a softcap of 30 bionics (see issue #2003).

[ejseto]

That "softcap" is an easily fixed, unrelated issue. Someone can just make the letters wrap around to capital letters. And something that is sometimes beneficial and sometimes neutral is still on the whole beneficial. The expectation value of utility is still positive unless you're so strong you kill everything, 100% of all monsters, in 1 hit.