So. I've been wrestling with soft attacks every so often, trying to get somewhere useful, and to be honest I think the most useful thing at this point is a quick summary and to move on. There are so many different things to consider, all of which play into other aspects of the game, that I can't usefully do much specific without actually creating a game from whole cloth. And that's more Dan's thing.
So here's the main conclusions I've reached.
The main thing, the biggie: how much granularity does the status system provide? It's not really about the attack, or the resilience, or the recovery; it's about whether you can be "blind" vs. "not blind", or whether you measure 10 different degrees of blindness.
Higher granularity means low swinginess, high scalability and a high tracking burden. Low granularity means high swinginess, low scalability and a low tracking burden. Anyone picking a system for soft attacks needs to decide which of those is the priority. The details of the system involved are really a secondary matter.
For a game without much in the way of levelling - the sort of thing where a new character is about as effective as a veteran - scalability is not an issue. In those cases, the decision might come down to other elements of the game. If it's supposed to be highly realistic, the cost of tracking might be acceptable - particularly as in such a game, quite a lot of tracking is probably already necessary: if you have to track damage to a number of specific locations, ammunition, energy expended, morale, fatigue and thirst, adding in blindness, sleepiness and stunnedness might seem perfectly reasonable. On the other hand, if it's a fast-paced game featuring "alive", "wounded" or "dead" statuses and very little else, nobody wants detailed tracking of blindness.
If the game has strong levelling, then scalability is important. Having low-level soft weapons that cripple high-level targets creates loopholes, cheap tactics and can undermine the setting. On the other hand, having high-level soft attacks that are often ineffective makes makes it not worth using them. There are ways around this other than granularity, though. If the attack rolls necessary scale with level, and the status effect isn't too severe, then a binary system may be acceptable.
Example: Scaly Monitors
In a new build of Monitors, skill is modelled by simple skill pluses linked to level. A modified roll of 11+ on a d20 is a success. Roj, a 5th-level iguana xenologist, has +5 Pistol and carries a tranq pistol to capture interesting fauna on the ice-world of Kraant, as well as a standard-issue blaster for self-defence. When his research group are confronted by a 5th-level cryoboar with +5 Stamina (which subtracts from his roll) he has to roll an 11 or better to successfully stun it, halving its action rate, which will make it easier to capture, kill or escape from. On the other hand, he could turn to the trusty blaster, which should kill the boar in five or six shots.
A little later, Roj runs into a cryophant. Oops. With a Stamina of 10, he needs a 16 or better to stun it. A lucky shot would make life substantially easier, but attempting to beat the critter would still be very risky. Once aggro'd, the cryophant's hit points mean it'll survive to reach combat sooner or later, at which point Roj may be wishing he'd gone straight for the blaster rather than waste turns trying to drug it, since he'd have to endure fewer rounds of punishing combat.
D&D seems to aim for this with its scaling attack rolls, defensive stats and DCs. However, the major status effects in D&D (stun, blind, slow and paralyse) are very powerful, and this means makes it viable to focus on stun-locking, or to spam soft attacks at a boss until once gets through, followed by unloading the party's most damaging attacks while it's vulnerable. While those are perfectly valid ways of playing, it can become repetitive and feel cheesy - and designers and DMs alike tend to compensate with immunities and very high defences.
Swinginess is amplified against small numbers of relatively powerful targets, since these are typically highly resistant to compensate for their numbers. With only "success" or "failure" results, a single soft attack can make the difference between a difficult battle and a cakewalk (depending, of course, how powerful the status effect is).
Penalties can vary in severity, in duration or both. The most and least severe penalties inflicted by a soft attack are a significant factor. Soft attacks with a weak maximum penalty can be allowed a high success rate without risking 'ruining' boss fights. Soft attacks with a strong maximum penalty need to rarely impose that penalty on powerful enemies, otherwise they become disparately useful. The duration complicates matters further: a minor penalty that lasts for a long time has limited effect, but may be a pain to remember, depending how the system models it. A severe penalty that also lasts a long time is more swingy as an individual attack, since a single good result can cripple the target for the duration of a fight; however, a severe penalty that's short-lived can be very powerful if it comes at just the right time, and otherwise has limited effect, so it's also swingy. There's probably not much to choose between 'em, to be honest.
Soft attacks should never be crippling, either by themselves or through synergystic effects. This, as Dan mentioned, basically allows them to circumvent the standard combat system. The only situation where I might be inclined to favour this would be niche games, where hard-attack combat isn't what you're supposed to be doing.
Example: All's Fair
For example, I can visualise a game about fey interlopers. You have a variety of supernatural abilities that allow you to beguile, bemuse, bewilder and bewitch NPCs that come between you and your goals. Striking people blind, sending them to sleep or rooting them to the spot fit perfectly well with fey folklore, rather better than hacking your way through hordes of guards. The smooth way to play the game is to slip gently through the NPCs you encounter, eliminating them with tactically-applied magic. Actual combat is a fall-back if you mess up, and something to be avoided. The tactical challenge isn't whittling down individual opponents, it's dealing with the situation as a whole, picking the right spells to use in a situation, and avoiding drawing down a whole crowd of enemies on your head at once. The only point where drawn-out combat occurs might be dealing with other magic-using entities, where arcane duels might take place.
For a lot of games, a highly granular soft attack system is not going to be appropriate, despite its advantages.
- As we agreed before, a game with highly abstracted combat doesn't want very granular soft attacks, because it contradicts other aspects of the combat system.
- For games that try to minimise the tracking burden, perhaps to create a streamlined and accessible system or to speed up play, granular soft attacks are also inappropriate.
- For games that aren't especially concerned about "realism" or "fairness", the granularity may simply not be a priority. If the setting is full of randomness and arbitrariness, with luck and the whims of the powerful playing a significant part, then it doesn't necessarily matter whether blind spells are swingy, since just about everything else is too. Sometimes you get lucky, sometimes you die.
For most games, though, I'm inclined to feel that at least a three-way distinction is useful, with "semi-successful" sitting between "effective" and "entirely ineffective". This would reduce swinginess by allowing soft attacks to have limited effects on powerful creatures, without either rendering them useless or allowing them to stun-lock the big bads. Exactly how the result would be established would depend on the system as a whole.
The other thing is that any soft attack system depends on the frequency of soft attacks in the game. If only rare equipment can blind, stun or paralyse creatures, a relatively high-maintenance resolution system isn't too problematic if it gives pleasing results. If they're going to come up in every fight, though, the smoothness of play is more important.
And that's probably it from me, to be honest. I had thought of scribbling a bit more about recovery systems, but I'm not sure there's much point. So a rather desultory end to a rather confused project.