Josh Sawyer's tweets and teasers, part III

[kmbogd]

Every dmg bonus/malus has an actual "step" associated to it.

 

 

Since damage bonuses are positive in nature their step is positive and equal to the percentage damage increase. Some examples of steps of dmg bonuses:

• Weapon specialization bonus(+10% dmg) => step=0.1
• Crit damage bonus (+25% dmg) => step=0.25
• Sneak attack bonus (+50% dmg) => step=0.5

 

In the case of maluses the formula for the step is a bit different: 1-1/(1-malus). Here are some examples of steps of dmg maluses:

• Graze malus (-50%) => step=1-1/(1-0.5)=1-1/0.5=1-2=-1
• Light under-penetration (-25%) => step=1-1/(1-0.25)=1-1/0.75=1-4/3=-0.333
• Severe under-penetration (-75%) => step=1-1/(1-0.75)=1-1/0.25=1-4=-3

 

Now, all the steps of the dmg bonus and maluses (except might) are summed up together.

 

If the sum is >=0 you have the formula that you mentioned: rolled_dmg * (Step_SUM + 1) * MightModifier. For example if the sum steps turns out to be 0.75 then your damage would be rolled_dmg (0.75+1) * MightModifier.

 

If the sum is <0 you have the following formula: rolled_dmg/(1-Step_SuM) * MightModifier. For example if the sum of the steps turns out to be -0.75 then yout damage would be rolled_dmg/(1-(-0.75)) * MightModifier = rolled_dmg/1.75 * MightModifier

 

 

 

Might has a step of it's own computed in the same way as for the other bonuses/maluses. If the the step is positive (at least 10 might): MightModifier=1+MightStep. If the step is negative (below 10 Might): MightModifier=1/(1-MightStep).

Edited May 4, 2018 by kmbogd

[Boeroer]

Yes, one has to know that with maluses like graze and such it's not like you lose 50% of your rolled base damage. You lose more as long as you have a positive dmg modifier after all (see above). Because of that some people might have thought that graze is "multiplicative" like MIG, which it isn't. It works like all the other "additive" dmg mods. The only exceptions are MIG and Power Level bonus.

 

The percentage-values with all maluses are totally misleading for the average player.

 

For the same reason blunted criticals (-25% crit damage) does not lead to a crit that does the same danage as a hit - a crit with a gun leads to less damage than with a hit. This shows that even the devs/desginers at obsidian don't really know how their mechanics work - at least not all of them.

The intention was that a crit with a gun should do the same damage as a hit. So somebody thought: "well if a crit does +25% then I only need to add a malus of -25% and the crit does the same damage as the hit..." - well - no. Now a crit with a gun can do +25% and -33% which leads to an overall crit damage of -8%. Bravo!

 

Again: very nice research of the damage formulars by the way guys.

Edited May 4, 2018 by Boeroer

[JerekKruger]

That seems like a really weird way of calculating damage penalties. Any idea why Obsidian went with it?

[dunehunter]

That seems like a really weird way of calculating damage penalties. Any idea why Obsidian went with it?

 

PoE(Path of Exile) style?

[Madscientist]

Thank you

 

This was the formula I did not know:

If the sum is <0 you have the following formula: rolled_dmg/(1-Step_SuM) * MightModifier.

 

So this means:

- If there is only a single modifier the effect is exactly what it says: sneak attack = +50% damage, graze = -50% damage

- If there are several bonusses and no penalties the bonusses are added: 50% sneak attack + 40% biting whip = +90% damage

- If there are several penalties and no bonusses the penalties have diminishing returns, which means that several penalties result in a bigger penalty than every single one of them but the total penalty is smaller than the sum of all penalties ( because otherwise you could have negative damage )

- If you have both bonusses and penalties, the penalties have a bigger wight than the bonusses

EDIT: The example I have gives was wrong.

 

result:

If you have several bonusses and penalties it is impossible to determine the effect of each modifier unless you are a math expert or you use a calculator. It is hard to tell if a specific combination of modifiers result in more or less damage than normal. The combat log will be confusing for many players because the can see +X% damage from something and -Y%damage from something else but the numbers do not add up.

 

I think the speed calculation has the same issues. I have to look there again.

So much about PoE2 mechanics are easier to understand than PoE1.

 

At some point I have suggested:

- keep the old formula from PoE1: final damage = base damage * ( 1 + sum of all modifiers ( including might) )

- If 1+sum is very small or negative you do a minimum damage that is X% of base damage.

 

result of my suggestion:

- Bonusses and penalties have the same wight

- You can easily determine the final modifier by adding all individual modifiers

- The bonus does always what the description says ( relative to base damage )

 

I cannot make or play a game with my suggestion so it may have downsides I do not know.

Edited May 4, 2018 by Madscientist

[Boeroer]

The downside? Way too easy!

[Madscientist]

I have found the ultimate example why the new damage formula is dumb.

Imagine you have a 1% bonus and a 50% penalty.

Now look at the sum of all modifiers: 1 + 0.01 - 1 = 0.01

 

With a 50% penalty and no bonus you do 50% damage (so far, so good)

With a 50% penalty and 1% bonus you do 1% damage

 

generally speaking:

Under certain conditions adding a bonus to a penalty can result in lower damage than if you had only the penalty.

Does anybody know another game where adding a bonus to something results in a drop of this something?

 

EDIT: Sorry, but this example is false.

Please ignore the stuff above

Edited May 4, 2018 by Madscientist

[JerekKruger]

Hence why I don't understand why they went with it. The old system worked fine as far as I can tell.

[dunehunter]

I have found the ultimate example why the new damage formula is dumb.

Imagine you have a 1% bonus and a 50% penalty.

Now look at the sum of all modifiers: 1 + 0.01 - 1 = 0.01

 

With a 50% penalty and no bonus you do 50% damage (so far, so good)

With a 50% penalty and 1% bonus you do 1% damage

 

generally speaking:

Under certain conditions adding a bonus to a penalty can result in lower damage than if you had only the penalty.

Does anybody know another game where adding a bonus to something results in a drop of this something?

 

As @kmbogd said, the fomula is:

 

If the sum is >=0 you have the formula that you mentioned: rolled_dmg * (Step_SUM + 1) * MightModifier. For example if the sum steps turns out to be 0.75 then your damage would be rolled_dmg (0.75+1) * MightModifier.

 

If the sum is <0 you have the following formula: rolled_dmg/(1-Step_SuM) * MightModifier. For example if the sum of the steps turns out to be -0.75 then yout damage would be rolled_dmg/(1-(-0.75)) * MightModifier = rolled_dmg/1.75 * MightModifier

 

Why do u get such conclusion? It is totally wrong and misguiding. In the case u raise: 1) 50 penalty, the fomula is rolled_dmg/1.5 = 0.67 * rolled_dmg; 2) 1% bonus and 50% penalty, your sum = -49% so u use the rolled_dmg/(1+0.49) = 0.67 * rolled_dmg;

Edited May 4, 2018 by dunehunter

[Insidous]

I have found the ultimate example why the new damage formula is dumb.

Imagine you have a 1% bonus and a 50% penalty.

Now look at the sum of all modifiers: 1 + 0.01 - 1 = 0.01

 

With a 50% penalty and no bonus you do 50% damage (so far, so good)

With a 50% penalty and 1% bonus you do 1% damage

 

generally speaking:

Under certain conditions adding a bonus to a penalty can result in lower damage than if you had only the penalty.

Does anybody know another game where adding a bonus to something results in a drop of this something?

 

That's not how it works.

 

With a 50% penalty and 1% bonus you would calculate the steps_sum to (-1) + 0.01 = -0.99

 

Since this is negative you apply rolled_dmg / (1-steps_sum ) = rolled_dmg / 1.99 = a factor of 0.5025

[Madscientist]

OK, looks like I did something wrong.

 

Sorry:

[morhilane]

Hence why I don't understand why they went with it. The old system worked fine as far as I can tell.

 

They probably didn't "went with it", some programmer got the job to refactor POE1 engine to the Deadfire one and probably screwed something up while moving stuff around.

 

POE1 also had damage calculation problems with multiplicative in beta (and a while after release). That tells me they aren't strong on unit testing their algorithms though (or their unit tests are crap which is much more likely).

[jf8350143]

Ashley Johnson is doing the narrator's voice during the start of the game. So that's why she only voiced a sidekick instead of a companion.

[Nail]

I agree.

 

If only one employee of Obsidian watched this forum they must have stumbled over this topic a dozen times.

 

Also BAdler answered a question of MaxQuest during beta1, explaining this behavior. So he def. knew.

I'm most certain they do. Big brother is watching...

[Insidous]

What I like about this system is when you have a double damage buff (+100%) and a half damage debuff (-50%) at the same time, they even out to normal damage as I think it's reasonable. It just makes the math in the combat log hard to follow at times

[SaruNi]

Lash, and particularly Bleak Walker lash, seems like more of an issue there than Might.

 

Bleak Walker has great defense and a huge multiplicative lash bonus at minimal zeal cost. It's unbalanced. 

 

But their defensive bonuses do depend on making the right dispositional choices... and perhaps those make the final game more difficult?

Edited May 4, 2018 by SaruNi

[Madscientist]

Lash, and particularly Bleak Walker lash, seems like more of an issue there than Might.

 

Bleak Walker has great defense and a huge multiplicative lash bonus at minimal zeal cost. It's unbalanced. 

 

But their defensive bonuses do depend on making the right dispositional choices... and perhaps those make the final game more difficult?

 

Why would depositions make the game more difficult?

Do you think there will be many situations where you cannot avoid benelovent or diplomatic choices?

 

Even if you chose the wrong deposition sometimes, all it does is lowering your defensis a bit. Unless you wish to tank things on PotD this should not be a big issue.

[SaruNi]

 

Lash, and particularly Bleak Walker lash, seems like more of an issue there than Might.

 

Bleak Walker has great defense and a huge multiplicative lash bonus at minimal zeal cost. It's unbalanced. 

 

But their defensive bonuses do depend on making the right dispositional choices... and perhaps those make the final game more difficult?

 

Why would depositions make the game more difficult?

Do you think there will be many situations where you cannot avoid benelovent or diplomatic choices?

 

 

 

Well, cruel actions will lower your reputation (as well as force you into fights that could be avoided). Perhaps some companions will leave or attack you. Perhaps you'll be forced to slaughter the entire village. Perhaps you'll miss out on some quests and XP. 

[Dr <3]

I just never understood why they made might bonus multiplicative in first istance.

 

Might was altready the most powerful stat.

in poe1 classes with intrinsec high dmg modifiers ( chiper, rogue) could choose to go with low might, since the dmg boost granted by might was kind of "diluited" by the other boni.

 

Now is the opposite: more dmg boost you have and greater becomes the dmg bonus given by might. And with the changes to the lash system this is even more evident ( the lash dmg is calculated after the might multiplication).

 

So, mutiplicative might benefits much more classes with lots of dmg bonus / lashes than the others. And in some cases ( es: bleack walker-rogue/ priest weapon-rogue/ chiper-rogue/ ecc) the dmg can real grow to excessive values.

 

For these reasons i'm in favor to revert might to give only additive dmg. So it will revert to be an useful stat for everyone instead then a very powerful stat for everyone and absolutely OP for some builds.

[dunehunter]

I just never understood why they made might bonus multiplicative in first istance.

 

Might was altready the most powerful stat.

in poe1 classes with intrinsec high dmg modifiers ( chiper, rogue) could choose to go with low might, since the dmg boost granted by might was kind of "diluited" by the other boni.

 

Now is the opposite: more dmg boost you have and greater becomes the dmg bonus given by might. And with the changes to the lash system this is even more evident ( the lash dmg is calculated after the might multiplication).

 

So, mutiplicative might benefits much more classes with lots of dmg bonus / lashes than the others. And in some cases ( es: bleack walker-rogue/ priest weapon-rogue/ chiper-rogue/ ecc) the dmg can real grow to excessive values.

 

For these reasons i'm in favor to revert might to give only additive dmg. So it will revert to be an useful stat for everyone instead then a very powerful stat for everyone and absolutely OP for some builds.

FoD is nerfed again in the game, it only gives u 25% lash. But for the problem of Might and Lash, a helwalker can stack 35 Might and +80% lash now, with all his autoattacks.

[George_Truman]

Every dmg bonus/malus has an actual "step" associated to it.

 

 

Since damage bonuses are positive in nature their step is positive and equal to the percentage damage increase. Some examples of steps of dmg bonuses:

• Weapon specialization bonus(+10% dmg) => step=0.1
• Crit damage bonus (+25% dmg) => step=0.25
• Sneak attack bonus (+50% dmg) => step=0.5

 

In the case of maluses the formula for the step is a bit different: 1-1/(1-malus). Here are some examples of steps of dmg maluses:

• Graze malus (-50%) => step=1-1/(1-0.5)=1-1/0.5=1-2=-1
• Light under-penetration (-25%) => step=1-1/(1-0.25)=1-1/0.75=1-4/3=-0.333
• Severe under-penetration (-75%) => step=1-1/(1-0.75)=1-1/0.25=1-4=-3

 

Now, all the steps of the dmg bonus and maluses (except might) are summed up together.

 

If the sum is >=0 you have the formula that you mentioned: rolled_dmg * (Step_SUM + 1) * MightModifier. For example if the sum steps turns out to be 0.75 then your damage would be rolled_dmg (0.75+1) * MightModifier.

 

If the sum is <0 you have the following formula: rolled_dmg/(1-Step_SuM) * MightModifier. For example if the sum of the steps turns out to be -0.75 then yout damage would be rolled_dmg/(1-(-0.75)) * MightModifier = rolled_dmg/1.75 * MightModifier

 

 

 

Might has a step of it's own computed in the same way as for the other bonuses/maluses. If the the step is positive (at least 10 might): MightModifier=1+MightStep. If the step is negative (below 10 Might): MightModifier=1/(1-MightStep).

 

See guys they didn't dumb things down after all!

[Enoch]

A few Tumblr posts today:

 

Two with old-to-us information on multi v. single classes.

 

One interesting one on difficulty:

 

 

 

Q:  Cohcarnage didnt reload on potd.Did you tuned down the difficulty?

 

A:  We didn’t tune down the gameplay effects of Path of the Damned, but I do think that sections of the critical path on PotD are not as challenging because the encounters aren’t as large and varied as they should be.

 

I take responsibility for it.  I didn’t want to prioritize PotD tuning over bugfixing because I think it’s preferable to have fewer bugs and easier PotD combat at launch than the opposite.  That said, there are side areas that are quite challenging on Veteran and PotD and there are sections of the critical path that are also challenging on those difficulties.  It’s just not consistent.

 

I’ve written up bugs for key areas that I think need to beefed up for an early patch.  They aren’t dramatic changes, but they do require testing, so I didn’t want to push for them prior to launch.  I apologize to anyone who is disappointed.  I think we can get to an appropriately consistent, challenging level of PotD difficulty with a modicum of work and testing.

[bleedthefreak]

I agree with his line of thinking on bug fixing vs. difficulty balance on PotD.

[bonarbill]

 

A few Tumblr posts today:

 

Two with old-to-us information on multi v. single classes.

 

One interesting one on difficulty:

 

 

 

Q:  Cohcarnage didnt reload on potd.Did you tuned down the difficulty?

 

A:  We didn’t tune down the gameplay effects of Path of the Damned, but I do think that sections of the critical path on PotD are not as challenging because the encounters aren’t as large and varied as they should be.

 

I take responsibility for it.  I didn’t want to prioritize PotD tuning over bugfixing because I think it’s preferable to have fewer bugs and easier PotD combat at launch than the opposite.  That said, there are side areas that are quite challenging on Veteran and PotD and there are sections of the critical path that are also challenging on those difficulties.  It’s just not consistent.

 

I’ve written up bugs for key areas that I think need to beefed up for an early patch.  They aren’t dramatic changes, but they do require testing, so I didn’t want to push for them prior to launch.  I apologize to anyone who is disappointed.  I think we can get to an appropriately consistent, challenging level of PotD difficulty with a modicum of work and testing.

 

 

Makes sense.  Rarely will people play the highest difficulty, but a lot of people will complain about bugs.

[The Sharmat]

Counterpoint: I play on the highest difficulty and don't care about other consumers.