Weapon 'Damage Band' impact

[Torture]

Hey chaps. I'm hoping some of the cleverer folk on here might be able to explain something to me.

To set a brief background. I was running with Abom knuckles main hand, FGC offhand before the move to 25's. So far still no luck with either KS or BB, but I was lucky enough to grab Pugius from our first 10 man LK kill, and the heroic FGC from the resulting hardmode run.

So, I find myself with a heroic FGC as the 'best' of my current weapons and my initial thought was 'that goes in the main hand then' replacing Pugius and leaving me with 2 FGC, 1x264, 1x251.

However, I pondered how sensible replacing a 258 + 251 with 264 + 251 really was. Surely it was better to run with the highest overall items, i.e Pugius main hand and the new FGC offhand. But that meant I was putting my highest DPS weapon in my offhand which also felt a bit odd.

So, I obviously ran it through the sims and rawr and it turns out I was right in my assumption that Pugius Main hand and FGC offhand netted the best results.

Upon closer inspection, Pugius has an incredibly high 'minimum damage' in it's weapon damage band. Far higher than the FGC and I realised I knew nothing about how those damage bands work.

Is a weapon with 100-900 damage inferior to a 500-600 simply because the latter will provide a higher median damage spread?

Before I played WoW I played Anarchy online. In that game your weapon damage bracket (i.e 200-500) was determined by the armour of the mob. If the mob had low armour you'd be hitting the top end, if it had high armour then the 'minimum damage' number became far more important than the 'maximum damage' number. On high armour mobs a 200-300 weapon would outperform a 100-1000 weapon as the mob armour was effectively limiting you to minimum damage.

Could someone either explain to me, or point me in the direction of a good source of information on just how the damage spreads on weapons work and the relative importance of high max damage vs high minimum damage, and whether these change with circumstance (i.e would a higher 'minimum damage' weapon be more suitable on certain encounters).

Just realised it was something I knew nothing about, so I thought I'd ask

/Tort

[Foomoo]

Without any reference to specific encounters, if 2 "things" have the same expected value, and thing A has a larger variance than thing B, then for most purposes B > A.

So if two weapons have the same DPS and same speed, then the one with the smaller damage range is "better".

[Omanley]

the gain of dps is overvalued i think.

http://www.wowhead.com/item=51801 + http://www.wowhead.com/item=51916
127 agility 214 atp ( socket included ) , 95 crit-r, 42 haste, 31 armor-pen, 1 red socket

vs

http://www.wowhead.com/item=51916 + http://www.wowhead.com/item=50787
123 agility, 208 atp ( socket bonus included ), 95 crit-r, 66 armor-pen and 1 red socket

=> 4 agility, 6 atp, 42 haste vs 35 armor pen
in terms of dps difference:
-10/+21 vs +10/-21

dont forget that a higher offhand dps give you better ss damage ( although not as high as mainhand dps ), better white damage ( same here ), but also better lavalash damage

and in terms of stats, pugius/hc-fgc easily wins, making up for that little mainhand dps inferiority.

[Torture]

the gain of dps is overvalued i think.

http://www.wowhead.com/item=51801 + http://www.wowhead.com/item=51916
127 agility 214 atp ( socket included ) , 95 crit-r, 42 haste, 31 armor-pen, 1 red socket

vs

http://www.wowhead.com/item=51916 + http://www.wowhead.com/item=50787
123 agility, 208 atp ( socket bonus included ), 95 crit-r, 66 armor-pen and 1 red socket

=> 4 agility, 6 atp, 42 haste vs 35 armor pen
in terms of dps difference:
-10/+21 vs +10/-21

dont forget that a higher offhand dps give you better ss damage ( although not as high as mainhand dps ), better white damage ( same here ), but also better lavalash damage

and in terms of stats, pugius/hc-fgc easily wins, making up for that little mainhand dps inferiority.

Yup, that's pretty much exactly what the sims concluded

It's not really my current circumstances that I'm concerned with. It's more that in resolving my current circumstances a question was raised in my head to which I had no answer (i.e the significance of the various elements of a weapon's damage band) and I just fancied filling in that gap in my knowledge. I only used my personal weapons as an illustration of where the question originated, but thank you none the less

[Izzirez]

As far as weapon dps goes and the 'damage bands', weapons with higher top ends also have lower low ends, evening out the damage in the end. Say your weapon has 900 top end and is a 250 dps weapon, it will have higher damage SS, WF's, etc, but it will also have an equal amount of low damage SS, WF's, etc, compared to a weapon that is, say, 805 top end, but has a higher low end damage.

Overall, it comes down to speed > dps > stats.
Slow weapons always come first, then you check the weapon dps, then you check the stats. Of course there is some variation with this because of the complex way our stats affect us, but that is the basic idea.

[Billnye]

The damage range on a weapon does not affect actual damage output in a similar way to you described Anarchy Online's system.

When calculating an attack's damage, a random number between the low end and high end of the range is chosen. That random number is then used as the base for the rest of the attack's damage. The number that is chosen has nothing to do with any other stats that you or your target may have; it's a simple die roll.

This means that a weapon with a range of 500-1000 will put out, on average, the same amount of DPS as a weapon with a range of 650-850, or as 100-1400 (assuming the weapons have the same stats and speed, other than damage range). The average of each of the ranges is 750, so that is what the average of the random rolls will be on each separate weapon. The highest possible damage from a single attack will be from the 100-1400 weapon, but that weapon will also have the lowest possible damage from a single attack. In the end, the overall DPS from each weapon will be the same, you just have a larger variation on the damage done by each attack with the weapon that has a larger damage range.

tl;dr - The actual low end and high end numbers of the damage range aren't really important. What's important is the average of the two (also found by multiplying the speed by the weapon DPS).

[Torture]

The damage range on a weapon does not affect actual damage output in a similar way to you described Anarchy Online's system.

When calculating an attack's damage, a random number between the low end and high end of the range is chosen. That random number is then used as the base for the rest of the attack's damage. The number that is chosen has nothing to do with any other stats that you or your target may have; it's a simple die roll.

This means that a weapon with a range of 500-1000 will put out, on average, the same amount of DPS as a weapon with a range of 650-850, or as 100-1400 (assuming the weapons have the same stats and speed, other than damage range). The average of each of the ranges is 750, so that is what the average of the random rolls will be on each separate weapon. The highest possible damage from a single attack will be from the 100-1400 weapon, but that weapon will also have the lowest possible damage from a single attack. In the end, the overall DPS from each weapon will be the same, you just have a larger variation on the damage done by each attack with the weapon that has a larger damage range.

tl;dr - The actual low end and high end numbers of the damage range aren't really important. What's important is the average of the two (also found by multiplying the speed by the weapon DPS).

Perfect, thank you. Nice to know there aren't any underlying mechanics I wasn't aware of. That answers the question

[Deleted]

In many ways, sims will favour higher top end damage even if the average DPS is not better. Take two weapons for example. One does 500DPS and always hits for exactly the same amount, the other does 500DPS on average but has a high top end and a low bottom end. While both will even out to the same thing, the second one is cabable of hitting higher - even it it requires some RNG.

[Izzirez]

In many ways, sims will favour higher top end damage even if the average DPS is not better. Take two weapons for example. One does 500DPS and always hits for exactly the same amount, the other does 500DPS on average but has a high top end and a low bottom end. While both will even out to the same thing, the second one is cabable of hitting higher - even it it requires some RNG.

This is why you sim it for many thousands of hours. All rng will be evened out because of the sample size.

[Deleted]

No I don't mean averages - I just mean that the one with the higher top end damage has the highest theoretical damage - even though it means little in practice.

[shakkenbake]

No I don't mean averages - I just mean that the one with the higher top end damage has the highest theoretical damage - even though it means little in practice.

But do to the nature of random numbers, it has just as much of a chance to be the theoretical lowest damage of the two. Your logic is flawed. Random numbers are random and averaged numbers are averaged.

Here's something for thought though.. random number generators are never completely random; They must be created through some sort of algorithm that normally uses the exact time and date (maybe even IP address) to determine the outcome of the 'roll'.

So in essence, these random numbers can be predicted if the algorithm is known. And if you could figure out the algorithm, and find a reliable way to predict when the greatest time and date is to generate the highest 'random' numbers you could in theory use the weapon with the widest range of damage to increase the effect of this.

This would be extremely difficult though, and real life practice of this would be next to impossible (and quite possibly land you a ban). In theory though it works out.

[Billnye]

No I don't mean averages - I just mean that the one with the higher top end damage has the highest theoretical damage - even though it means little in practice.

Any sim that doesn't choose a random number from the weapon's damage range but, instead, chooses a number on the high end of the scale.. is a bad sim. It's just not correct. EnhSim doesn't even have a place to input your damage range. It simply asks for weapon DPS, because the author knew that the high damage attacks and the low damage attacks would average out in the end.

Yes, the highest damage hits will always come from a weapon with a larger range, but if those are the attacks that the sim is using most often then the sim is just plain inaccurate. If the sim is accurate, then it does exactly what WoW's combat code does. It takes a random number between the low end and the high end of the weapon's range. And.. if it takes a random number every time, then the results will AVERAGE OUT to a certain number.

A higher weapon will have the highest theoretical DAMAGE (that is, when comparing single attacks), but not the highest theoretical DAMAGE PER SECOND. Because sims measure DPS, a weapon with a large damage range and a weapon with a small damage range will have the same results.

Basically, a sim is just as likely to give a lower output for a high-range weapon (because they also have lower low-ends in addition to higher high-ends) as it is to give a higher output. That is, it is not likely to happen at all if the sim is even anywhere near accurate.

[Fornaw]

But do to the nature of random numbers, it has just as much of a chance to be the theoretical lowest damage of the two. Your logic is flawed. Random numbers are random and averaged numbers are averaged.

Here's something for thought though.. random number generators are never completely random; They must be created through some sort of algorithm that normally uses the exact time and date (maybe even IP address) to determine the outcome of the 'roll'.

So in essence, these random numbers can be predicted if the algorithm is known. And if you could figure out the algorithm, and find a reliable way to predict when the greatest time and date is to generate the highest 'random' numbers you could in theory use the weapon with the widest range of damage to increase the effect of this.

This would be extremely difficult though, and real life practice of this would be next to impossible (and quite possibly land you a ban). In theory though it works out.

An interesting thought exercise, however any random number generator where this is at all exploitable is a bad random number generator. No matter what your seed(s) are the expected value should be the same, that would be a key aspect of a good RNG.

Personally I would prefer weapons with small damage ranges for consistency in the short term fights (valkyrs etc) but it's a back burner issue. Damage ranges in WoW are usually determined by DPS, Speed and Weapon Type. Fists usually have the smallest damage ranges, but the difference is pretty negligible compared to the actual damage after AP.