The Savage Attacker feat is underpowered

[BigONotation]

I've houseruled it to be once per turn one damage die of a melee attack can be auto-max roll. In my opinion the existing feat should include +1 Str or be rewritten.

[Easy_Lee]

You can use http://rumkin.com/reference/dnd/diestats.php to calculate some statistics. I checked it out with 1d12 (2d12 drop lowest) to see how much damage it would add on a typical turn. The average damage on a D12 is 6.5. Rolling twice and dropping the lowest brings it up to 8.49. So that's a two damage increase per turn, overall. It doesn't scale up with number of attacks, either.

So yeah, pretty lame feat. Even worse, it makes you roll more dice and take longer every time you make an attack. Nobody at the table is going to appreciate you taking this feat. I don't think adding 1 damage will help it, I think it just needs to be removed.

I'd change it to let you roll a guaranteed 20 once per short/long rest. Now that would be a powerful feat option.

[Cambrian]

Keep in mind the feat's returns are not static:
The more attacks and the larger the variance of the dice used, the better the Feat is.

It is however mathematically a pain in the tuckus to calculate the expected returns and the mathematically optimum time to reroll an attack.

[BigONotation]

Keep in mind the feat's returns are not static:
The more attacks and the larger the variance of the dice used, the better the Feat is.

It is however mathematically a pain in the tuckus to calculate the expected returns and the mathematically optimum time to reroll an attack.
Even so I think my solution is elegant and makes the Greataxe much more attractive. It's currently much less effective than the Greatsword or Maul.

[Strill]

Keep in mind the feat's returns are not static:
The more attacks and the larger the variance of the dice used, the better the Feat is.

It is however mathematically a pain in the tuckus to calculate the expected returns and the mathematically optimum time to reroll an attack.

You're not reading it. Its returns absolutely ARE static. It's only once per turn. The number of attacks is irrelevant. Easy_Lee already calculated that it's +2 damage in the best case.


I've houseruled it to be once per turn one damage die of a melee attack can be auto-max roll. In my opinion the existing feat should include +1 Str or be rewritten.Is that really worth a feat? Keep in mind that +2 STR typically gives you around +20% overall DPR. I don't think that what you're suggesting would be worth it. Maybe if it were +1 STR AND you can maximize one melee die per turn, then it'd be on-par with +2 STR.

[Oscredwin]

You're not reading it. Its returns absolutely ARE static. It's only once per turn. The number of attacks is irrelevant. Easy_Lee already calculated that it's +2 damage in the best case.

With one attack on a Great Axe you have one chance to roll a 1 and get 5.5 bonus damage. With four attacks, you have 4X the chances of getting a larger bonus (or any bonus at all). Still not a great feat, but it does scale with number of attacks.

[Cambrian]

Even so I think my solution is elegant and makes the Greataxe much more attractive. It's currently much less effective than the Greatsword or Maul.Yeah the unreliability and low impact of the feat makes it more like a half feat than a full feat. Your suggestion is very reasonable and elegant compared to changing the actual mechanics of the feat.

[Rezby]

Since we're granted: upon rolling damage for a melee weapon attack, you can reroll the weapon's damage dice and use either total.

I'd argue it includes sneak attack, divine strike, and the X smite spell buffs, as well as hex or hunter's quarry. All dice used for damage caused by a weapon can be rerolled. It's a feat, with a single line of benefit. Yes, its pretty powerful, but you're taking a feat called ASSLOADS of damage*.

* Not actually called assloads of damage

[EugeneVoid]

Yeah, when you have advantage, you can roll damage for weapons twice and take better result would be situational, but potentially very strong.

[Strill]

Since we're granted: upon rolling damage for a melee weapon attack, you can reroll the weapon's damage dice and use either total.

I'd argue it includes sneak attack, divine strike, and the X smite spell buffs, as well as hex or hunter's quarry. All dice used for damage caused by a weapon can be rerolled. It's a feat, with a single line of benefit. Yes, its pretty powerful, but you're taking a feat called ASSLOADS of damage*.

* Not actually called assloads of damage

It specifically says the "weapon's" damage dice.

[Mechaviking]

So Rogues can use it when they melee somebody with their sneak attack, and that´s it?

[Strill]

So Rogues can use it when they melee somebody with their sneak attack, and that´s it?

It only applies to the weapon's damage dice. It wouldn't apply to sneak attack.

[Mechaviking]

It only applies to the weapon's damage dice. It wouldn't apply to sneak attack.

Yup, this feat is a waste of ink, paper and printing power.

[Rezby]

It specifically says the "weapon's" damage dice.

RAW vs RAI, my friend. RAW, the feat is nothing. +2 ish to damage. RAI, I think the feat is meant to be more along the lines of what I suggested as an interpretation.

[Sidmen]

Yup, this feat is a waste of ink, paper and printing power.
It's more than doubled one of my player's effective damage output. While some math may assess it at a +2 typical damage, I've seen it increase a result of 1 or 2 to 10-12 too many times to take your post seriously.

Of course, I have a player that consistently rolls either really low or really high - with nothing in between. (And yes, we roll in the open).

[Strill]

RAW vs RAI, my friend. RAW, the feat is nothing. +2 ish to damage. RAI, I think the feat is meant to be more along the lines of what I suggested as an interpretation.

Compare it to Great Weapon Fighting Style, which explicitly does not apply only to the weapon's damage dice.


It's more than doubled one of my player's effective damage output. While some math may assess it at a +2 typical damage, I've seen it increase a result of 1 or 2 to 10-12 too many times to take your post seriously.

Of course, I have a player that consistently rolls either really low or really high - with nothing in between. (And yes, we roll in the open).Congratulations, you have loaded dice? Saying that the feat is good because you're lucky is a non-argument.

[Sartharina]

... I loved this feat in the playtest.

Not sure why it has the 'once per turn' limitation, though, instead of 'once per attack', so that fighters could benefit from it more.

[Sidmen]

Congratulations, you have loaded dice? Saying that the feat is good because you're lucky is a non-argument.
Using averages when talking about a feat whose true utility is in compensating for incredibly bad luck is also a non-argument.

[Easy_Lee]

As an update to my post above, and to be fair, it's true that multiple attacks can increase the feat's damage. If you can figure out when to use the bonus (a damage threshold of X, perhaps), then you can potentially gain more than +2 average damage from it.

However, I still think it's a weak feat. By comparison, polearm master is a good feat, and it provides anywhere from 1d4 + attribute to 1d4 + attribute + 2 (duelist, quarterstaff) extra damage per round in addition to extra opportunity attacks. It takes your bonus and/or reaction, sure, but that's 7.5 to 9.5 extra damage per round, or more if you get the opportunity attack.

By comparison, let's look at Savage attacker. At a maximum of four attacks, and used perfectly, 4d12 with your greataxe becomes 5d12 drop lowest. That's 29.97 average damage, where 4d12 averages 26.00. So even with perfect usage, savage attacker only adds 4 damage average.

If you were able to reroll all of the damage from an attack, including things like sneak attack, smite, elemental damage, etc, then savage attacker would be worthwhile. As is, it's a purely offensive feat that benefits your character about as much as dual wielder (up to 4 damage vs 2 damage and 1 AC). At least dual wielder lets you draw two weapons at once.

[Sartharina]

You can choose to reroll the weapon dice after you already roll. Saying it's a "+2" is therefore inaccurate. It can range anywhere from a +0 to a +11, and given the nature of the feat, it's VERY context-sensitive.

[Strill]

Using averages when talking about a feat whose true utility is in compensating for incredibly bad luck is also a non-argument.What utility do you get from that? How does it help you kill enemies any more efficiently?

-----------------------

I think I might know why this feat is so weak. Apparently at some points during the playtest, instead of extra attacks, you added additional damage dice. Savage Attacker would've made a lot more sense in that context.

[Easy_Lee]

You can choose to reroll the weapon dice after you already roll. Saying it's a "+2" is therefore inaccurate. It can range anywhere from a +0 to a +11, and given the nature of the feat, it's VERY context-sensitive.

On the average, when used perfectly, it's between +2 and +4 damage, nothing else. Saying that it does more in specific circumstances is misleading. In specific circumstances, your bonus attack from polearm master duelist quarterstaff crits and does 2*(4+2+5) = 22 damage. That doesn't mean the feat adds 22 damage; the average benefit is what we're concerned with.

[BigONotation]

... I loved this feat in the playtest.

Not sure why it has the 'once per turn' limitation, though, instead of 'once per attack', so that fighters could benefit from it more.

Another viable solution, per attack would make it worthwhile.

[Easy_Lee]

Another viable solution, per attack would make it worthwhile.

True, but who wants to sit around while you roll all those dice?

[Demonic Spoon]

Using averages when talking about a feat whose true utility is in compensating for incredibly bad luck is also a non-argument.


The problem is that in any individual combat you're probably going to be making a lot of attacks. Your overall damage output in a combat is not hugely subject to random chance. There's very rarely some disastrous consequence that's going to come about because you rolled poorly on one particular damage roll.

I think that this feat just shouldn't exist. The feat is there to add +damage to your melee attacks...wrapping the +damage in an admittedly more interesting "reroll damage dice" mechanic doesn't change the fact that you take the feat to just straight up do more damage. That's boring and is never going to compete well with a +2 bonus to Strength - either the feat is powerful enough to obsolete the +2 strength, or it's weak enough that the +2 strength is generally outright better.

We have Great Weapon Master as the feat that makes you better at hitting things with big axes, and it does that in a cool and interesting way. I don't think Savage Attacker is a good feat, but I don't think it adds enough to the game to warrant houserule fixes. The example above of making the boost per attack is just going to obsolete the Strength ability boost, which is bad.

[Sartharina]

True, but who wants to sit around while you roll all those dice?It doesn't take that long - and if it's every attack, you do it at the same time.


I think I might know why this feat is so weak. Apparently at some points during the playtest, instead of extra attacks, you added additional damage dice. Savage Attacker would've made a lot more sense in that context.And we got feats at the same rate as in 3.5 at this point - but it was only reroll one set of damage dice. And there were no 2d6 weapons - just 1d10 and 1d12.

I'd rather have seen this feat and the Great Weapon Combat Style be merged into "Reroll your weapon dice, take higher".

[hymer]

never going to compete well with a +2 bonus to Strength

I generally agree, however this particular bit isn't all there is to it.
It's not always useful or possible to take +2 <damage ability>. Once it's maxed is the obvious thing, but there are also moon druids, to whom this is nearly the only option for increasing damage in wild shape, unless the table allows them to take half-feat ability modifiers along to wild shape.

[Sartharina]

I generally agree, however this particular bit isn't all there is to it.
It's not always useful or possible to take +2 <damage ability>. Once it's maxed is the obvious thing, but there are also moon druids, to whom this is nearly the only option for increasing damage in wild shape, unless the table allows them to take half-feat ability modifiers along to wild shape.
I think they lose the feat in Wild Shape.

[hymer]

I think they lose the feat in Wild Shape.

I'll be the first to admit that wild shape rules aren't exactly clear. But it seems feats in general should be kept:

"You retain the benefit of any features from your class [...] or other source and can use them if the new form is physically capable of doing so." PHB p. 67.
Seems to me you keep Savage Attacker from that.

[Ramshack]

I too have seen to many 1's and 2's rerolled into 10, 11 and 12s. Sure sometimes you get a 6 or 8 or whatever, but the feat really is quiet strong. Especially when used with the bigger die. Plus the player excitement when there like oh man i rolled terrible. 1 +4?? no wait reroll, yay 10+4! Stack this with the Great Weapon Fighting Style and you can reroll every 1 and 2 and still reroll another die.

[Demonic Spoon]

I too have seen to many 1's and 2's rerolled into 10, 11 and 12s. Sure sometimes you get a 6 or 8 or whatever, but the feat really is quiet strong. Especially when used with the bigger die. Plus the player excitement when there like oh man i rolled terrible. 1 +4?? no wait reroll, yay 10+4! Stack this with the Great Weapon Fighting Style and you can reroll every 1 and 2 and still reroll another die.


"I have seen X happen quite a lot" is a poor argument against math.

[Sartharina]

"I have seen X happen quite a lot" is a poor argument against math."But the math says X!" is a poor argument against actual experience.

[archaeo]

"I have seen X happen quite a lot" is a poor argument against math.


"But the math says X!" is a poor argument against actual experience.

Well, gosh, this is a pickle. It's almost like both of these poor arguments are accurate descriptions of a game where the RNG is tempered by a human game master!

[Strill]

Well, gosh, this is a pickle. It's almost like both of these poor arguments are accurate descriptions of a game where the RNG is tempered by a human game master!

In what situation could the GM be affecting the result of these dice rolls?

[Oscredwin]

It seems like this is a weak feat that will be beloved at some tables where no one would ever think to look up builds online and used to ban the 3.5 monk for being too powerful. It makes the game more fun. It's not so bad as to a be a trap option (like 3.5 toughness), and something had to balance polearm master to keep the average down.

[archaeo]

In what situation could the GM be affecting the result of these dice rolls?

Maybe it would be better to just say "tempered by human interaction." In the context of the present argument, even if Savage Attacker provides relatively modest average benefits, it can occasionally provide rather substantial bumps in power above and beyond the capability of numerous other damage-boosting abilities. It also comes with a huge psychological effect, providing a ton of insurance against bad luck, a quality that many will find desirable.

[Strill]

Maybe it would be better to just say "tempered by human interaction." In the context of the present argument, even if Savage Attacker provides relatively modest average benefits, it can occasionally provide rather substantial bumps in power above and beyond the capability of numerous other damage-boosting abilities. It also comes with a huge psychological effect, providing a ton of insurance against bad luck, a quality that many will find desirable.

So in other words, people who don't know how to make informed decisions will like it. Got it.

[Sidmen]

So in other words, people who don't know how to make informed decisions will like it. Got it.

It's more accurate to say: people who make informed decisions based on things other than mathematical averages will like it.

Boosting a pitiful damage roll to an amazing damage roll is a morale boost to the player and sometimes the whole group. That it will only happen 1 in 10 times is irrelevant, because when it does happen it's effects are more than simply dealing more damage.

Humans are not machines.

[archaeo]

So in other words, people who don't know how to make informed decisions will like it. Got it.

Or, you know, people who don't like bad die rolls when their party is counting on them to put out a ton of melee damage? People who want to raise their average damage with the massive number of feats a Fighter gets instead of taking a couple of attribute points? People who think the "savage" flavor makes sense for their characters from a roleplaying perspective?

But I guess we can just call these people "uninformed," if you like. That seems like a fair characterization of players who aren't winning at this elf game.

[Demonic Spoon]

"But the math says X!" is a poor argument against actual experience.

That depends on the topic.

Savage Attacker does literally nothing else other than give you more damage. It doesn't have any utility, and there are no game situations where it is more or less useful barring the situation that you roll low damage. If the feat were more nuanced in the situations that it gave you more damage then experience probably would matter more. Polearm master, mage slayer, and sentinel, for example, all vary in effectiveness based on what and where you're fighting.

When the feat effectively just gives you more damage on every roll, it is perfectly acceptable to talk about it in terms of raw math.

[Strill]

Or, you know, people who don't like bad die rolls when their party is counting on them to put out a ton of melee damage? People who want to raise their average damage with the massive number of feats a Fighter gets instead of taking a couple of attribute points? People who think the "savage" flavor makes sense for their characters from a roleplaying perspective?

But I guess we can just call these people "uninformed," if you like. That seems like a fair characterization of players who aren't winning at this elf game.

Then those people should get +2 STR instead. It does all the things you described, only better.

[Ferrin33]

Then those people should get +2 STR instead. It does all the things you described, only better.

To be fair, savage attacker allows you to add to damage beyond the 20 strength. I agree it's a terrible choice though.

[Demonic Spoon]

Or, you know, people who don't like bad die rolls when their party is counting on them to put out a ton of melee damage? People who want to raise their average damage with the massive number of feats a Fighter gets instead of taking a couple of attribute points? People who think the "savage" flavor makes sense for their characters from a roleplaying perspective?


Having a thing on your character sheet that says "Savage Attacker" doesn't do anything to give your character savage flavor. Great Weapon Master is much better at that. Sacrificing accuracy for power, and cleaving straight through one target into another has a lot more savage flavor than rerolling some low dice.

More strength will let you both hit more often and make a larger part of your damage a static bonus rather than relying on the weapon die - that does more for doing a consistent large amount of melee damage than Savage Attacker

A fighter who wanted to do those things with his high number of attribute increases would be better served either getting GWM or raising his strength. If a feat is only useful after your strength is maxed and after you've taken another feat, its design should be called into question.

[Pex]

On the average, when used perfectly, it's between +2 and +4 damage, nothing else. Saying that it does more in specific circumstances is misleading. In specific circumstances, your bonus attack from polearm master duelist quarterstaff crits and does 2*(4+2+5) = 22 damage. That doesn't mean the feat adds 22 damage; the average benefit is what we're concerned with.

Mathematically but not player logically. A generic player cares if a feat he takes helps him a significant number of times. When a player gets to reroll a 1 or 2 on his damage die he's going to be quite happy about that. It's not going to matter to him if over a 100 combats the feat only gave him +2 damage of what his overall total would have been without the feat. He cares that against monster X he rerolled a 2 into a 5, a 1 into 8, a 3 into 9. All the instances where he chose never to reroll or a rolled number only got one pip better he's not going care about as long as he gets to reroll enough 1's and 2's and see the feat doing its job.


It seems like this is a weak feat that will be beloved at some tables where no one would ever think to look up builds online and used to ban the 3.5 monk for being too powerful. It makes the game more fun. It's not so bad as to a be a trap option (like 3.5 toughness), and something had to balance polearm master to keep the average down.

Those players aren't wrong for not using builds online or thinking monks are too powerful for them. They just have a different style of taste.

Edited for better wording.

[Demonic Spoon]

Mathematically but not player logically. A generic player cares if a feat he takes helps him a significant number of times. When a player gets to reroll a 1 or 2 on his damage die he's going to be quite happy about that. It's not going to matter to him if over a 100 combats the feat only gave him +2 damage of what his overall total would have been without the feat. He cares that against monster X he rerolled a 2 into a 5, a 1 into 8, a 3 into 9. All the instances where he chose never to reroll or a rolled number only got one pip better he's not going care about as long as he gets to reroll enough 1's and 2's and see the feat doing its job.


It's those kinds of players who are hurt most by this kind of feat. They may be happy with its performance, but the problem runs deeper than that. Such players are likely to make other choices during character design which make them less powerful than other characters might be (which isn't necessarily bad - I'm not very much of an optimizer myself and do the same thing). When combined with other factors, things like picking a subpar feat can cause said player to drop behind other players in effectiveness. Joe Barbarian may be perfectly happy with the performance of his Savage Attacker feat and not notice that it is bringing him down. However, he will notice if other players start to pull ahead of him and he stops being relevant. And then the game gets less fun for him. And he has no idea why.

One of the big plusses of 5E is that there are very few trap options and it's hard to make a character that will get to the point of not being relevant unless that's actually your goal, so players can just build their character to fit a concept without having to worry about optimization if they don't want to.

This is not to say that Savage Attacker is such a bad feat that it is likely to be the difference between relevance and irrelevance in combat. Honestly, it's not actually that big a deal. However, that doesn't mean that the feat is well-designed or good for the game. It's not.

[Easy_Lee]

Suffice it to say that, if a player of yours chooses the feat, please consider homebrewing to make it stronger.

[Hytheter]

Even so I think my solution is elegant and makes the Greataxe much more attractive. It's currently much less effective than the Greatsword or Maul.

How does Savage Attacker make the Greataxe more attractive? Doesn't it do the exact same thing for the other weapons?

[Demonic Spoon]

How does Savage Attacker make the Greataxe more attractive? Doesn't it do the exact same thing for the other weapons?

1d12 is a lot more swingy than 2d6. Since you'll roll low more often with a single d12, being able to reroll lows is more attractive.

[Doug Lampert]

1d12 is a lot more swingy than 2d6. Since you'll roll low more often with a single d12, being able to reroll lows is more attractive.

Right: By my calculation:
2d6 with a reroll and take the higher roll gives an average of 8.37191358,
so it adds 1.3719 damage.

1d12 with a reroll and take the higher roll gives an average of 8.486111111.

Savage attacker actually makes 1d12 better than 2d6 till you get your second attack. So level 4 basically unless variant human.

[MaxWilson]

Suffice it to say that, if a player of yours chooses the feat, please consider homebrewing to make it stronger.

Maybe you could house-rule that Savage Attacker can be "upgraded" to Great Weapon Mastery any subsequent level. So at low level you have fun with your savage swings but you're not stuck having permanently wasted a feat when extra attacks come online.

[Sartharina]

Or have it grant +1 STR, making it a half-feat.

[Cambrian]

Suffice it to say that, if a player of yours chooses the feat, please consider homebrewing to make it stronger.What does everyone think of the OP's "half-feat" suggestion? Is the feat unreasonable as is if it also offers +1 STR/DEX?

[Yakk]

Suggestion: rolls of 1 become the max value on the die.

This is worth (1/die_size)*(die_size-1) per die, or a bit less than +1 damage per die.

Sadly, it helps 2d6 more than 1d12.

[Pex]

It's those kinds of players who are hurt most by this kind of feat. They may be happy with its performance, but the problem runs deeper than that. Such players are likely to make other choices during character design which make them less powerful than other characters might be (which isn't necessarily bad - I'm not very much of an optimizer myself and do the same thing). When combined with other factors, things like picking a subpar feat can cause said player to drop behind other players in effectiveness. Joe Barbarian may be perfectly happy with the performance of his Savage Attacker feat and not notice that it is bringing him down. However, he will notice if other players start to pull ahead of him and he stops being relevant. And then the game gets less fun for him. And he has no idea why.

One of the big plusses of 5E is that there are very few trap options and it's hard to make a character that will get to the point of not being relevant unless that's actually your goal, so players can just build their character to fit a concept without having to worry about optimization if they don't want to.

This is not to say that Savage Attacker is such a bad feat that it is likely to be the difference between relevance and irrelevance in combat. Honestly, it's not actually that big a deal. However, that doesn't mean that the feat is well-designed or good for the game. It's not.

The only thing that matters is a player being happy with his character's performance. If the feat works for him, that's his business if you would not make that choice for your character. It's the same issue with Vital Strike in Pathfinder. People here condemn it to Asmodeus' Palace, yet a player in my group does have it and is quite happy with it. He liked it so much he took Improved Vital Strike. It did not bother him at all he spent another feat on the concept. He feels no loss not using it when he makes a full attack. He's thrilled he gets to use it when he's only making one attack anyway because he moved.

Infinite uses of Savage Attacker for average damage increase comparisons are irrelevant. A player is not going to keep track. He experiences the jumps of low numbers to high and everyone cheers. His character does not suddenly become incompetent for having it, he notices improvement, and couldn't care a rat's posterior what another character does. He's not the Suck nor a terrible player just because you wouldn't take the feat.

However, I force myself to avoid a hint of hypocrisy considering the Thesis Wish thread where I want to change it while others think it's fine as is. If Savage Attacker is not good enough for you, that is your prerogative. Suggestions of improvement for those who think it needs improvement is a valid discussion.

Some ideas, mainly redoing the feat text.

It allows a reroll of a Natural 1 on attack rolls, a mini-halfling trait.

If you fail to hit your target by 1, you get a glancing blow that just deals weapon damage, no extra damage from attribute, feats, class abilities, spells, or weapon enchantment.

You cannot have disadvantage on melee weapon attack rolls.

After an opponent attacks you, you can use your reaction to attack that opponent once with a melee weapon if within reach.

[Strill]

What does everyone think of the OP's "half-feat" suggestion? Is the feat unreasonable as is if it also offers +1 STR/DEX?

Well +2 STR is generally worth around +20% overall damage, so for that to be worth it, it would need to improve overall damage by about +10%. So then if it gave +1 STR, AND was usable on every attack, then yes it might be worth it.

[JFahy]

"But the math says X!" is a poor argument against actual experience.

This is pretty much the mentality that Moneyball is about - people valuing
their experiences (and what they think those experiences mean) more than
mathematical data. Gotta be careful with that.

1-2 extra damage sounds disappointing if you're attacking a raid boss with a
million hit points and 39 people helping you, but we should be able to come up
with a more realistic context than that. Here's one.

If you're hitting for 1d6+4 against something with 9 hit points, as you might
find yourself doing in Hoard of the Dragon Queen, Savage Attacker is
"your hits have a 55% chance instead of a 33% chance of putting an opponent
down". When you're getting swarmed by bad guys that's a pretty big deal.

At higher level one-shots are less common and the 'big deal' diminishes, but
SA will still help you put down wounded targets, get phat crits that affect morale,
and force harder concentration checks. The effect there can be bigger than the
'aggregate effect on damage' suggests.

[Easy_Lee]

What does everyone think of the OP's "half-feat" suggestion? Is the feat unreasonable as is if it also offers +1 STR/DEX?

If I did the math right, savage attacker used perfectly adds 4 damage average for a four-attack round with a greataxe (26->29.97). That's a 15% damage increase. So if you add +1 STR on top of that, and the assumption that +2 STR adds 20% damage is correct, then I'd say it's a decent feat choice. That puts it better than some options (dual wielder), but worse than others (polearm master, crossbow expert), a good place for a homebrew feat to be.

[archaeo]

Then those people should get +2 STR instead. It does all the things you described, only better.

I don't think a +2 provides quite the same "hit" that being able to reroll a bad damage die does, but whatever, we're not going to agree here.


Having a thing on your character sheet that says "Savage Attacker" doesn't do anything to give your character savage flavor. Great Weapon Master is much better at that. Sacrificing accuracy for power, and cleaving straight through one target into another has a lot more savage flavor than rerolling some low dice.

This is a matter of personal taste more than anything else. It appeals to a certain kind of player, who will regard it as very strong as it gives a once-per-turn reroll to save from getting lousy damage. Other players will naturally be more concerned with that straight boost to DPR. And neither one of them is wrong, since I still don't see evidence of the kind of difference that will really make a player wince at their terrible feat choice or anything.

But "savage attacker" has good "savage" flavor, in my opinion.


What does everyone think of the OP's "half-feat" suggestion? Is the feat unreasonable as is if it also offers +1 STR/DEX?

This seems like the nicest solution, though it would go from being a fair choice to being one of the top tier half-feats for melee bros. I'm pretty sure only Heavy Armor Master is better, and mostly just because you're likely to be taking half-feats at level 4.


This is pretty much the mentality that Moneyball is about - people valuing
their experiences (and what they think those experiences mean) more than
mathematical data. Gotta be careful with that.

"Money" is the operative term in that phrase, JFahy. Improving an MLB team could mean a multimillion dollar windfall; improving your DPR by some nominal number of average damage means that, at best, you're going to manage to end a combat a few turns earlier than you would have otherwise.

Not that there's anything wrong with aiming for high DPR! It's just, you know, even less important than baseball, which is saying something.

[Sartharina]

"Money" is the operative term in that phrase, JFahy. Improving an MLB team could mean a multimillion dollar windfall; improving your DPR by some nominal number of average damage means that, at best, you're going to manage to end a combat a few turns earlier than you would have otherwise.

Which can be the difference between a curbstomp victory and a TPK.

[Segev]

Let's examine the assumptions behind "the math."

It's not accurate to say that this is going to average 8.26, because there's a matter of selection involved. If this were, in fact, rolling twice and taking the higher die every time, that would be accurate on a 1d12. If you have only one attack per turn, this is again accurate, because you'll always use it on your one attack.

I don't have the feat in front of me, but I will assume that it's one you choose to use after rolling the damage for the attack, but can't choose to use on an earlier attack after rolling damage on a later one. This means that you'll ALWAYS use it on the last attack you have in a round if you have not already done so. But, you might use it sooner, if you have more than one.

Let's assume a d12 die, which averages 6.5 normally and, if you're just using this feat on the roll no matter its result, the average of that roll will be 8.26.

If you ALWAYS use it on the first roll, then that first roll averages 8.26, and all subsequent ones average 6.5. Not all that hot; it increases your expected damage/round by less than 2.

However, if you establish a minimum number below which you will choose to use this feat, you will increase the expected value of your overall damage. I don't really want to take the time right now to calculate various scenarios, but I would estimate that you would see the biggest return if you had a linear function of the number of attacks you've got left in the round, such that the first attack only has this feat applied if the number is particularly low, while the second attack gets it applied even if the number rolled is slightly higher, until you will always roll it if the result is less than the maximum on the last roll.

The expected value for the round is the sum of the averages of all your dice. If you could choose any one of the dice to re-roll after all were rolled, you could maximize the expected value due to this re-roll, but the ambiguity of later rolls complicates this.

Still, because there is more judgment than "one attack per round gets to roll twice," it's not just the average result of two d12s-take-the-highest. You can choose to save the one re-roll for when the die roll is low, unless none of the die rolls are particularly low (in which case you re-roll the last one).

Because you're selecting for a lower-than-average roll, you don't have circumstances where a 10 replaces a 9 all that often. You'll be replacing, usually, an on-average lower result from the set making up your sum. This reduces the number of low rolls making up your average sum, as well as introducing a potentially-higher roll. This is non-trivial to calculate without exhaustive probability break-downs, so I will leave those to others if they care. (I'd do it in matlab, but my computer no longer has it, sadly.)

[MustacheFart]

Honestly, I would homebrew it as:

1 per Short/Long Rest. After successfully hitting with a melee attack you may use this feat to automatically deliver max damage.

This would fall in line with similar abilities that already exist in 5th ed. Also it would give melee players the ability to deliver a large amount of reliable burst on a limited basis. Combo it with a critical hit for a lot more damage.

[Strill]

Honestly, I would homebrew it as:

1 per Short/Long Rest. After successfully hitting with a melee attack you may use this feat to automatically deliver max damage.

This would fall in line with similar abilities that already exist in 5th ed. Also it would give melee players the ability to deliver a large amount of reliable burst on a limited basis. Combo it with a critical hit for a lot more damage.

That's great for a Rogue or a Paladin, but not worthwhile for anyone else.

[Demonic Spoon]

The only thing that matters is a player being happy with his character's performance. If the feat works for him, that's his business if you would not make that choice for your character. It's the same issue with Vital Strike in Pathfinder. People here condemn it to Asmodeus' Palace, yet a player in my group does have it and is quite happy with it. He liked it so much he took Improved Vital Strike. It did not bother him at all he spent another feat on the concept. He feels no loss not using it when he makes a full attack. He's thrilled he gets to use it when he's only making one attack anyway because he moved.

Infinite uses of Savage Attacker for average damage increase comparisons are irrelevant. A player is not going to keep track. He experiences the jumps of low numbers to high and everyone cheers. His character does not suddenly become incompetent for having it, he notices improvement, and couldn't care a rat's posterior what another character does. He's not the Suck nor a terrible player just because you wouldn't take the feat.


But that's not my point. My argument was specifically not that the player will be disappointed with his choice in feat. My argument was that options that are outright bad compared to other options further an environment where some players contribute substantially more than others, and pressure players to optimize more than they would otherwise do, which isn't fun.

[MustacheFart]

That's great for a Rogue or a Paladin, but not worthwhile for anyone else.

Not true. It most optimal for the two classes you specified but it is still beneficial to fighter and barbarian (ie. champion fighter or either path barbarian thanks to extra crit dice OR battlemaster's superiority dice which would be maxed as well). A feat doesn't have to be optimal for everyone but should be beneficial for more than it is optimal, which my example is. Plus my goal wouldn't be to make a crappy feat the best feat but simply a bit better.

[Ralanr]

Honestly, I would homebrew it as:

1 per Short/Long Rest. After successfully hitting with a melee attack you may use this feat to automatically deliver max damage.

This would fall in line with similar abilities that already exist in 5th ed. Also it would give melee players the ability to deliver a large amount of reliable burst on a limited basis. Combo it with a critical hit for a lot more damage.

I like it. Makes the feat fair for greataxes, greatswords, and mauls.

[Ferrin33]

Not true. It most optimal for the two classes you specified but it is still beneficial to fighter and barbarian (ie. champion fighter or either path barbarian thanks to extra crit dice OR battlemaster's superiority dice which would be maxed as well). A feat doesn't have to be optimal for everyone but should be beneficial for more than it is optimal, which my example is. Plus my goal wouldn't be to make a crappy feat the best feat but simply a bit better.

I suggest at least allowing it to be used after the damage is rolled so you can still see if it's worth using there with your suggested change.

[Easy_Lee]

1 per Short/Long Rest. After successfully hitting with a melee attack you may use this feat to automatically deliver max damage.

Savage attacker alert assassin would be quite powerful.

[MustacheFart]

I suggest at least allowing it to be used after the damage is rolled so you can still see if it's worth using there with your suggested change.

Of course. Just like the original they could decide to use it after they've already rolled.

[MustacheFart]

Savage attacker alert assassin would be quite powerful.

For 1 time per short rest/long rest...assuming they can get a surprise round lol.

[Rummy]

I like the solution of keeping it as is, but making it also add +1 Str, like Athlete.

[Malifice]

I too have seen to many 1's and 2's rerolled into 10, 11 and 12s.

Confirmation bias not backed up by the math.

Do you ever see 12's and 11's rerolled succesfully?

[Segev]

Confirmation bias not backed up by the math.

Do you ever see 12's and 11's rerolled succesfully?

Er... why would they use their re-roll on max or near-max rolls? They'd save it for low rolls, likely with more willingness to use it the fewer attacks they have left in the round. One could calculate an optimal threshold below which the power should be used based on how many attacks you have left in the round. The only one I'm positive of is, "use it on the last one if you still have it, unless the last one already rolled max." But I doubt, if you have 3 attacks/round, that you'd use it on your first attack if the damage came up within 1-2 of maximum. You'd save it for the next one, in case it came up lower.

That's why this can't be measured statistically as if it's just "one of your attacks is rerolled." The informed choice as to when to use it - after seeing a roll come up poorly - skews it higher in its impact on expected damage, but also makes it harder to calculate the effects mathematically.

[MaxWilson]

That's why this can't be measured statistically as if it's just "one of your attacks is rerolled." The informed choice as to when to use it - after seeing a roll come up poorly - skews it higher in its impact on expected damage, but also makes it harder to calculate the effects mathematically.

But it's easy to calculate a ceiling on its effectiveness, as a prior poster already did. Just make it behave like "five rolls drop lowest" for four attacks, and you're guaranteed that Savage Attacks cannot be better than that no matter how optimally the player uses it.

Amusingly, this is ALMOST the secretary problem (http://en.wikipedia.org/wiki/Secretary_problem) except that the probability distribution isn't unknown, so you can actually improve upon the Secretary Problem's standard solution.

[Easy_Lee]

Thought I'd add that I did the math for greatswords for four swings/round. It's not 10d6 drop two because we drop entire attacks. Instead, I used 5d13 drop one. Average damage without is 28, with is 32.3. We can ignore stat boosts and other benefits since they're either fixed or we don't roll them.

This assumes all attacks hit. If only two attacks hit, the feat adds just 3.23 damage average. For three hits, it adds 3.87 damage.

In conclusion, this feat adds about 2 damage per round to a greatsword, eventually scaling up to 4 per round, but only if used perfectly. For that reason, it is statistically inferior to TWF, which adds 1AC, 1 damage per hit, and an extra draw and is still considered one of the worst feats around. I'd house rule or avoid savage attacker.