Power Attack - double weapon dice, or double ability mod bonus?

[Greywander]

I'm considering some armor tweaks that would make flat damage reduction a lot more prominent, and the idea occurred to me to perhaps balance it out by giving the option to trade out extra attacks for stronger attacks. So you can basically choose between doing a flurry of regular strength attacks, or do a single stronger attack. If, like the fighter, you get more than one extra attack, you can do a combo, or pile them all into a super attack.

For the implementation, I was thinking that by expending one of your extra attacks, you could double either the weapon dice or the ability modifier bonus of the one attack. If you have multiple extra attacks, then expending two would triple, three would quadruple, etc., so it stacks linearly.

The idea is that it ends up being weaker than just making several normal attacks against a target with no damage reduction, but against a target with high damage reduction it might absorb all or nearly all the damage from a regular attack. A power attack would then allow you to deal more damage in a single strike, and thus get more damage past the flat damage reduction.

So, to double the dice or double the modifier bonus? Doubling the dice makes it less dependent on stats, so even with a low STR or DEX you'll still do nearly full damage. It also makes crits stronger, and it greatly favors magical weapons that add additional damage dice (e.g. Flame Tongue). Doubling the stat bonus makes it more dependent on stats, and stat modifiers also tend to be more consistent (you can almost always expect it to be in the +3 to +5 range).

Doubling the damage dice seems like it could be a lot stronger, particularly with magic weapons, but since a power attack is replacing an extra attack then that seems fair. By making it less reliant on stats it becomes more accessible to characters who may not have maxed out that stat instead of being nearly useless. It also means that less martially focused characters will tend to favor one big attack over many weaker attacks, which, coincidentally also fits in line with a new weapon type I was also thinking of including: single shot weapons that have more damage dice but take a full action to reload.

[ReallySeamus]

I'm considering some armor tweaks that would make flat damage reduction a lot more prominent, and the idea occurred to me to perhaps balance it out by giving the option to trade out extra attacks for stronger attacks. So you can basically choose between doing a flurry of regular strength attacks, or do a single stronger attack. If, like the fighter, you get more than one extra attack, you can do a combo, or pile them all into a super attack.

For the implementation, I was thinking that by expending one of your extra attacks, you could double either the weapon dice or the ability modifier bonus of the one attack. If you have multiple extra attacks, then expending two would triple, three would quadruple, etc., so it stacks linearly.

The idea is that it ends up being weaker than just making several normal attacks against a target with no damage reduction, but against a target with high damage reduction it might absorb all or nearly all the damage from a regular attack. A power attack would then allow you to deal more damage in a single strike, and thus get more damage past the flat damage reduction.

So, to double the dice or double the modifier bonus? Doubling the dice makes it less dependent on stats, so even with a low STR or DEX you'll still do nearly full damage. It also makes crits stronger, and it greatly favors magical weapons that add additional damage dice (e.g. Flame Tongue). Doubling the stat bonus makes it more dependent on stats, and stat modifiers also tend to be more consistent (you can almost always expect it to be in the +3 to +5 range).

Doubling the damage dice seems like it could be a lot stronger, particularly with magic weapons, but since a power attack is replacing an extra attack then that seems fair. By making it less reliant on stats it becomes more accessible to characters who may not have maxed out that stat instead of being nearly useless. It also means that less martially focused characters will tend to favor one big attack over many weaker attacks, which, coincidentally also fits in line with a new weapon type I was also thinking of including: single shot weapons that have more damage dice but take a full action to reload.

Interesting! Regarding your specific question, a lot would depend on exactly what sort of damage reduction system this is intended to counter, but I think double the damage dice would likely be overpowered. It would turn every strong attack hit into a crit, and the way AC and to-hit numbers scale, I think it would quickly reach a point where the optimal strategy would be to do the strong attack every time. My suggestion would be that the strong attack allows you to add your proficiency bonus to damage. That would probably make a difference in terms of getting through damage resistance without being effectively a crit and without being ability score-dependent.

I like the concept of sacrificing something than your to-hit modifier in order to get a damage bonus. Trading out action economy benefits is an intriguing way to do it. Maybe it would make sense to have it eat your bonus action rather than use up an extra attack? The extra attack isn't even an option for several levels, and it seems like a tactical decision that should be available from day one. Using the bonus action would make the strong attack only be usable once per turn and most classes have options for the bonus action, so it would still require situational sacrifices and shrewdness. Another potential cost would be to have the strong attack be made at a disadvantage.

[Greywander]

Interesting! Regarding your specific question, a lot would depend on exactly what sort of damage reduction system this is intended to counter,

Still a WIP, but the basic idea is that armor has three stats: deflection, absorption, and coverage. Deflection is a flat AC that the armor gives you, but you can choose to use your DEX-based AC instead (even in full plate). Absorption is the amount of damage reduction you get, and coverage is how high you need to roll above the target's AC to bypass their damage reduction (basically attacking the gaps in their armor).

My tentative stats for a breastplate are 13 deflection, 8 absorption, and 4 coverage. So a breastplate provides a whopping 8 points of DR, but you only need to roll a 17 or higher to bypass that. A gambeson might have 11 deflection, 2 absorption, and 8 coverage. You can layer armor, though only one absorption can apply at a time. If you combine a breastplate with a gambeson, then you have an AC of 13, absorb 8 damage on an attack roll of 13-16, absorb 2 damage on a roll of 17-20, and take full damage on a roll of 21 or higher. Because coverage only cares about your final AC and not the deflection rating of that one specific piece of armor, wearing a breastplate (or having high DEX) does extend the range of attack rolls that a gambeson can absorb damage from.

There's a lot I'm still working out, and this will definitely require some playtesting to get the numbers right, but I'm thinking plate could go as high as 14 deflection, 10 absorption, and 10 coverage. Attacks from goblins would just bounce off, though a dragon would actually have a fair shot at bypassing the DR, and a dragon's damage would be high enough that the DR wouldn't fully negate it. I think something else to keep in mind is that monsters with more DR won't necessarily have less HP, and the same goes for players.

but I think double the damage dice would likely be overpowered.

That's what I thought at first. But if you just attack twice then you double both the damage dice and the flat modifiers. By power attacking, you're basically doing one attack at about 150% strength (for a non-magical weapon) instead of two attacks at 100% strength each. You won't notice the loss as much with something like a Flame Tongue, but with a non-magical weapon a power attack will feel significantly weaker than just making two attacks against a target with no damage reduction. Against a target with damage reduction, we need to make sure we can get enough damage past the damage reduction, or else our attacks might go from completely ineffective to little more than limp-wristed slaps. Like, technically we're doing damage, but we're still going to run out of HP before they do.

It would turn every strong attack hit into a crit, and the way AC and to-hit numbers scale, I think it would quickly reach a point where the optimal strategy would be to do the strong attack every time.

Are you sure you're not thinking of GWM/SS here? AC and to-hit numbers shouldn't have any effect on which is better. Well, except in the armor system above, where rolling high enough can bypass DR. But I hadn't told you about that when you wrote this, so for all you knew DR would always apply on every hit.

I just realized that if we double the damage dice, then the break-even point is when the DR exactly equals your ability modifier. At exactly that point, you'll do two sets of damage dice worth of damage whether you use extra attack or power attack. If DR is higher than your ability modifier, then power attacking is better. That's... actually really easy to remember, though you shouldn't expect the DM to tell you what the enemy's DR is (though if they're wearing armor you can infer it from that).

My suggestion would be that the strong attack allows you to add your proficiency bonus to damage. That would probably make a difference in terms of getting through damage resistance without being effectively a crit and without being ability score-dependent.

Hmm... A d8 or d10 is about the same as an ability mod of +5, which is about the same as a proficiency bonus of +6, so all of these are more or less interchangeable. The difference is that proficiency bonus starts lower and ends higher, while the die remains static throughout, and ability mod is between the two. So I don't think you'd see much of a difference on a normal weapon, it's only when magic weapons like the Flame Tongue come into play that it would make a difference. So I suppose the question is: should the Flame Tongue do even more damage, or should the bonus be the same regardless of the weapon? It would also make a difference for crits, too.

An example of the single-shot weapons I was considering was making a basic crossbow deal 3d6 + DEX mod, while requiring an action to reload. This is slightly stronger than attacking twice, assuming you don't have extra attack and have a below average DEX mod. If you have extra attack, then it's better to just use a bow that deals 1d6 + DEX mod damage and make two attacks on each turn. But if a power attack doubles damage dice then the crossbow could attack for 6d6 + DEX mod damage, and have you reload the following turn. This comes out to slightly worse than making four attacks across two turns with the bow, and even worse if you have high DEX. But, the crossbow seems to benefit more from a power attack than the bow would, which seems like a desirable outcome (it gives each weapon a slightly different niche).

I'm still inclined to double the dice, but adding proficiency instead is definitely something I'd like to investigate. Sometimes things don't go the way we expect them to during actual play.

Maybe it would make sense to have it eat your bonus action rather than use up an extra attack? The extra attack isn't even an option for several levels, and it seems like a tactical decision that should be available from day one.

Why not both? Or rather, either/or? Before getting Extra Attack, spending a BA is one way to still get a power attack. After getting Extra Attack, you have more options for how you want to do it.

I've been working off an on on a big overhaul, and one of the things I'll be experimenting with is giving the players more bonus actions at higher levels, with the caveat that each BA can only be used once per turn. So having lots of things to spend BAs on is a good thing. My one worry is that this would be too optimal of a use of a bonus action, to the point that you'd never go a turn without using a BA this way.

Another potential cost would be to have the strong attack be made at a disadvantage.

I mean yeah but then we're basically back to having an alternative version of GWM/SS. For that, I still like the idea of just not adding proficiency to the attack roll, then adding double proficiency to the damage roll.

[ReallySeamus]

That's an interesting armor system idea. It's probably too in-the-weeds for my table, but I like the layers to it and could definitely see myself adapting something similar - the idea of basically have "tiers" of AC success is really engaging.

The reason I think that the AC/to-hit factors are relevant is that there's a difference in value between 1 attack with some chance of hitting with double damage and 2 attacks each with the same likelihood of hitting for normal damage. It changes the percentage likelihood of each outcome and therefore could change the calculus about which is more valuable. For example, let's say you're using a greataxe:

• If a natural 11 or higher hits, on a strong attack: miss entirely 50% of the time; deal 2d12 (13) 50% of the time.
• If a natural 11 or higher hits, normal two attacks: miss entirely 25% of the time; deal 1d12 (6.5) 50% of the time; deals 2d12 (13) 25% of the time.
• If a natural 13 or higher hits, on a strong attack: miss entirely 60% of the time; deal 2d12 (13) 40% of the time.
• If a natural 13 or higher hits, normal two attacks: miss entirely 36% of the time; deal 1d12 (6.5) 48% of the time; deals 2d12 (13) 16% of the time.

Okay, when I run those numbers through, I think that actually seems more reasonable than I initially thought. Based on the system you've described though, where damage resistance can be pretty high and an individual strong attack isn't less likely to hit than a normal attack, my instinct is that it's likely that players would be incentivized to make this the default option, but that's not necessarily a bad thing. Tying this strong attack to the proficiency bonus makes sense to me thematically with the idea being that the character is drawing on their experience and skill to overcome an armor system specifically designed to stop them, but I now agree that doubling damage dice could work.

I do wonder if making this sort of thing a limited resource might be a sensible "cap" that lessens the impact of any slight imbalances, like you can use it a number of times equal to your proficiency bonus per long rest to emphasize the idea that it's a particularly challenging activity.

You've clearly got a lot of awesome ideas - I hope you don't mind if I hijack some of them when I'm tinkering at my table. :)

[Greywander]

It's probably too in-the-weeds for my table, but I like the layers to it and could definitely see myself adapting something similar - the idea of basically have "tiers" of AC success is really engaging.

As they say, different strokes for different folks. There is no universal system, set of house rules, or homebrew that is the best for every table. Also, my goals are a bit... eclectic. Honestly, I should probably just abandon D&D entirely and learn a new system, or make my own. But when you have a really cool idea for a project, sometimes your brain won't let it go until you make it.

The reason I think that the AC/to-hit factors are relevant is that there's a difference in value between 1 attack with some chance of hitting with double damage and 2 attacks each with the same likelihood of hitting for normal damage.

The difference is only in perception. Two attacks that deal 10 damage each and have a 50% chance to hit will deal an average of 10 damage per turn over the course of many turns. One attack that deals 20 damage with a 50% chance to hit also deals an average of 10 damage per turn over many turns. The overall expected damage output is identical.

Now, what does change is reliability. The one strong attack will never deal 10 damage, even though that's the average. It's an all or nothing deal; you either do the full 20 damage, or no damage. This makes it more swingy. The two weaker attacks are more reliable, as you'll deal at least the expected damage 75% of the time. That's something you can plan around.

And that's if the strong attack is straight up double damage. Since I'm only suggesting to double the dice and not the flat bonuses, it's more like a 50% damage increase.

Okay, when I run those numbers through, I think that actually seems more reasonable than I initially thought.

I know, right? When I first considered doubling the dice I thought it might be too strong, but the more I thought about it the more I realized how comparable it was to doubling the ability mod or adding proficiency.

Tying this strong attack to the proficiency bonus makes sense to me thematically with the idea being that the character is drawing on their experience and skill to overcome an armor system specifically designed to stop them, but I now agree that doubling damage dice could work.

Yeah, I think it's mostly down to how we want it to interact with other mechanics. Crits, low ability scores vs. high ability scores, weapons and features that add extra damage dice, etc. Proficiency bonus would be mostly unaffected by any of these, making it consistent, but I'm not sure if that's a good thing. It makes power attacking a bit stronger on weaker weapons, e.g. daggers, and a lot weaker on weapons with extra damage, e.g. a Flame Tongue greatsword.

You've clearly got a lot of awesome ideas - I hope you don't mind if I hijack some of them when I'm tinkering at my table. :)

Feel free, I'm happy to share. At least then some of it might see actual play. ;)

[Yakk]

Double dice.

This makes bigger weapons power attack better naturally. A 1d4 dagger power attack will be weaker than a 1d12 greataxe one in a nice way. And rollling more dice is also fun.

Note you can restrict it to double weapon damage dice, and not include stuff like flame tongue or sneak attack dice.

...

One thing I've looked into tweaking is making this mandatory; you can only make one hit on a foe per turn. You can throw more attacks at them, but that just adds weapon damage dice.

That change lets you be much more generous with extra attacks. It also makes heavy weapons better for single target foes.

So, imagine a TWF. They have 3 attacks, each for 1d6+5 damage. If they focus on one target, they can do up to 3d6+5 damage (15.5).

The greataxe user who focuses on one target does 2d12+5 (18).

If I limit it to hits (ie, once you have hit a foe, you can add more damage dice from attacks), then twf is also more accurate.

[Herbert_W]

How do you want making a power attack to feel?

It seems that you're thinking about power attack as a purely mechanical function: something that players do when they see that it's the optimal decision, and don't when they see that it isn't. It's a tradeoff between comparables that's either worth it or not. It's a knowledge (or guessing) game where you infer your enemy's absorption and deal more average damage if you can get it at least roughly right.

That's not necessarily a bad thing, but it can be. Some players like that sot of decision. Some players enjoyed pulling out a spreadsheet to see exactly how many points they should out into 3e's power attack (which traded accuracy for damage on a linear basis) - and others did not.

This stands in contrast with choices between incompariables; i.e. things which are both good (or bad) but which can't be measured on the same scale. For example, a barbarian's Reckless Attack enables a barbarian to deal extra damage at the risk of taking more damage. It decreases the party's chances of loosing while increasing the individual player's chances of being downed. This is a roleplaying choice as well as a tactical decision.

The two options that you've presented here would both make power attack into a mechanical thing through which players can express system mastery. Maybe that's what you want, but also maybe you might want to have this be a roleplaying choice too - in which case there's a much wider range of options to consider.

The difference is only in perception . . . Now, what does change is reliability . . .

I'm nitpicking, but there is another way that there's a significant tactical difference between the two: when a monster is at low HP, putting it down before the end of your turn is possible and much more valuable than letting it take a turn with a few HP. You do want to deny that monster one last turn's worth of actions, right? If a monster's health is very low such that a single weak attack would fell them, making two weak attacks is superior as it gives you two chances to hit. If a monster needs either two weak attacks or one strong attack to go down, then a strong attack is the superior option - it deals the same average damage, but it doesn't carry as much risk of leaving a monster alive with a sliver of health.

[Yakk]

Here are the mechanics for "automatic power attack"
Extra Attack: When you take the attack action you can make a 2nd attack on a creature you haven't hit yet this turn, or you can add a set of weapon damage dice from a weapon you are holding to a creature you already hit.

Two weapon fighting: If you have made an attack with a light weapon held in one hand, you can make an extra attack (see the extra attack feature) with a light weapon held in the other hand. Just like extra attack, if it is on a creature you have already hit you roll your weapon damage dice and add it to the damage total. If it is on a creature you have not hit, you make an attack the usual way.

...

A 2 attack TWF (1d6 weapons) with 18 attack stat and +3 proficiency on AC 16, single target.
Hit on a 9+, crit on a 20.

First attack could hit (60%) with 5% of that crits, or miss (40%). If it hits, you deal 3d6+4 (14.5) damage. If it crits, you deal 6d6+4 (extra 10.5) damage.
If it misses, second attack could hit or crit (60%/5%). If it hits, 2d6+4 (11), crits (extra 7).
If 2nd misses, 3rd attack is 60% 1d6+4 (7.5), 5% extra 1d6 (3.5).

.6 * 14.5 + .05 * 10.5 + .4 * ( .6 * 11 + .05 * 7 + .4 * (.6 * 7.5 + .05 * 3.5) = 12.8 damage per round.

A 2 attack greatsword user (2d6 weapon) with same stats.
60% 4d6+4 5% extra 14
40% of (60% 2d6+4 5% extra 7)
.6 * 18 + .05 * 14 + .4*(.6 * 11 + .05 * 7) = 14.3 damage per round

How about at 4 attacks using a +3 weapon? 6 proficiency, 22 stat (+6) against AC 22, 19-20 crit range (for the fun of it).

Hits on a 7+, crits on a 19-20
70% hit rate, 10% of which are crits, 30% miss rate.
Static damage is +9.

TWF is 70% 5d6+9 10% extra 5d6, then 30% (70% 4d6+9, 10% extra 4d6 then 30%(...))
THF is 70% 8d6+9 10% extra 8d6, then 30%( 70% 6d6+9, 10% extra 6d6, then 30%(...))

Static damage is .7 + .3*(.7+.3(.7+.3*(.7 + .3 * .7) = 99.7% chance for twf, and 99.2% for thf. I'll call that 100%.

So only working out dice:
.8*5 + .3*(.8*4 + .3*(.8 * 3 + .3*(.8 * 2 + .3*(.8*1)))) = 5.22 dice on average for TWF (each of which are 1d6)
.8*4 + .3*(.8*3 + .3*(.8 * 2 + .3*(.8 * 1 ))) = 4.1 dice on average for THF (each of which are 2d6)

THF does 4.1*7 + 9 = 37.7 average damage
TWF does 5.2 *3.5 + 9 = 27.2 average damage

In comparison, raw multiple attacks does 11.9 (THF) * 4 = 47.6, while TWF does 9.1 * 5 = 45.5. This makes two weapon fighting comparatively worse, because two weapon fighting relies on static damage more than the dice to deliver its damage.

[Deepbluediver]

For the implementation, I was thinking that by expending one of your extra attacks, you could double either the weapon dice or the ability modifier bonus of the one attack. If you have multiple extra attacks, then expending two would triple, three would quadruple, etc., so it stacks linearly.

The idea is that it ends up being weaker than just making several normal attacks against a target with no damage reduction, but against a target with high damage reduction it might absorb all or nearly all the damage from a regular attack. A power attack would then allow you to deal more damage in a single strike, and thus get more damage past the flat damage reduction.

That is actually a really interesting way to approach things- I'm intrigued. Especially since I was never that fond of the basic power-attack in the first place. I also like how this helps address the issue of critical-misses/fails (if used) penalizing high-level and multi-attack based builds.

If I had any reservations, it would be allowing you sacrifice multiple attacks to ramp up the damage to x3 or x4 levels. For the same reason I don't allow crit-hits to go above x3 in my homebrewed setting, I worry that damage-spikes like this can make gameplay unfun in both directions. On the one hand, yes randomness is part of the game. On the other hand though, having your BBEG fantastic-encounter that you've been building up to for several sessions get bounced in a round-and-a-half, or having the character you've invested time, effort, and emotion into get snackered without you even being able to respond is not fun (IMO). I guess it's not that much different than the kind of damage that you could do if you hit with all 3 or 4 of your normal attacks anyway, but something about doing it this way just makes me hesitant.

One final thought- how do you feel about this same mechanic as applied to accuracy? Basically, sacrifice an attack, get something like Advantage (roll two d20, keep the highest) or double your Ability bonus on the attack roll? And I'm split if you should be allowed to combine it with the damage boost.
Since Power Attack is already situational, maybe make it an either/or option kinda thing. Just brainstorming a little.


P.S. To answer the original question, I favor doubling damage-dice over the ability score, for damage. Something about rolling whole handfulls of dice is just FUN for me. So I would prefer going from 1d10+5 to 2d10+5 instead of to 1d10+10. That's just my personal take on it though- it does change the damage-curve to a different shape instead of just scaling it higher and some people might not like that.

[PhantomSoul]

If I had any reservations, it would be allowing you sacrifice multiple attacks to ramp up the damage to x3 or x4 levels. For the same reason I don't allow crit-hits to go above x3 in my homebrewed setting, I worry that damage-spikes like this can make gameplay unfun in both directions. On the one hand, yes randomness is part of the game. On the other hand though, having your BBEG fantastic-encounter that you've been building up to for several sessions get bounced in a round-and-a-half, or having the character you've invested time, effort, and emotion into get snackered without you even being able to respond is not fun (IMO). I guess it's not that much different than the kind of damage that you could do if you hit with all 3 or 4 of your normal attacks anyway, but something about doing it this way just makes me hesitant.

Agreed -- at the very least there needs to be diminishing returns, but given potential interactions with other things (e.g. Portent), letting things get too wild is a big risk... and even without those it can diminish the fun for some contexts (even if doing it on a difficult-but-not-highly-strategic-or-puzzle-like random encounter or a mook would likely be fine)

One final thought- how do you feel about this same mechanic as applied to accuracy? Basically, sacrifice an attack, get something like Advantage (roll two d20, keep the highest) or double your Ability bonus, etc? And I'm split if you should be allowed to combine it with the damage boost.
Since Power Attack is already situational, maybe make it an either/or option kinda thing. Just brainstorming a little.

I'd definitely make it either/or, just because those feel like different goals (big hit vs. reliable hit).

I've toyed with the idea of letting (some/all) Fighters use 2d10 instead of 1d20 (their choice, declared before the roll) to give a reliable hit option that isn't also crit fishing (maybe barring some subclasses where 10 on either die, for example, could be a crit).

P.S. To answer the original question, I favor doubling damage-dice over the ability score, for damage. Something about rolling whole handfulls of dice is just FUN for me. So I would prefer going from 1d10+5 to 2d10+5 instead of to 1d10+10. That's just my personal take on it though- it does change the damage-curve to a different shape instead of just scaling it higher and some people might not like that.

Agreed, it's just more enjoyable -- plus it isn't then just a bonus for better stats (and a +5 is already roughly the average of 1d10, which is the high end for weapons). Given higher-damage dice favour better/martial weapon proficiencies and strengthier weapons overall, there's also that as a motivation for preferring using the dice for me.

[Deepbluediver]

Agreed, it's just more enjoyable -- plus it isn't then just a bonus for better stats (and a +5 is already roughly the average of 1d10, which is the high end for weapons). Given higher-damage dice favour better/martial weapon proficiencies and strengthier weapons overall, there's also that as a motivation for preferring using the dice for me.

Good point- I hadn't even thought of that. Helping to emphasize better weapon proficiencies over simply pumping your stats ever-higher.

[Yakk]

The only reason why you'd double stat is that "power attack is strength based", ie, "more strength more power".

But it feels better for heavier weapons to hit harder when you power attack.

[Greywander]

Double dice.

This makes bigger weapons power attack better naturally. A 1d4 dagger power attack will be weaker than a 1d12 greataxe one in a nice way. And rollling more dice is also fun.

An excellent point, I think this is the way to go.

One thing I've looked into tweaking is making this mandatory; you can only make one hit on a foe per turn. You can throw more attacks at them, but that just adds weapon damage dice.

An interesting idea. I think this makes a lot of sense if you're not using damage reduction. Focus fire is already optimal, and this would make it less clear cut whether it's better to focus on one enemy or spread your attacks out. Since I am looking at using damage reduction, the dynamic would be a bit different. This is something I'll have to consider if I want to do both.

How do you want making a power attack to feel?

It seems that you're thinking about power attack as a purely mechanical function: something that players do when they see that it's the optimal decision, and don't when they see that it isn't.
[...]
Maybe that's what you want, but also maybe you might want to have this be a roleplaying choice too - in which case there's a much wider range of options to consider.

I don't know that D&D, or a derivative, is the proper system for roleplaying in combat. Combat is very crunchy; it is THE place where system mastery is expressed. The ability to freely refluff things so long as the mechanical effect is the same goes a long way toward allowing a certain amount of roleplay without compromising the mechanical gameplay.

You make a good point about incomparables. In my mind, these are things where the relative value changes depending on the situation, so it's next to impossible to compute the optimal choice ahead of time. In the moment, one might be clearly the better choice than the other, and you can form strategies around using one or the other or both, but the two produce completely different effects. For example, the Dash action can't really be compared to the Use an Object action. It's more difficult to do this with actions that produce similar effects, though. Whether it's better to cast Fireball or Disintegrate will depend somewhat on the situation, but both deal damage and in many cases either option will produce the same outcome.

I'm nitpicking, but there is another way that there's a significant tactical difference between the two: when a monster is at low HP...

Is that a bad thing, though? A big part of this is reading the tactical situation and deciding which option is the best tool for the job. If anything, it makes it more interesting, as otherwise once you figure out how much DR a monster has you can compute whether it is better to always power attack or not. This can make it better to switch once the enemy is low on HP, so the same option isn't always better all the time.

If I had any reservations, it would be allowing you sacrifice multiple attacks to ramp up the damage to x3 or x4 levels. For the same reason I don't allow crit-hits to go above x3 in my homebrewed setting, I worry that damage-spikes like this can make gameplay unfun in both directions.

It definitely gets more swingy. It's certainly a cause for concern. Would it be better to restrict it, or trust the players to make the best choice for themselves at the time? I'm not sure, but I tend to lean toward trusting the players to make their own choices.

One final thought- how do you feel about this same mechanic as applied to accuracy? Basically, sacrifice an attack, get something like Advantage (roll two d20, keep the highest) or double your Ability bonus on the attack roll? And I'm split if you should be allowed to combine it with the damage boost.
Since Power Attack is already situational, maybe make it an either/or option kinda thing. Just brainstorming a little.

I like it. And I do think you should be able to combine them, but that would require having two extra attacks. So it would only be an option for higher level characters. I'll have to look at the math, though.

I think I might have it roll an extra d20 and take the highest, making it stack with advantage, but with diminishing returns. It also increases crit chance, which makes sense. For disadvantage, on the other hand, you'd take the highest of the disadvantage roll and the raw d20 roll from the precision attack, making it actually slightly better than just attacking normally.

I've toyed with the idea of letting (some/all) Fighters use 2d10 instead of 1d20 (their choice, declared before the roll) to give a reliable hit option that isn't also crit fishing (maybe barring some subclasses where 10 on either die, for example, could be a crit).

I've played with a similar idea, where 2d10 is the standard roll for everything, but you have the option to be reckless and roll 1d20 or cautious and roll 3d6. A cautious roll tends towards a middling result, making it a good choice if your skill is high enough that an average roll succeeds. A reckless roll is equally likely to produce an extreme result as it is an average result, so it can at least give you a chance when you're outclassed. In the end, there didn't really seem to be a reason to ever roll normally, as it would always be more optimal to roll cautiously or recklessly, and the system would need to be designed with this in mind, whereas D&D is designed around d20 rolls.

[PhantomSoul]

I've played with a similar idea, where 2d10 is the standard roll for everything, but you have the option to be reckless and roll 1d20 or cautious and roll 3d6. A cautious roll tends towards a middling result, making it a good choice if your skill is high enough that an average roll succeeds. A reckless roll is equally likely to produce an extreme result as it is an average result, so it can at least give you a chance when you're outclassed. In the end, there didn't really seem to be a reason to ever roll normally, as it would always be more optimal to roll cautiously or recklessly, and the system would need to be designed with this in mind, whereas D&D is designed around d20 rolls.

Cool for comparison -- I definitely think that's a logical outcome from 1d20 being the middle of three options! I'd tend to just give one (I imagine 1d20 and 2d10 for simplicity) -- and I'd have that option be a specific Class Feature (with it being the Fighter for simplicity). I'd rather that (available at a lower level, probably 6-7 to avoid stacking small dip appeal) over Reliable Talent for Rogue Skills too, personally!

[Deepbluediver]

Sorry for not responding sooner- I had a lot going on and this basically fell off my radar. Anywho....

Is that a bad thing, though? A big part of this is reading the tactical situation and deciding which option is the best tool for the job. If anything, it makes it more interesting, as otherwise once you figure out how much DR a monster has you can compute whether it is better to always power attack or not. This can make it better to switch once the enemy is low on HP, so the same option isn't always better all the time.

Is there a way for players to tell just how much HP an enemy has left? Is it worth burning a Standard/Swift/Free Action to discern that? Would it be worth it if one character could convey that information to another? etc etc etc

One thing that I think is interesting is comparing different types or genres of games to one another to see how they highlight the comparisons, or how different mechanics contribute to the overall experience. For example, when I used to play World of Warcraft (WoW), some boss-monsters would " Enrage" after a certain time-limit, which meant that their damage increased dramatically and it would very rapidly lead to a raid-wipe/TPK. It was basically a DPS-check to make sure your group was appropriately geared for the encounter and to discourage a group from just stacking healers and slow-rolling every boss-fight to the death. I can definitely recall at least one instance where our group hit the "Enrage" timer with the boss at low health, and the raid leader called for all the tanks and healers to swap to damage- basically seeing if we could kill the boss before he wiped all of us (it worked, by the way).
But that was in an online MMORPG, where we could see the boss' HP in real-time and the game was more about reactions and positioning than strategy and resource-use. Could you replicate this in D&D? I'm not sure you can, and I'm not sure you really want to, either.

If it works for your group, that's good. I just encourage everyone to consider what sort of experience you are going for when you design mechanics.

It definitely gets more swingy. It's certainly a cause for concern. Would it be better to restrict it, or trust the players to make the best choice for themselves at the time? I'm not sure, but I tend to lean toward trusting the players to make their own choices.

Yes, but the GM is responsible for designing the world that the players operate in. As noted above, I think it's important to consider how players will approach this, including when/if they need to pick Feats for their build.

I've played with a similar idea, where 2d10 is the standard roll for everything, but you have the option to be reckless and roll 1d20 or cautious and roll 3d6. A cautious roll tends towards a middling result, making it a good choice if your skill is high enough that an average roll succeeds. A reckless roll is equally likely to produce an extreme result as it is an average result, so it can at least give you a chance when you're outclassed. In the end, there didn't really seem to be a reason to ever roll normally, as it would always be more optimal to roll cautiously or recklessly, and the system would need to be designed with this in mind, whereas D&D is designed around d20 rolls.

Yes, D&D tends to reward specialization. Or at least not emphasize taking the middle ground- a party full of Melee specialists and dedicated Casters tends to out-perform an entire party full of Gishes, etc etc etc. It doesn't mean you can't give players options, it just means you have to think about what the cost is to make it worth-while for players to take one path over another.
For example, suppose your odds for a roll are 3d6, 2d10+X, or 1d20. What does "X" have to be to get players to make that kind of roll instead of the "Safe" roll or the "Gamble" roll? That's essentially the cost of taking the middle ground.