46 replies to this topic

[Karl Streiger]

OK thanks again for your work. I have made an update...and found some mistakes in the code
Just some "Nerdtastic fun"


the overall precision of all weapons of all participants. (have to remove the Zero Bar) - you can't get any valuable information from this chart, so don't try it.

But if you have a good idea wich weapons should be compared (maybe Gauss; AC2; CUAC10; SPL; ERLarge) let me know

[xWiredx]

You running data analytics on this via python script or something?

I'd be interested to know the difference in accuracy stats between clan and IS weapons. Do lower burn times necessarily reflect in a higher accuracy number for IS lasers? Does the burst fire of clan ACs/UACs hurt their accuracy rating? Do the stream-fire clan LRMs have a worse accuracy compared to IS missile blobs?

How much damage per fire do clanners do compared to IS players per weapon? Maybe the most interesting few weapons only, since this one could be a longer process. Clan ERLL vs IS LL and ERLL. Clan UAC5 vs IS UAC5. Clan LRM 15 vs IS LRM 15. You get the idea here.

I'd also be interested to see if there is any weapon that people are more consistently accurate with, like a ranking of average accuracy of all the weapons. Something very general.

[wolf74]

sent you my Poor stats

[BabyCakes666]

easy to read graph man we need one of those

[Snowbluff]

xWiredx, on 17 June 2016 - 10:48 AM, said:

Does the burst fire of clan ACs/UACs hurt their accuracy rating? Do the stream-fire clan LRMs have a worse accuracy compared to IS missile blobs?

Looking at my own stats, no and not really. However, damage efficacy is lower.

[Karl Streiger]

Snowbluff, on 18 June 2016 - 07:33 AM, said:

Looking at my own stats, no and not really. However, damage efficacy is lower.

Yep the chance to hit something is better with the UAC20 but it doesn't deal accordingly damage.

I will prepare something maybe Monday - Tuesday
(direct comparison - Clan vs Is weapon (LPL, ER)
effect of dps weapons vs slow high alpha guns

[Karl Streiger]

another update (this time with readable charts)
tH aka toHit (the accuravy value for a weapon (you can read your in the stats)
tD aka to Damage (the effective damage dealt by a weapon system. calculation - damage dealt (stats) divided by max damage of a weapon multiplied with number of hits)
This last value is not very accurate because additional damage dealt by critical hits as well as reduced damage because of range drop off are part of the "damage dealt"


To read form upper left to lower right

• compariosn between AC/2 vs Gauss (does the increase RoF of the AC/2 has negative side effects - apparantly not
  
• C-ER Large Laser vs E-ERPPC what I had expected
  
• Clan vs IS
  
  • Large Pulse Laser - no real differences
    
  • while the UAC 10 hits more often the IS AC10 deals more damage
    
  • LRM and SRM are almost identical

[Tarogato]

Weapon stats are completely borked and I don't think you're going to get anything accurate out of them. Probably not worth the effort.

But sure, I'll pitch in.

https://docs.google....#gid=1520896091

[Widowmaker1981]

Seriously.. the weapon stats are displaying less than 10% of overall games played for whatever reason - if anyone thinks you can get any sort of even vaguely useful data out of that they are insane. complete waste of time.

[Karl Streiger]

Widowmaker1981, on 20 June 2016 - 02:21 AM, said:

Seriously.. the weapon stats are displaying less than 10% of overall games played for whatever reason - if anyone thinks you can get any sort of even vaguely useful data out of that they are insane. complete waste of time.

Again and Again and Again.... i don't give a damn about "overall" games - nor time equipped.
I only use:

• Shots Fired
  
• Shots Hit
  
• Damage Dealt (this is the only challenge)

And even when your stats are borked, they have to be as borked as those stats of all the other guys (>50 now).
Funny thing they must be bugged by the same margin.
Looking at the Gauss alone, there were several guys that had exactly the same accuracy with the Gauss Rifle down to the the first decimal number (afaik 67.1%)

Sometimes there are values off the chart (100% to Hit for example - but the numbers are low - for example I have 100% with the NARC but only 5shots fired)
and my final approach does only use the absolute sum of all "shots" fired - ignoring the individual.
All the later charts are only for plausibility - and they look valid for me

[Corrado]

sent.

[Wintersdark]

Karl Streiger, on 20 June 2016 - 02:46 AM, said:

Again and Again and Again.... i don't give a damn about "overall" games - nor time equipped.
I only use:

• Shots Fired
  
• Shots Hit
  
• Damage Dealt (this is the only challenge)

And even when your stats are borked, they have to be as borked as those stats of all the other guys (>50 now).
Funny thing they must be bugged by the same margin.
Looking at the Gauss alone, there were several guys that had exactly the same accuracy with the Gauss Rifle down to the the first decimal number (afaik 67.1%)

Sometimes there are values off the chart (100% to Hit for example - but the numbers are low - for example I have 100% with the NARC but only 5shots fired)
and my final approach does only use the absolute sum of all "shots" fired - ignoring the individual.
All the later charts are only for plausibility - and they look valid for me

Yup. I've said this before and will again. Stats may not include every game, but you can make a reasonable assumption that they're going to be vaguely accurate in terms of percentages, it's basically a random sampling of your matches.

While this definitely reduces their value for personal record keeping, it doesn't impede their usage to determine the sorts of things Karl wants to know.

[totgeboren]

Sent.
It's strange, I also have lower accuracy with the A C-LRM15 than without that A. Same with the C-LRM5. Seems I'm not the only one.

Edited by totgeboren, 20 June 2016 - 03:21 AM.

[Karl Streiger]

totgeboren, on 20 June 2016 - 03:20 AM, said:

Sent.
It's strange, I also have lower accuracy with the A C-LRM15 than without that A. Same with the C-LRM5. Seems I'm not the only one.

You are not the only one.... i will evaluate it next

[Widowmaker1981]

Wintersdark, on 20 June 2016 - 03:06 AM, said:

Yup. I've said this before and will again. Stats may not include every game, but you can make a reasonable assumption that they're going to be vaguely accurate in terms of percentages, it's basically a random sampling of your matches.

While this definitely reduces their value for personal record keeping, it doesn't impede their usage to determine the sorts of things Karl wants to know.

Thats assuming its random. I think its actually more likely to be not including any matches where quirks affect the weapon in question.. i.e. most of the actually good games you have with the weapon.. When the stats are missing this much, its not representative anymore. Its just gibberish.. but carry on.

[Bishop Steiner]

Widowmaker1981, on 20 June 2016 - 03:55 AM, said:

Thats assuming its random. I think its actually more likely to be not including any matches where quirks affect the weapon in question.. i.e. most of the actually good games you have with the weapon.. When the stats are missing this much, its not representative anymore. Its just gibberish.. but carry on.

since pretty much every match since the reset I have used an ac20 has been in my HBK-4G, I have to say that seems unlikely. Or I'd have no AC20 stats at all. I show 296 recorded HBK-4G matches..... 155 recorded ac20 matches.

So while there is definitely something wonky, Quirks do not appear to be the culprit (plus it never tracked weapons terribly accurately before quirks, either)

[totgeboren]

Widowmaker1981, on 20 June 2016 - 03:55 AM, said:

Thats assuming its random. I think its actually more likely to be not including any matches where quirks affect the weapon in question.. i.e. most of the actually good games you have with the weapon.. When the stats are missing this much, its not representative anymore. Its just gibberish.. but carry on.

But this issue of missing data was around long before the Quirkening.
Also, I have 3 registered games using an C-AC20 on mechs with no quirks relating to the c-uac20, with 16 shots fired and 14 hitting. That's both waaaay less games than I have used that gun in (it's at least 50 games), and my hit-rate with the U-AC20 is not even close to 90%, because I suck using it!

Lol got the c-ac20 and the c-uac20 mixed up. Only having 3 games using the c-ac20 sounds entirely reasonable, and by the looks of it I'm pretty accurate using it! The c-uac20 has 64 games played, with slightly lower accuracy (~80%), also entirely reasonable. So... hmmm..

Edited by totgeboren, 20 June 2016 - 05:29 AM.

[Karl Streiger]

So about the LRMs. Can't say for sure but Clan LRMs with Artemis doesn't seem to work right.
Artemis for Clan LRMs doesn't seem to be worth the money and tonnage.
About the damage - curious about those with rating below 100% was there a time when LRMs didn't deal 1 dmg - afaik since "release" they dealt 1 damage.

Name	Shots fired	Shots Hit	Accuracy	Delta	dmg total	dmg rel.
LRM 5	1,239,605	441,008	35.58%		465,500	105.55%
LRM 5 + ARTEMIS	417,055	154,534	37.05%	4.15%	168,171	108.82%
C-LRM 5	805,376	292,615	36.33%		288,327	98.53%
C-LRM 5 + ARTEMIS	141,378	47,951	33.92%	-6.65%	47,718	99.51%
LRM 10	2,515,406	889,367	35.36%		933,452	104.96%
LRM 10 + ARTEMIS	1,202,804	443,788	36.90%	4.35%	465,544	104.90%
C-LRM 10	862,944	331,932	38.47%		329,686	99.32%
C-LRM 10 + ARTEMIS	182,617	62,836	34.41%	-10.55%	61,829	98.40%
LRM 15	1,658,602	542,884	32.73%		568,958	104.80%
LRM 15 + ARTEMIS	2,579,558	926,931	35.93%	9.78%	979,698	105.69%
C-LRM 15	2,155,990	716,083	33.21%		725,897	101.37%
C-LRM 15 + ARTEMIS	790,208	269,036	34.05%	2.51%	270,806	100.66%
LRM 20	252,874	73,512	29.07%		85,195	115.89%
LRM 20 + ARTEMIS	1,306,120	429,460	32.88%	13.11%	467,106	108.77%
C-LRM 20	1,078,534	325,572	30.19%		333,164	102.33%
C-LRM 20 + ARTEMIS	241,865	73,496	30.39%	0.66%	72,911	99.20%

Considering the distribution it doesn't seem that bad anyhow. Judge for yourself



About the validation of data discussion: stats from 56 players so far

Edited by Karl Streiger, 20 June 2016 - 04:50 AM.

[Wintersdark]

My training grounds testing has confirmed that Artemis does practically nothing in most cases. People always argue it (Hurr! Training grounds mean nuthin! I use Artemis all the time and it's awesome! Bloody morons.) but comparing alrm15 and lrm15 (clan side) I find I need virtually the same amount of ammo to get a kill; if there is a difference, it's smaller than the normal deviation from trial to trial.


Edit: "does nothing" in this instance is in relation to Clan LRM grouping. Although Karl's data shows any added tracking doesn't appear to matter much either.

Edited by Wintersdark, 20 June 2016 - 05:09 AM.

[Widowmaker1981]

Wintersdark, on 20 June 2016 - 05:08 AM, said:

My training grounds testing has confirmed that Artemis does practically nothing in most cases. People always argue it (Hurr! Training grounds mean nuthin! I use Artemis all the time and it's awesome! Bloody morons.) but comparing alrm15 and lrm15 (clan side) I find I need virtually the same amount of ammo to get a kill; if there is a difference, it's smaller than the normal deviation from trial to trial.


Edit: "does nothing" in this instance is in relation to Clan LRM grouping. Although Karl's data shows any added tracking doesn't appear to matter much either.

Artemis has always been a waste on LRMs, because most of the time you want to be using indirect fire where it doesnt work, so the tonnage and slots arent worth it. If you are mostly using LRMs in a direct fire (LOS) mode, you should be using ACs or lasers in the first place, since they are enormously better...