50 réponses à ce sujet

[Egermano]

Update 4/30/2012:

I did a little research regarding player damage reduction (DR). What I found is that:

- DR from multiple sources stacks additively, including Justicar's Biotic Sphere
- 133% DR is needed for full damage immunity - see the formula
- Cover provides an additional 40% DR only if facing the cover - see my post on page 2
- There is no cap on DR
- Other variables, like total shield points, and shield bar standing effect the damage a little (+/-5%). Still exploring.

The DR formula I found is:
[Damage Suffered] = [BD] * 0.75 * (1 - [Total DR] / 100) + [BD] * 0.25

Where BD is base damage in points, Total DR is the sum of all DR effects in percentage. A Krogan Sentinel with 40% Tech Armor DR, 20% Rage DR standing in Biotic Sphere with 20% DR would have 80% Total DR. A Batarian with maxed out Blade Armor doing heavy melee while standing in maxed out Biotic Sphere has around 140% DR and is immune (tested, see social.bioware.com/forum/1/topic/343/index/11485213).

For reasoning and proof see my posts below.

Original post:

Initially, I wanted to test wether Krogan Vanguard Rage DR is bugged (it is, along with the melee bonus, see the appropriate threads), but in the end, I did a little more than that. I tested Krogan Vanguard's DR against my Krogan Sentinel's DR which happens to be specced for 40% DR from Tech Armor and 20% DR from Rage, resulting a nice, linear rise in additive DR: 0%, 20%, 40%, 60%. Multiplicative, Rage DR * Tech Armor DR would be 52%, but that's not the case, as it can be seen below. Damage source was a single Geth Rocket Trooper rocket on Bronze. I collected 3+ screenshots for every DR combination with Krogan Sentinel, and concluded that the damage source is appropriate, as it's a constant damage depending on DR, there's no randomness in shields lost (damage from rockets is reduced if the player is partially covered, but that's easy to avoid). Shield standings after a rocket hit:

Damage received scales linearly with additive DR:

My Sentienel's shield strength is 1200, so we can estimate the damage received from the fill percentage of the shield bar. Due to fraps downscale and the fancy shinyness there can be errors measuring the pixels. Shine is two pixels wide, fraps downscale is 2x, so I determined a four pixel error margin, which translates to approximately +/- 4% error in the following numbers:

0% DR: 763.63
20% DR: 647.27
40% DR: 538.18
60% DR: 421.81

The linear function f(x)=k*x+n in our case would be [Damage Received] = [Mitigable Damage]*[DR] + [Unmitigable Damage]. There is no pair of Mitigable Damage (MD) and Unmitigable Damage (UD) which solves all four equations, but there are a range of solutions wich are well in the error margin, like 570 for MD and 193 for UD. 

This doesn't prove anything yet, and I don't think the formula is this simple, but it's something to start with. It's more like that MD and UD are a percentage of a Base Damage (BD) value. For the sake of nice round numbers, in our case BD could be 800, then MD would be 75% of BD (600) and UD would be 25% of BD (200). Also, there might be a small, constant DR, depending on race/class - an armor value, just like armored mobs have. More precise data is required to determine these details.

It would be interesting to see how current system works for DR values over 100%. Stability and accuracy can overflow resulting a decrease in accuracy or negative (downwards) recoil. If DR overflows (and there is no check), MD*DR will be negative, and if that's larger than UD then total damage will be negative too, possibly healing the player. If there is a check for DR overflow, damage will be constant for increasing DR over 100%, and equal UD. If there is no check for DR overflow, but there is a check for negative damage, then damage will be 0. Currently 110% DR is possible with Krogan Sentinel and Biotic Sphere (if Biotic Sphere is additive).

Feel free to add your thoughts, or prove me wrong. Bioware staff is welcome to share the facts too

Modifié par Egermano, 29 avril 2012 - 10:10 .

[Wordlywisewiz]

even with just 40 precent from barrier and 40 precent from the bubble throw down a geth turret and your almost invincible

[peddroelm]

bump

Modifié par peddroelmz, 17 avril 2012 - 05:46 .

[Ben Gadura]

Nice thread! Just wanted to point out that you can edit your save in the singleplayer and give shepard all defensive powers.I believe the formulas are the same for both SP and MP, so stuff like this could be easier tested offline.

[Xaijin]

The difference being that DR has a flat ceiling, whereas Accuracy is actually simply a value change that can be pushed past its intended value due by affecting a linear value in a three dimensional space. I sincerely doubt you'll get healing from it

A damn shame it's additive as opposed to multiplicative, Krogans might have a better chance on Gold.

[Ravennus]

Nice work! Great information

[Autochthon]

Xaijin wrote...

The difference being that DR has a flat ceiling, whereas Accuracy is actually simply a value change that can be pushed past its intended value due by affecting a linear value in a three dimensional space. I sincerely doubt you'll get healing from it

A damn shame it's additive as opposed to multiplicative, Krogans might have a better chance on Gold.

If it wrere multiplicative krogans would take more damage.

Additve DR > Multiplicative. Read the post he says the numbers difference (multiplicative DR would be 8% less DR)

[Bioticgodx]

Nice. Very imformative.

[DaDiddles]

This is pretty interesting. From my play experience, I've always felt rather underwhelmed with damage reduction. I tend to die far more frequently from stunlock, and damage reduction really doesn't help you there. I notice very little difference in survivability with the krogan battlemaster when barriers are on versus off.

I'd be curious to see what the scale is for silver and gold, as I'm betting the "fixed damage" that can't be reduced is much greater on the higher difficulties, which would explain why Krogans are nearly indestructible on bronze, but by gold they tend to die almost as fast as any other race.

[Egermano]

Xaijin wrote...

The difference being that DR has a flat ceiling, whereas Accuracy is actually simply a value change that can be pushed past its intended value due by affecting a linear value in a three dimensional space. I sincerely doubt you'll get healing from it

Yeah, I doubt it too, but since accuracy and stability don't have overflow check we can assume other bugs too. Here's the reasoning behind the healing: if DR is given in percentage internally, it has to be converted before it's applied. The conversion is 1-(DR/100). For DRs above 100% that's a negative number. Applied to the damage the result will be negative too, because it involves multiplication, and if this result is high enough total damage can be negative too. There may be multiple checks to avoid this of course, but the system described in the OP doesn't prevent it automatically.

[molecularman]

Nice job! Curious how sphere would work here

[Egermano]

DaDiddles wrote...

This is pretty interesting. From my play experience, I've always felt rather underwhelmed with damage reduction. I tend to die far more frequently from stunlock, and damage reduction really doesn't help you there. I notice very little difference in survivability with the krogan battlemaster when barriers are on versus off.

DR doesn't work on its own, you have to be mobile. If you are in a melee spree and you can keep meleeing stuff without missing your chance of survival is very high. Miss once while surrounded you're in trouble. I like tank characters, as I like to be on my foot rather then in cover. DR helps a lot, but you have to know your limits. If all mobs are in melee range they won't shoot you as often. So you got to avoid situations where there are multiple distant mobs in los.

I'd be curious to see what the scale is for silver and gold, as I'm betting the "fixed damage" that can't be reduced is much greater on the higher difficulties, which would explain why Krogans are nearly indestructible on bronze, but by gold they tend to die almost as fast as any other race.

I did some testing on gold too, but it's very hard to get clean rocket shots while soloing... =) Geth Rocket Trooper rocket damages between 1500-1600, and I calculated the unmitigable portion around 400. Krogan Sentinel with 40% DR got damaged for 1083, Krogan Vanguard with 20% DR for 1300. That's bronze doubled.

[Phazael]

You can easily push the basic DR value over 100. Krogan with Tech Armor and rage standing under a Justicar Bubble does it. In my experience, they always take a minimum damage, even if you are at 100% DR and certain armor piercing attacks will bypass DR entirely. But, yeah a Raging Tech Armored Krogan standing under a defensive Biotic Bubble next to a recharging Geth Turret is damn near immortal.

[Egermano]

Phazael wrote...

certain armor piercing attacks will bypass DR entirely.

I think most of us would like to know what mobs have these piercing attacks. I didn't really pay attention, but now that you say it makes sense.

[Phazael]

I know for a fact that the Cerb snipers and Geth Prime rifle shots bypass it, so I would assume that anything else that counts as a sniper weapon on mobs would as well, which would basically include Ravegers. Also, the Banshee Warp attack seems to ignore about half the DR and the damage over time does not care about any defense you have. In fact, anything that works on a damage over time basis seems to not care at all about DR, which includes the acid pool from a dead Raveger. Strangely, pyro damage seems to get blocked by it. My Asari Justicar can get tripple flammed when she has bubble and reave up and tough it out a while, for example.

[Zenning]

Phazael wrote...

I know for a fact that the Cerb snipers and Geth Prime rifle shots bypass it, so I would assume that anything else that counts as a sniper weapon on mobs would as well, which would basically include Ravegers. Also, the Banshee Warp attack seems to ignore about half the DR and the damage over time does not care about any defense you have. In fact, anything that works on a damage over time basis seems to not care at all about DR, which includes the acid pool from a dead Raveger. Strangely, pyro damage seems to get blocked by it. My Asari Justicar can get tripple flammed when she has bubble and reave up and tough it out a while, for example.

Err.. Are you sure Nemesises and primes bypass Damage reduction?  I haven't done any straight testing, but based on experience this isn't the case.  

I'll test it to make sure later.

[Egermano]

Zenning wrote...

Err.. Are you sure Nemesises and primes bypass Damage reduction?  I haven't done any straight testing, but based on experience this isn't the case.  

I'll test it to make sure later.

Nemesis strips full Shield/Barrier in one shot on any difficulty. So yeah, that bypasses DR on shields. This is from memory, so I have to confirm it. I've tanked Primes with tank specced (Fitness +HP/Shields, Tech Armor max DR) Turian Sentinel, I firmly remember that full 4-shot series didn't deplete the shields on gold.

Edit: Nemesis doesn't skip DR either. It damages for a crapton of damage though. 1250 unmitigated damage on bronze (measured 1247.13). With Barrier 20% DR damage measured was 1063.44, which upscaled by the calculations in the OP (with 25% unmitigable BD) is 1251.35, again, well within error margin.

So the formula [Damage Received]=[75% Base Damage]*[DR]+[25% Base Damage] works on Nemesis' attack too.

I estimate that on Silver Nemesis damages for 1875, and 2500 on Gold (shield gates works). No wonder that it seems unmitigated. The damage is so high that it rips through most defenses.

Modifié par Egermano, 17 avril 2012 - 09:20 .

[Egermano]

I unlocked Asari Justicar so I could test Biotic Sphere (BS). My test was more thorough now, I made full resolution screenshots instead of half-resolution videos. That doesn't exclude errors unfortunately, as HP bar still isn't pixel-perfect because all those shiny effects, and even if it was pixel-perfect, rounding errors will always occur (my Justicar's 990 Barrier can't fit on 331 pixels).

Anyways, I made test with all nine BS (20%, 30%, 40%) plus Reave (15%, 25%, 40%) combinations, and what I found confirmed that DR stacks additively, and that damage suffered shrinks linearly with increasing DR.



The data I gathered fits the formula I found well:
[Damage Suffered] = [Base Damage] * 0.75 * max( 0, 1 - [Total DR] / 100) + [Base Damage] * 0.25



The table doesn't show that different BS+Reave combinations with same Total DR yield the same result.

The error in my calculations clearly rise towards high DR, that might be the sign of diminishing returns. Still, every result but the last at 80% DR is still within a reasonable error margin. Will have to do two-player tests with a Krogan Sentinel and Asari Justicar to be sure, but I'm leaving this project for a while.

[HEEGZ]

Thanks for this, very helpful.

[R0Gun13]

There was a topic a few days ago with vids. Someone took barbarian soldier with blade armour in a justicar bubble and did falcon punch at the same time as being hit by rockets. No health loss!

Im sure if u search u can find it, sorry on my phone and lazy ;-)

I think the person worked out that falcon punch was giving about 30-35% DR. To make 100% DR & taking 0 damage.

Modifié par R0Gun13, 20 avril 2012 - 06:22 .

[Axialbloom]

R0Gun13 wrote...

There was a topic a few days ago with vids. Someone took barbarian soldier with blade armour in a justicar bubble and did falcon punch at the same time as being hit by rockets. No health loss!

Im sure if u search u can find it, sorry on my phone and lazy ;-)

I think the person worked out that falcon punch was giving about 30-35% DR. To make 100% DR & taking 0 damage.

Seeing as that contradicts what the OP said, we really need that video.

[peddroelm]

Egermano wrote...

That doesn't exclude errors unfortunately, as HP bar still isn't pixel-perfect because all those shiny effects, and even if it was pixel-perfect, rounding errors will always occur (my Justicar's 990 Barrier can't fit on 331 pixels).

also - while the game keeps health and armor HP values in float format (high precision) it rounds shield//barrier HP values to integer after each damage calculation ...This obviosly will add some errors to your tests and on an unrelated note makes DOT effects not do consitant damage vs shields//barrier - due to all the rounding involved ...

Modifié par peddroelmz, 21 avril 2012 - 06:14 .

[TSCIGAR]

Re: nemesis and phantoms and DR, I know the phantom shots are affected, and I'm pretty sure the nemesis shot is too. I'll have to test tomorrow- I'm turning in now. Will report back- just wanted to have a comment so I can easily find the thread again.

[broshiggie]

Axialbloom wrote...

R0Gun13 wrote...

There was a topic a few days ago with vids. Someone took barbarian soldier with blade armour in a justicar bubble and did falcon punch at the same time as being hit by rockets. No health loss!

Im sure if u search u can find it, sorry on my phone and lazy ;-)

I think the person worked out that falcon punch was giving about 30-35% DR. To make 100% DR & taking 0 damage.

Seeing as that contradicts what the OP said, we really need that video.

http://social.biowar.../index/11485213

[Egermano]

I already tested Nemesis attack. It is affected by DR. It's a huge attack though, 1250 unmitigated on bronze (2500? on gold), so it will strip full shields even with high DR if not specced for Fitness HP bonus.

I will test Batarian heavy melee after Diablo beta is over BS is 40% DR, Blade Armor is 30% DR, so heavy melee needs at least 55% DR to reach 125% DR that is needed to bypass the 25% "unmitigable" part in my formula. If this is true, that means there is no check for negative damage in the formula (no max function to cap DR at 100%).

Modifié par Egermano, 21 avril 2012 - 04:45 .