30 replies to this topic

[krolmir]

Ever have one of those matches where it seemed like your damage was just off? I'm not talking at the end of the round though; mid-brawl, a knuckle head stands still in front of you for like 30 plus seconds, and you carefully line up repeated shots; the paper doll registers a hits, however the color doesn't change and you just laid down 20 damage with 2 ERPPCs? Well its true, it didn't register, but not for the reasons you may think (HSR).

Today, after a match where I clearly should have obliterated and Orions CT, watched his paper doll register repeated hits to the CT, but no change in colors; I thought to myself what if ERPPC's are bugged? So I dropped in the testing grounds and this is what I found.

First mech I shot was the AWS-8Q. as a gauge I shot it with my ERLL first, which caused 2% damage. I then Shot it with my 2 ERPPC's which caused 2% damage. Whiskey Tango Foxtrot? I shot the mech repeatedly, at different ranges, and got the same result. The next mech I shot was the CTF 1X, for the most part the numbers made sense. Then next mech I shot was the CN9-A, it didn't take damage normally. With the Single ERLL doing 3% and 2 ERPPC's doing only 4%. The JR7-D was next, it took damage appropriately. After the Jenner was the AS7-D, damage was nominal as well. The CPLT-A1 did not take damage correctly, with the 1 ERLL doing 2% damage versus the 2 ERPPC's 3% on average. The CDA 2-A was next, the damage was within expected ranges, however the damage from the ERPPC's seemed 1-2% lower than it should have been. Last up was the CMD-1B which also took damage from the ERPPC's that was about 2% lower than it should have been. With the ERLL posting 5% damage and the 2 ERPPC's posting 9% damage done.

Math breakdown: ERLL = 9 damage, ERPPC = 10x2 = 20 damage
For testing purposes the ERLL was the constant.
I post my screenshots backwards so start at the bottom and work your way up, my bad?

Edited by krolmir, 24 November 2013 - 07:25 AM.

[krolmir]

How many chassis and variants of which are affected in game is hard to say given the extremely limited nature of our testing grounds. PGI needs to look into this fast.

[JohnnyWayne]

The displayed %age is seemingly a rounded value, therefore the same damage differes from shot to shot.

[AdamBaines]

Just a quick question. How are you firing the 2 ERPCs? Are you firing them at the same time? Or one by one? Do you see a difference when firing them together vs. 1 at a time?

[Rhys Erlykov]

Regardless, more data is needed. Set up the variables as close together as you can for each mech. Shoot the same spot each time. Shoot with one LL, then one ERPPC. Then 2 LLs, then 2 ERPPCs. Do this many times, build a table of data, then present it to us.

Then sacrifice a goat to Odin in the hopes that a PGI developer reads the results (if they show that something is indeed amiss).

[stjobe]

Just for the Awesome, 100% health is 494+237=741 health points. 1% is 7.41 damage.

Now the damage display shows only integer values, so 99% health could be anything from 493.9 to 486.6, depending on how it rounds the numbers (I'd say it rounds down, since I often get to 99% just from being brushed by other 'mechs in the starting jostle).

A fresh AWS getting 8 points of damage will be at 98% (741 - 8) / 741 = 0.989
A fresh AWS getting 20 points of damage will be at 97% (741 - 20) / 741 = 0.973
A fresh AWS getting hit by 8+20 damage will be at 96% (741-28) / 741 = 0.962

So yes, it looks like the dual PPCs do the same percentage damage that the LL does, but that's because you do just above the damage needed for the percentage to reduce twice with the LL (8 damage vs 7.41 leaves 6.82 damage before it reduces again), and just below what is needed for it to reduce thrice with the PPCs (20 damage vs 22.23 is 2.23 damage short).

In short, it's probably not a bug with the weapons, but just the health display rounding the numbers.

Edited by stjobe, 24 November 2013 - 08:10 AM.

[Rhys Erlykov]

Thanks for the math breakdown, Jobe.

[krolmir]

I fired them 2 at a time because at 20 damage is should be double + the damage of 1 ER or LL, it makes for an effective gauge. The numbers I posted are the average numbers that I have attained. I noticed that on mechs where I would get irregular percentages, that both sets of numbers would sway the same direction. EX. so if the ERLL damage was 1% the first round and the ERPPC's was at 2% the next round, if the ERLL went to 2% the ERPPC's when fired whould also jump 1% to 3% total. This is why I when test I always fired in the same order ERLL, then ERPPc's. This is also why all my screen shots are sequential, helps to show the actual proportions.

JohnnyWayne, on 24 November 2013 - 07:33 AM, said:

[size=4]post='2941772']
The displayed %age is seemingly a rounded value, therefore the same damage differes from shot to shot.

Well that still doesn't explain the AWS-8Q, it always showed the same damage 2% vs. 2% 1 ERLL vs. 2 ERPPC's; or the Centurion, Catapult, etc.

[krolmir]

I test fired on each mech not just once, but repeatedly. If i need to make a table I shall.
The ERLL does 9 points of damage which is 45% of the damage of 2 ERPPC's 20 damage.
It could be the health display rounding but I could understand a 1% either way, but 2% or more off is a bit odd....

Edited by krolmir, 24 November 2013 - 08:21 AM.

[krolmir]

Well, another way to look at the math is take total health pool 741/9 = 82.3 repeating for ERLL, this gives you the number of shots until complete destruction; for the two ERPPC's 741/20 = 37.05. 82.3/37.05 = .45 This is the same ratio you get when you divide the ERLL's 9 damage by that of 2 ERPPC. Now to get to the AWS point at which the health display would round up or down requires at least 11.115 damage (7.41x1.5). No matter what it points to an issue with either the damage application or the heath indicator. I will take an AC/20 mech into the testing grounds to verify displayed percentages.

[stjobe]

krolmir, on 24 November 2013 - 08:17 AM, said:

It could be the health display rounding but I could understand a 1% either way, but 2% or more off is a bit odd....

It's just the way the math works out when you alternate shots that make it seem like that:



First column pair is the health progression of an Awesome getting hit by a single LL, second column pair is getting hit by a pair of PPCs, and the third is the interesting one, getting hit by alternating single LL and pair of PPCs.

If you drop the part after the decimal point you get what you see on the health display - until the seventh shot it does seem that the single LL does the same 2% as the dual PPCs, but as I said that's just an artefact of the rounding down.

[krolmir]

Now this is interesting. Took my Orion-k with AC/20 out and only shot the AC/20. The results are very interesting....


So the AC/20 registers 3% on everyshot every time, but its only 20 damage same as 2 ERPPC's? Only on the 4th AC/20 shot does it register as 2%, on the 5th it goes up to 3% for another 3 shots, repeating.

Edited by krolmir, 24 November 2013 - 09:11 AM.

[stjobe]

krolmir, on 24 November 2013 - 09:07 AM, said:

Now this is interesting. Took my Orion-k with AC/20 out and only shot the AC/20. The results are very interesting....

[snipped pictures]

So the AC/20 registers 3% on everyshot every time, but its only 20 damage same as 2 ERPPC's?

See the middle column (2xPPC) in my post above yours.

It's math. Percentages and rounding.

There's no bug.

[Corvus Antaka]

testing grounds can not be trusted.

[Tolkien]

I'm glad to see someone looking at this, since there certainly are hitbox issues that keep slipping through QA. With the centurion, be aware that shooting them high up can lead to shots just disappearing:
http://mwomercs.com/...s-bug-verified/

[krolmir]

Well the math doesn't line up either, by your chart, every 7th shot should drop the LL to 1% yet ingame in testing grounds the first shot takes 2%, the next 3 do 1%, followed by another 2% repeating.\

Either way something is amiss.

Edited by krolmir, 24 November 2013 - 09:31 AM.

[stjobe]

krolmir, on 24 November 2013 - 09:30 AM, said:

Well the math doesn't line up either, by your chart, every 7th shot should drop the LL to 1% yet ingame in testing grounds the first shot takes 2%, the next 3 do 1%, followed by another 2% repeating.\

Either way something is amiss.

Possibly because I used maxed-out armour values and the AWS in the testing grounds is stock (as in it doesn't have max armour). I'd do another chart with the stock AWS-8Q but I can't be arsed. It's pretty clear to me that there is no bug and that you just don't understand how the health display works.

[krolmir]

Went back in with my Orion K with the AC/20, only this time fired the 4 medium backup lasers. Shot the AWS-8Q on the same map, once again doing 20 damage per shot. This time it to 97%, 94%, 91%, 88%, then the kicker 84%?

I understand that the health display can and will round numbers, the parameters may be extemely sloppy, or their is another issue.

Edited by krolmir, 24 November 2013 - 10:00 AM.

[krolmir]

Colonel Pada Vinson, on 24 November 2013 - 09:11 AM, said:

testing grounds can not be trusted.

[Tolkien]

krolmir, on 24 November 2013 - 10:00 AM, said:

What fraction of anomalies found on the testing grounds would you attribute to errors in the testing grounds (as opposed to errors in the game+testing grounds)?