Combat Mechanics? Attack Speed, Recovery

[Daemonjax]

 

 

2-mins testing suggests that it does behave that way. My rogue does not have its damage multiplied by 0.5/2.0 when it grazes/crits. That said, no, I'm not absolutely certain.

 

 

I was testing this myself today and can confirm you're right.  The higher you get your additive multiplier (rogues get start at 1.5 for sneak attack and ciphers start with 1.4 with talent, and then there's other talents and weapon damage enchants), the less crits/grazes matter. 

Edited April 8, 2015 by Daemonjax

[tonberrian]

My damage calculator is marginally useful now! I haven't yet implemented attack speed, but you should be able to compare weapons in the same speed tiers fairly easily.

[Etherlord]

Here's my $0.02, reposted from Steam: http://steamcommunity.com/app/291650/discussions/0/618458030677032991/

 

Arbalest or War bow

So because of the DR mechanic, I've quickly drown a conclusion, that it's benefitious to use high damage weapons, even if they are slow. I taught my ranger the soldier specialization, and got impressed by the ability to one-shot enemies. However later in the game I got really dissappointed by average results and loaded an old save before my decision of condemning the war bow. I would end at it, but my friend, RPG player with big experience in ADnD told me arbalest is actually better.

So I started to count.

War bow is three times faster than arbalest - that is, in the same amount of time you can either:

• shoot an arrow, recover, shoot again, recover, shoot the third time and fully recover
    
• shoot a quarrel, recover

So I'd expect arbalests to have 3 times higher damage in order to have higher DPS.

how average damage depends on accuracy and deflection on DR 7

 

how average dmg depends on DR when ACC = DEF

 

 

how average dmg depends on DR when ACC = DEF +30

How these values were counted?
Most values were counted in Excel. Here's the main spreadsheet: download
I took MIN/MAX damage values from my ranger as well as her ACC, and I took DEF value from a real NPC. Then I calculated chances to graze, hit, crit, based on ACC and DEF, calculated damage for graze (av. dmg/50% - DR), hit (av. dmg - DR), crit, and multiplied these damage values by their chances to occur.

I'm not sure if I calculated critical damage correctly. For war bow it's easy:

average damage * 1.5

, but I had to guess how "-0.3 crit damage multiplayer" works - my guess was:

av_dmg + av_dmg*0.5*(1-0.3) = av_dmg + av_dmg*0.5*0.7

In cases where DR was high enough, I considered the mechanic that limits Damage reduction to 80%, by creating a simple table with all possible damage outcomes, damage reduced by DR completely and damage reduced by 80% and taking the highest value from these two, and taking an average of the highest values as you can see here: arbalest grazes against DR 16

I didn't consider any other bonuses / penalties.I also didn't consider how the game rounds values

Do I miss something? If not, here are my conclusions:

• because crit damage is a percentage boost, faster or higher damage weapon will not benefit from it more, unless DR is so high, that in some cases not only base, but also crit bonus of a damage will be absorbed by DR (not the case in my tests, the DR would need to be much higher than 16)
    
• because arbalest has a penalty to crits, war bow benefits more from higher accuracy
    
• because DR is a flat value, war bow benefits more from higher accuracy, as it decreases chances of grazes and criticals (so increases damage)
    
• war bow has higher dps in any case except DR higher than 16 - in that case you probably want to switch to a wand or rod for their alternative damage type
    
• while dps of war bow is higher, you must be aware, that arbalest deals its damage sooner; this way you can possibly kill a target quickly enough, to deny some damage for the enemy; also in case of your ranger's quick death, his total damage may be higher with arbalest, than with war bow
    
• if you use a loading spell a lot, arbalest crits will be a decent way to start any encounter.
• keep in mind that because arbalest will overkill enemies by higher amounts of damage, it means it will waste more damage and so it's real DPS decreases.

Final conclusions:

• weapons seem very balanced in Pillars of Eternity; if you see two weapons of same quality and think one of them is better in all situations - you probably don't understand all game mechanics
    
• weapons like arbalest, crossbow and firearms are designed with melee fighers in mind - the purpose of these weapons is to make a single strong attack at the start of the battle and then switch to another weapon, not wasting time to reload it
    
• similarily to above, your ranged characters could have one of these weapons to start a fight and then switch to a bow or a wand.

If you find it interesting, I may also write about my experience with pistols and blunderbusses.

Edited April 8, 2015 by Etherlord

[Etherlord]

^^^ above post edited. Sorry for double post.

Edited April 8, 2015 by Etherlord

[sunburnedcrow]

Is this summary correct?

There are, potentially, six parts to each ranged attack

1) Attack delay

2) Attack Animation

3) Recovery delay

4) Recovery Animation

5) Reload delay (?)

6) Reload Animation

Dexterity bonus/malus affects Attack Animation, Recovery Animation and Reload Animation by the formula (No. of frames divided by (1+ Dexterity bonus/malus)

The other bonus/malus to attack speed in game only affects Recovery Animation, or Reload Animation if stated. These bonuses are additive, i.e. if you have a -50% attack speed and a +20% attack speed bonus, the multiplier would be (1+0.5-0.2)=1.3

If the above is correct, I have some questions:

1) Are the frame delays the same throughout all weapons?

2) Do the dexterity bonus affect the frame delays.

Personally, I think the frame delays are a quirk of the engine or just error when measuring.

OR

We could just ask the developers once they are less busy with fixing bugs (which should be the priority now)

[peddroelm]

Confident Aim – A percentage of a fighter’s Grazes are converted to

Hits. Additionally, the minimum damage for any melee weapon they use

is increased by 20% of the range between minimum and maximum.

 

With this pistol damage range becomes 22 * 1.2 = 26.4 - 30 .. Pretty narrow but not quite min=max we need for proper testing .. 

 

I did a cursory  examination of spells in manual - didn't find any that would operate on weapon damage range .. 

 

Perhaps one of the unique weapons already has tight damage range that Confident Aim could tighten to constant ? 

 

Perhaps one of the unique items in the game operates on weapon damage range (before the additive damage modifiers are added ) ...

 

Anybody has a list of all POE unique items stats ? Any modder capable of upping Confident bonus to 100% for testing purposes ? 

[Matt516]

Confident Aim – A percentage of a fighter’s Grazes are converted to

Hits. Additionally, the minimum damage for any melee weapon they use

is increased by 20% of the range between minimum and maximum.

 

With this pistol damage range becomes 22 * 1.2 = 26.4 - 30 .. Pretty narrow but not quite min=max we need for proper testing .. 

 

Read the description again. Range between min and max is 8. 20% of 8 is 1.6. It's not 26.4 to 30, it's 23.6 to 30.

 

EDIT: Also, only works with melee weapons.

Edited April 8, 2015 by Matt516

[Perkus]

Arbalest or War bow

 

War bow is three times faster than arbalest

 

Let me stop you right there.  Where did you get the idea that the firing rate is exactly 3x between these two?  According to our work on this spreadsheet, at Dex 10, with 0 speed modifers (attack, recovery, reload), arbalest takes 294 frames, or 9.80s @ 30fps.   War Bow takes 129 frames, 4.30s.  The ratio between them is 9.8/4.3 = 2.28.  Pretty big difference right there.  Any reload mod benefits only the arbalest.  Armor recovery penalty also benefits the arbalest (ratio down to 1.97 @ -50%).  Only attack speed benefits the war bow.  And you'd need to be naked with a 90% attack speed modifer to get to a 3.00 attack rate ratio.  So unless that's your typical scenario, the rest of your calculations are significantly off.

 

And why are you even comparing it to an arbalest instead of an Arquebus?  My spreadsheets suggest the Arquebus is almost always significantly higher DPS than the arbalest.  It's not that much slower, has a 6 DR bypass along with +3 average damage, which is an enormous advantage.

[Myrten]

My damage calculator is marginally useful now! I haven't yet implemented attack speed, but you should be able to compare weapons in the same speed tiers fairly easily.

Instead of making separate chart you could make separate sheet on our chart and use our data

[Myrten]

Is this summary correct?

There are, potentially, six parts to each ranged attack

1) Attack delay

2) Attack Animation

3) Recovery delay

4) Recovery Animation

5) Reload delay (?)

6) Reload Animation

Dexterity bonus/malus affects Attack Animation, Recovery Animation and Reload Animation by the formula (No. of frames divided by (1+ Dexterity bonus/malus)

The other bonus/malus to attack speed in game only affects Recovery Animation, or Reload Animation if stated. These bonuses are additive, i.e. if you have a -50% attack speed and a +20% attack speed bonus, the multiplier would be (1+0.5-0.2)=1.3

If the above is correct, I have some questions:

1) Are the frame delays the same throughout all weapons?

2) Do the dexterity bonus affect the frame delays.

Personally, I think the frame delays are a quirk of the engine or just error when measuring.

OR

We could just ask the developers once they are less busy with fixing bugs (which should be the priority now)

Your assumptions are correct.

 

Frame delays are like 5 +-3 frames. There is no evidence that they are affected by anything.

[Myrten]

Here's my $0.02, reposted from Steam: http://steamcommunity.com/app/291650/discussions/0/618458030677032991/

 

Arbalest or War bow

 

So because of the DR mechanic, I've quickly drown a conclusion, that it's benefitious to use high damage weapons, even if they are slow. I taught my ranger the soldier specialization, and got impressed by the ability to one-shot enemies. However later in the game I got really dissappointed by average results and loaded an old save before my decision of condemning the war bow. I would end at it, but my friend, RPG player with big experience in ADnD told me arbalest is actually better.

 

So I started to count.

 

War bow is three times faster than arbalest - that is, in the same amount of time you can either:

• shoot an arrow, recover, shoot again, recover, shoot the third time and fully recover

    

• shoot a quarrel, recover

So I'd expect arbalests to have 3 times higher damage in order to have higher DPS.

how average damage depends on accuracy and deflection on DR 7

 

how average dmg depends on DR when ACC = DEF

 

 

how average dmg depends on DR when ACC = DEF +30

 

How these values were counted?

Most values were counted in Excel. Here's the main spreadsheet: download

I took MIN/MAX damage values from my ranger as well as her ACC, and I took DEF value from a real NPC. Then I calculated chances to graze, hit, crit, based on ACC and DEF, calculated damage for graze (av. dmg/50% - DR), hit (av. dmg - DR), crit, and multiplied these damage values by their chances to occur.

 

I'm not sure if I calculated critical damage correctly. For war bow it's easy:

average damage * 1.5

, but I had to guess how "-0.3 crit damage multiplayer" works - my guess was:

av_dmg + av_dmg*0.5*(1-0.3) = av_dmg + av_dmg*0.5*0.7

In cases where DR was high enough, I considered the mechanic that limits Damage reduction to 80%, by creating a simple table with all possible damage outcomes, damage reduced by DR completely and damage reduced by 80% and taking the highest value from these two, and taking an average of the highest values as you can see here: arbalest grazes against DR 16

 

I didn't consider any other bonuses / penalties.I also didn't consider how the game rounds values

 

Do I miss something? If not, here are my conclusions:

• because crit damage is a percentage boost, faster or higher damage weapon will not benefit from it more, unless DR is so high, that in some cases not only base, but also crit bonus of a damage will be absorbed by DR (not the case in my tests, the DR would need to be much higher than 16)

    

• because arbalest has a penalty to crits, war bow benefits more from higher accuracy

    

• because DR is a flat value, war bow benefits more from higher accuracy, as it decreases chances of grazes and criticals (so increases damage)

    

• war bow has higher dps in any case except DR higher than 16 - in that case you probably want to switch to a wand or rod for their alternative damage type

    

• while dps of war bow is higher, you must be aware, that arbalest deals its damage sooner; this way you can possibly kill a target quickly enough, to deny some damage for the enemy; also in case of your ranger's quick death, his total damage may be higher with arbalest, than with war bow

    

• if you use a loading spell a lot, arbalest crits will be a decent way to start any encounter.
• keep in mind that because arbalest will overkill enemies by higher amounts of damage, it means it will waste more damage and so it's real DPS decreases.

Final conclusions:

• weapons seem very balanced in Pillars of Eternity; if you see two weapons of same quality and think one of them is better in all situations - you probably don't understand all game mechanics

    

• weapons like arbalest, crossbow and firearms are designed with melee fighers in mind - the purpose of these weapons is to make a single strong attack at the start of the battle and then switch to another weapon, not wasting time to reload it

    

• similarily to above, your ranged characters could have one of these weapons to start a fight and then switch to a bow or a wand.

If you find it interesting, I may also write about my experience with pistols and blunderbusses.

 

Way in which damage bonuses are applied totally derails your theory - the more damage bonus the less accuracy is important. Futhermore reload bonuses can improve arbalest performance significantly.

 

You can take a look at or spreadsheet here: https://docs.google.com/spreadsheets/d/1vDm5MOrCK6S95h5u0EzZbGTv_u_rYqjuO0-zupqk0_A/edit#gid=1264774515

Edited April 9, 2015 by Myrten

[Myrten]

I've modified our spreadsheed to reflect discoveries made by Daemonjax. Now we only have DEX, Attack+Recovery Mods and Reload mod as modifiers affecting speed.

 

One question - does Attack Speed\Recovery Speed modifiers affect reload time?

 

Sorry for multi-posting, but there are several important issues here and I don't want them to be mixed up.

Edited April 9, 2015 by Myrten

[Tvättbjörn]

Here's my $0.02, reposted from Steam: http://steamcommunity.com/app/291650/discussions/0/618458030677032991/

Do I miss something? If not, here are my conclusions:

• because crit damage is a percentage boost, faster or higher damage weapon will not benefit from it more, unless DR is so high, that in some cases not only base, but also crit bonus of a damage will be absorbed by DR (not the case in my tests, the DR would need to be much higher than 16)

    

• because arbalest has a penalty to crits, war bow benefits more from higher accuracy

    

• because DR is a flat value, war bow benefits more from higher accuracy, as it decreases chances of grazes and criticals (so increases damage)

    

• war bow has higher dps in any case except DR higher than 16 - in that case you probably want to switch to a wand or rod for their alternative damage type

    

• while dps of war bow is higher, you must be aware, that arbalest deals its damage sooner; this way you can possibly kill a target quickly enough, to deny some damage for the enemy; also in case of your ranger's quick death, his total damage may be higher with arbalest, than with war bow

    

• if you use a loading spell a lot, arbalest crits will be a decent way to start any encounter.
• keep in mind that because arbalest will overkill enemies by higher amounts of damage, it means it will waste more damage and so it's real DPS decreases.

Final conclusions:

• weapons seem very balanced in Pillars of Eternity; if you see two weapons of same quality and think one of them is better in all situations - you probably don't understand all game mechanics

    

• weapons like arbalest, crossbow and firearms are designed with melee fighers in mind - the purpose of these weapons is to make a single strong attack at the start of the battle and then switch to another weapon, not wasting time to reload it

    

• similarily to above, your ranged characters could have one of these weapons to start a fight and then switch to a bow or a wand.

If you find it interesting, I may also write about my experience with pistols and blunderbusses.

• Faster weapons have higher dps against DR0 so accuracy generally benefits faster weapons more.
• Arbalest has prone on crits which might be worth more than the extra damage. Especially when the prone helps others deal more damage (sneak attack, lower deflection).
• Which weapon has higher dps depends on DR, accuracy vs deflection and damage boost. High damage boost benefits the war bow more. Of course there are talents and other buffs that can make one weapon better too.
• Arbalest benefits more from per encounter bonus damage abilities like the rogue ones.

Edited April 9, 2015 by Tvättbjörn

[Baki]

@Myrten: no, only Reload Time and Dex affect Reload Time.

[Myrten]

Could someone explain me what's the exact difference in frames between single and dual wield weapons and how does it depends on particular weapons?

[Baki]

Myrten, could you please explain demonjaxs findings exaclty? why is there only "attack + recovery mod"? i thought our earlier tests had shown that these are different things affecting different stats.

Edited April 9, 2015 by Baki

[tonberrian]

 

My damage calculator is marginally useful now! I haven't yet implemented attack speed, but you should be able to compare weapons in the same speed tiers fairly easily.

Instead of making separate chart you could make separate sheet on our chart and use our data

 

One, I'm not sure how to algorithm weapon set comparisons other than how I've built them on my sheet, so I'd still have to build like 8-10 more sheets on your sheet to get all the comparisons going, and I'd end up duplicating some data (and I'd want to tease apart some of your modifiers). It's easier to start from scratch than worry if I'm going to break something. Two, I actually am using your frame data, or at least I have it listed. I'm waiting on an explanation of what delay frames are where before I use it in calculations, though.

[Myrten]

Myrten, could you please explain demonjaxs findings exaclty? why is there only "attack + recovery mod"? i thought our earlier tests had shown that these are different things affecting different stats.

1)Dex affects Attack Frames, Recovery Frames and Reload frames by 1/(3%*(DEX-10)), Attack Delay and Recovery delay are not affected by anything andare both more or less 5 frames each.

2)Attack speed modifier and recovery speed modifiers both only affect Recovery Frames - they are additive and then multiplicative

3)Reload frames besides DEX is only affected by Reload speed.

[tonberrian]

Is there a reload delay? Or is there just an attack delay and a recovery delay?

[Myrten]

I have an important request:

 

Take a look at Source sheet here:

 

https://docs.google.com/spreadsheets/d/1vDm5MOrCK6S95h5u0EzZbGTv_u_rYqjuO0-zupqk0_A/edit#gid=1005560344

 

We need to fill\verify data - I've marked with red color data which is not verified or is missing so if you could please do this and mark verified cells with green it would be very helpful

 

 

Is there a reload delay? Or is there just an attack delay and a recovery delay?

Good questiom, we need someone to verify this.

[toser]

We need to fill\verify data - I've marked with red color data which is not verified or is missing so if you could please do this and mark verified cells with green it would be very helpful

 

Hatchet's and War Hammer's damage was wrong (11-16 and 14-20 instead 9-13 and 11-16). I also fixed the name of the "Hunters Bow" to "Hunting Bow".

 

http://i.imgur.com/1X6AxO3.png

http://i.imgur.com/5BdDUJm.png

http://i.imgur.com/6YaA748.png

[Myrten]

 

We need to fill\verify data - I've marked with red color data which is not verified or is missing so if you could please do this and mark verified cells with green it would be very helpful

 

Hatchet's and War Hammer's damage was wrong (11-16 and 14-20 instead 9-13 and 11-16). I also fixed the name of the "Hunters Bow" to "Hunting Bow".[/size]

 

http://i.imgur.com/1X6AxO3.png[/size]

http://i.imgur.com/5BdDUJm.png[/size]

http://i.imgur.com/6YaA748.png[/size]

 

Thanks, are you sure that everything else is correct? Because You've marked everything green on the left side

[Matt516]

Questions:

 

1) Is the spreadsheet currently in need of an Accuracy table that calculates effective accuracy damage modifier for any value of ACC-DEF?

 

2) Does the spreadsheet currently take chance-to-connect (AKA to graze, hit, or crit) into account when calculating effective DPS? Because if not, it should. Proper way to do that would be either "Damage = Base * (Accuracy Modifier + Chance-to-Connect * (All other modifiers))", or to make things a little simpler, "Damage = Base * Chance-to-Connect * (Accuracy Modifier* + All other modifiers)", where the asterisked Accuracy Modifier in this case refers to the effective modifier on any attack that connects (otherwise known as the base accuracy modifier divided by the chance-to-connect).

 

If the answer to either or both of the above is "no", let me know and I'd be happy to add them.

[Matt516]

Also, are we absolutely certain that damage modifiers such as graze, crit, Might, two-handed style, abilities, modals, etc.... are we certain that these, when added together, multiply by the base weapon damage for a mundane weapon of that type? As in, weapon quality and/or damage enchants are just another modifier? The reason I ask is that when you enchant weapons, the base weapon damage on the info sheet changes.

 

Also, is the modified damage on that info sheet just the damage per hit after all modifiers? Because if so, I think there's some stuff its leaving out. My Barbarian shows slightly higher damage on her 2H weapons in the inventory screen than is shown on their info screens. Maybe the modified damage on the weapon info sheet is only taking Might into account?

[Myrten]

Questions:

 

1) Is the spreadsheet currently in need of an Accuracy table that calculates effective accuracy damage modifier for any value of ACC-DEF?

 

2) Does the spreadsheet currently take chance-to-connect (AKA to graze, hit, or crit) into account when calculating effective DPS? Because if not, it should. Proper way to do that would be either "Damage = Base * (Accuracy Modifier + Chance-to-Connect * (All other modifiers))", or to make things a little simpler, "Damage = Base * Chance-to-Connect * (Accuracy Modifier* + All other modifiers)", where the asterisked Accuracy Modifier in this case refers to the effective modifier on any attack that connects (otherwise known as the base accuracy modifier divided by the chance-to-connect).

 

If the answer to either or both of the above is "no", let me know and I'd be happy to add them.

1)Well, we got options for -10,5,0,5,10,12 accuracy bonuses\maluses on weapon + global bonus\malus. So I think we got everything covered.

 

2)I'm not sure if I understand correctly what do you mean by chance-to-connect but I'm pretty sure it's correct - you can take a look at table starting at cell S12

 

Also, are we absolutely certain that damage modifiers such as graze, crit, Might, two-handed style, abilities, modals, etc.... are we certain that these, when added together, multiply by the base weapon damage for a mundane weapon of that type? As in, weapon quality and/or damage enchants are just another modifier? The reason I ask is that when you enchant weapons, the base weapon damage on the info sheet changes.

Base damage change is only graphical - damage bonus on weapon has the same effect as any other damage modifier, elemental damage on weapon on the other hand works different and formulas account for that. Edited April 9, 2015 by Myrten