Combat Metrics Bug Reports

Solinur

Gandork wrote: »

Yes, this is only on the PTS. There's some other really bizarre bugs on the PTS right now like potion duration being affected by your weapon rarity or the psijic shield stacking into the millions before crashing your game so maybe it'll get fixed with something else. Precise weapons currently show a 7000% crit tooltip, it should just be a display error but maybe stacking all these crit passives is messing up some background numbers. I am unfortunately rather ignorant on how addons work.

I'm going to install the PTS on my laptop tomorrow to see if I get the same problem on another computer. My laptop is a little quirky though since there's no official driver support for the GTX 970m I used to replace the 660m that was originally in it.

I finally see this issue too. Twin Blade and Blunt was required for me as well. I forgot that initially

Update: This indeed seems to be an API issue (CombatMetrics asks for the current Critical Strike Value) which takes unusually long (~33 ms) if this is done once, a less noticeable delay happens, if called multiple times in quick succession (in that case I've seen 4 calls by the current version of CMX) it causes a severe stutter. For now I mitigate it by doing only one such call within 100 ms. Also I'll file a bug report at EsoUI.

Solinur

So I now got to the bottom of this issue. Apparently the API function to determine the players Weapon Crit value gets slower the more modifiers are active.

In the case where we see issue its 7:

• Minor Savagery
• Major Savagery
• Perfect Strike (CP)
• Twin Blade and Blunt (Dual Wield)
• Dexterity (Medium Armor)
• Pressure Points (NB Assassination)
• Carnage (Khajit)

If none of them is present the function runs at 0.00024 ms (you don'T notice anything)
If all 7 are present it is 32 ms (you notice a small lag if it is called once, a severe on 3 or more calls)
If 6 of them are there it is about 5ms (you don't notice anything on a single call and maybe a small delay on more than 5)

Edit: I'll upload a Combat Metrics version that remedies this issue a bit.

Gandork

It's not visually choppy on my laptop (i7-3740QM) but my average FPS is dropping from a stable 82-84 to 64-78 when I pop a potion for major savagery so there still seems to be a performance impact; its just much less severe.
For both of those frame rates, I'm dropping refreshing path on a dummy then spamming relentless focus for lots of buffs/sec.

Edit: Oops I didn't see the new thread page before posting this.

Gandork

Glad you were able to figure this out, I was starting to doubt my own eyes. I did notice it a tiny bit on redguard with 6 modifiers but it was minor enough that I dismissed it as my imagination. I don't know why this bug seems to hit my desktop extra hard, the stuttering in the video I shared was just with surprise attack so you can imagine how much worse it was in an actual parse.

I just got done testing your new version on desktop and it's way better. It's still dropping my average FPS from 116-119 to 101-112 with much smaller stutters at a lower frequency but its definitely tolerable. In the same conditions with the previous version I was dropping to 40-70 FPS average with stuttering on every cast that frequently exceeded 200ms.

Would it be easy for me to tweak your 100ms limit to try to further mitigate the problem?

By the way, its awesome to see such great support for this amazing addon, you should get yourself a donation link

Solinur

Gandork wrote: »

Glad you were able to figure this out, I was starting to doubt my own eyes. I did notice it a tiny bit on redguard with 6 modifiers but it was minor enough that I dismissed it as my imagination. I don't know why this bug seems to hit my desktop extra hard, the stuttering in the video I shared was just with surprise attack so you can imagine how much worse it was in an actual parse.

I just got done testing your new version on desktop and it's way better. It's still dropping my average FPS from 116-119 to 101-112 with much smaller stutters at a lower frequency but its definitely tolerable. In the same conditions with the previous version I was dropping to 40-70 FPS average with stuttering on every cast that frequently exceeded 200ms.

Would it be easy for me to tweak your 100ms limit to try to further mitigate the problem?

By the way, its awesome to see such great support for this amazing addon, you should get yourself a donation link

You can easily edit the number. Look for the function GetNewStats in CombatMetrics/libs/LibCombat/LibCombat.lua

Let's hope ZOS can do something about this.

Update: ZOS is working on a fix

evoniee

I wanted to ask a question about critical resistance.
My toon have 749 crit resistance from cp, i tested it on random overland monster and i got 3.4% crit resistance, is it true?
I searching everywhere and people said critical resistance is 68 for 1% mitigation.

Solinur

ZOS is working on a fix .

evoniee wrote: »

I wanted to ask a question about critical resistance.
My toon have 749 crit resistance from cp, i tested it on random overland monster and i got 3.4% crit resistance, is it true?
I searching everywhere and people said critical resistance is 68 for 1% mitigation.

I haven't tested it to be honest. I might have just used the critical strike conversion (219 for lvl 50). Can you link any tests data in this direction?

Lightspeedflashb14_ESO

Solinur wrote: »

ZOS is working on a fix .

show previous quotes

evoniee wrote: »

I wanted to ask a question about critical resistance.
My toon have 749 crit resistance from cp, i tested it on random overland monster and i got 3.4% crit resistance, is it true?
I searching everywhere and people said critical resistance is 68 for 1% mitigation.

I haven't tested it to be honest. I might have just used the critical strike conversion (219 for lvl 50). Can you link any tests data in this direction?

https://forums.elderscrollsonline.com/en/discussion/279426/damage-mitigation-explanation-updated-for-morrowind-new-calculator/p1

CRITICAL RESISTANCE:

Now critical resistance works very differently from the other forms of resistance. What it does is it lowers the enemy players critical hit damage modifier. A characters base modifier is 1.5 which means that if you land a critical hit then you deal 50% more damage. There are ways to improve this, such as minor and major force or the Champion point system. What we are gonna be looking at of course is how we can reduce that modifier. This one is very special cause unlike mitigation that is a multiplicative reduction this one is actually a subtractive one. This is the formula:

CRITICAL MODIFIER=1.5+(Critical Damage Buff #1/100)+(Critical Damage Buff #2/100)+(etc etc)-(Critical Resistance/68/100)

DAMAGE TAKEN=Base Damage*(1.5+(Critical Damage Buff #1/100)+(Critical Damage Buff #2/100)+(etc etc)-(Critical Resistance/68/100))

If a player has base critical modifier and they hit an enemy player that has 100 points into Resistant then their modifier will be affected like this:

1.5-(1650/68/100)=1.5-0.24264705882~1.5-0.24=1.26

Finding out what the % per resistance was was not easy but from my testing I have found that it is 68(if someone else has a different number I would love to see it.). This means that 100 points into Resistant gives you ~24% critical hit resistance. The total amount of critical hit resistance possible is7518 or 110%. I now want to make something perfectly clear, THERE IS NO CRITICAL RESISTANCE HARD CAP! Ok? Ok! There isn't even enough critical resistance available to justify having it. Max resistance is 110% and highest possible critical damage modifier is 115%.

We have now gone through everything you need to know to calculate mitigation on any character you want or any situation you want. But before we get to the list of all the different sources there is one special ability I want to address cause the rules that it uses are not very straight forward, that ability is...

evoniee

Solinur wrote: »

ZOS is working on a fix .

show previous quotes

evoniee wrote: »

I wanted to ask a question about critical resistance.
My toon have 749 crit resistance from cp, i tested it on random overland monster and i got 3.4% crit resistance, is it true?
I searching everywhere and people said critical resistance is 68 for 1% mitigation.

I haven't tested it to be honest. I might have just used the critical strike conversion (219 for lvl 50). Can you link any tests data in this direction?

Ah, i dont save it because i thought its true.
Its easy to reproduce anyways. Sorry but im browsing on mobile right now. Later maybe i test it again.

Solinur

evoniee wrote: »

show previous quotes

Solinur wrote: »

ZOS is working on a fix .

evoniee wrote: »

I wanted to ask a question about critical resistance.
My toon have 749 crit resistance from cp, i tested it on random overland monster and i got 3.4% crit resistance, is it true?
I searching everywhere and people said critical resistance is 68 for 1% mitigation.

I haven't tested it to be honest. I might have just used the critical strike conversion (219 for lvl 50). Can you link any tests data in this direction?

Ah, i dont save it because i thought its true.
Its easy to reproduce anyways. Sorry but im browsing on mobile right now. Later maybe i test it again.

Anyway, I changed this now, but it might take a little, until this version gets published. I hope this is fine for now.

Gandork wrote: »

Glad you were able to figure this out, I was starting to doubt my own eyes. I did notice it a tiny bit on redguard with 6 modifiers but it was minor enough that I dismissed it as my imagination. I don't know why this bug seems to hit my desktop extra hard, the stuttering in the video I shared was just with surprise attack so you can imagine how much worse it was in an actual parse.

I just got done testing your new version on desktop and it's way better. It's still dropping my average FPS from 116-119 to 101-112 with much smaller stutters at a lower frequency but its definitely tolerable. In the same conditions with the previous version I was dropping to 40-70 FPS average with stuttering on every cast that frequently exceeded 200ms.

Would it be easy for me to tweak your 100ms limit to try to further mitigate the problem?

By the way, its awesome to see such great support for this amazing addon, you should get yourself a donation link

It seems to me like ZOS fixed the issue with this weeks PTS (the function runs at 0.016 ms). Should all be smooth once again

Altaryan

First, thank you for your amazing work on Combat Metrics.
I had a question for you is there any way to export the detailed log of a fight ?
I explored the combat metrics data Lua files but it seems it is not there. I would like to open a detailed log in Excel format further analysis.

visionality

Me and my friends are having issues with group damage/group healings numbers. The abolute numbers only slightly vary from person to person (mainly because the fight starts a bit differently for everyone), but the personal amount of dps or hps does not fit in at all

Typical example:
DD 1 reports (and posts) doing 50% of group damage - also showing in his absolute numbers
DD2 reports (and posts) doing 50% of group damage - also showing in his absolute numbers
Healer reoprts (and posts) doing 25% of group damage - also showing in his absolute numbers

Group damage is ofc not 125%, so some 25% of damage are lost in the group calculations.
Same problem with healing (my healer alone very often does 25k hps connected to 23k group hps, which feels a bit absurd)

Is this a known issue? Is there a fix for it or a way to avoid it?

Solinur

Altaryan wrote: »

First, thank you for your amazing work on Combat Metrics.
I had a question for you is there any way to export the detailed log of a fight ?
I explored the combat metrics data Lua files but it seems it is not there. I would like to open a detailed log in Excel format further analysis.

Thats not an easy thing to achieve.
Right now the log is encoded using my own base64 encoder. It would generally be possible to decode this but then you still only have lines with numbers in them which are unit and ability IDs. From there one would need a script that looks up the abilities and so on and transforms the data from lua to an excel or csv sheet and then do basically the same stuff that happens within combat metrics in lua.

Since I don't want to spend time on this myself someone else with scripting or programming abilities would be needed to do this. I would however offer assistance if anybody wants to take this on.

visionality wrote: »

Me and my friends are having issues with group damage/group healings numbers. The abolute numbers only slightly vary from person to person (mainly because the fight starts a bit differently for everyone), but the personal amount of dps or hps does not fit in at all

Typical example:
DD 1 reports (and posts) doing 50% of group damage - also showing in his absolute numbers
DD2 reports (and posts) doing 50% of group damage - also showing in his absolute numbers
Healer reoprts (and posts) doing 25% of group damage - also showing in his absolute numbers

Group damage is ofc not 125%, so some 25% of damage are lost in the group calculations.
Same problem with healing (my healer alone very often does 25k hps connected to 23k group hps, which feels a bit absurd)

Is this a known issue? Is there a fix for it or a way to avoid it?

This might be something that I need to look into. I believe the code generally works but at least one issue applies:
You only generate data on mobs that are known to you. This means you either hit or debuffed them or they hit you. So if the tanks didn't hit every little trash mob somehow, that units are unknown to him and the % is not attributing to the whole group but only to those mob he "knows". What you should look at in this case is the number of enemies, and the total group damage that was done to them (can be seen in the main report panel) and see if you have equal values there.
Another issue happens when you are out of range (like in vAS where you are out of range to St. Olms when you are on the opposite side of the room or in vCR when you go to the shadow world) in those cases you client doesn#t receive the "group" data anymore.

Lightspeedflashb14_ESO

so i got something, could you explain what is going on in this video, prolly ought to mute because the music is really loud and bad.

https://youtu.be/9I0XE-2gsi0?t=637


in it combat metrix says that he has penetrated with a maul two hander at max, 12441, with minor fracture as the only debuff. when he was using a sword or ace, he got a max of 10202. meaning that the maul added a total of 2239 pen to his build. this makes no since. it ought to have been 3356, as 18,200-1320 = 16880 and 16780 *.2 is 3376.

am i totally off base here with my math?

Solinur

Lightspeedflashb14_ESO wrote: »

so i got something, could you explain what is going on in this video, prolly ought to mute because the music is really loud and bad.

https://youtu.be/9I0XE-2gsi0?t=637


in it combat metrix says that he has penetrated with a maul two hander at max, 12441, with minor fracture as the only debuff. when he was using a sword or ace, he got a max of 10202. meaning that the maul added a total of 2239 pen to his build. this makes no since. it ought to have been 3356, as 18,200-1320 = 16880 and 16780 *.2 is 3376.

am i totally off base here with my math?

This is weird. I'm pretty sure I didn't put tracking of this and the dual wield version in there. so unless the penetration value from the char sheet is affected, it shouldn't be tracked at all.
Your math seems fine to me. I'll look into it.

Edit: That's actually what happens. For some reason the penetration value of the player changes. I don't get why this would happen at all.
For me it only changes for 269 on a almost naked template char on PTS but this is really very weird.

Edit2: Could it be that once maces had a flat penetration increase? After all that penetration stat is not on the vanilla char sheet it's only available via API. Maybe the calculation there is wrong. I guess I could work out the effect of the Maces and Mauls, but fixing the base stat seems like a lot of work (there are quite a few sources of Penetration). Similarly Penetrating Magic from Destro Staffs is also not supported by CMX.

Edit3: It seems to be just a display bug. The value that is shown there scales with the physical resistance of the player (its actually 20% of the players physical resistance). I'll file a bug so this gets fixed in the API.

hashield

install the latest version but so far it does not work says it needs to update "combat metric fight data"

Solinur

hashield wrote: »

install the latest version but so far it does not work says it needs to update "combat metric fight data"

I pushed an update to address the issue.
That aside you can always select to load outdated addons in the addon selections. Since this only refers to whether an author adjusted a number in the text file, having an addon shown out-of-date is no guarantee it doesn't work. Reversely even when an addon is shown as up-to-date there is no guarantee that it works as intended.

R0ckster

As I relogged to another character, all my saved fights completely disappeared. They disappeared on all characters, almost as if the addon reset itself while relogging.
I didn't delete anything or reinstall the addon, i simply relogged. Is there a way to recover my fight history or find out what happened?