126 replies to this topic

[MustrumRidcully]

Grus, on 12 June 2017 - 09:26 AM, said:

that mechanic was in back in closed beta and was wonderful. Valid tactic was to antisapate where the light mech was coming around and ram him to knock him over and then leg him. Now we have to just cal up into a wall to keep them from getting behind you, anf I hope you have arm mounted weapons because you can't aim down far enpugh with torso high mounts to hit the locus litterly in your crotch (Kodiak as reference)

It wasn't great,it was ridicilous, especally with all the teleportation bugs involved.

[Karl Streiger]

MustrumRidcully, on 13 June 2017 - 01:01 AM, said:

It wasn't great,it was ridicilous, especally with all the teleportation bugs involved.

It only warped while standing up - or slittering on the ground.... but you need the patience to wait until the mech stood on his legs again AND then fire all your guns.
But as usual people didn't stoped firing and so they were overheated when the light was back - or in several cases, were i rammed a Jenner with my Awesome - i had a open and orange CT afterwards - friendly destruction at its finest.

But its true - its absolute ridiculous (nice play of words) - than you can't hit that locust that stand 20m in front of you powered down because of overheating and you can do nothing with your Banshee or Atlas torso weapons.

Edited by Karl Streiger, 13 June 2017 - 01:08 AM.

[RestosIII]

If anything, I'd personally shrink a few light mechs. I'll admit that RIGHT NOW light mechs are the closest to balanced they've been in a long while, but they do NOT need any nerfing. If their opponent can actually aim, and doesn't have 200+ ping, lights are easy pickings. Especially when streaks get involved. And I'll say that 99.99% of the cases I've seen of hitreg complaints on lights in the last 1-2 years are just cases of pilots missing. They only get worked up over it because of the fact that a missed shot on a light is more painful, considering how little health light mechs actually have.

(Oh, and if anyone actually claims any of the 35 tonners needs nerfing, I don't understand you, but I pity you.)

[Shifty McSwift]

I might shrink them too if they weren't already having hitbox issues, I mean when they can happen on massive mechs why is it a surprise or secret that it also happens on tiny mechs?

Also saying that a particular weapon type like SSRM is what should be used vs them kind of just proves the point more, these heat seeking style weapons do full damage unless dodged entirely (and legitimately, in outrunning or putting something between etc), shredding these mechs that persist only because all the other weapon types have issues (beyond the lights speed and maneuverability) when hitting them on occasions. These kinds of pilots are getting free tankiness in most cases based on dodgy hit registration etc, and when they come up vs a weapon it doesn't apply to, suddenly they don't feel so tanky or skilled.

Damage can go where its not supposed to in the best of situations at times, it is just exacerbated by the super lights when they are moving at these super speeds and in most cases only taking a small percentage of the full damage of a weapon (barring direct hit PPFLD gauss ppc, and homing).

The design of some weapons is good in how you really shouldn't be using them in certain situations vs certain targets (high burn time vs high speed is a bad idea particularly at close range as one example), and in how some weapons just end up as powerful vs a certain mech type and average vs others. that is all well and good balancing etc. But that is distinctly separate from the issues of hitboxes and cross global server communications and the mechanics of hit registry, which granted, can never be perfect, but which the effects of should be minimalised rather than potentially abused or abusable.

[Aramuside]

JediPanther, on 12 June 2017 - 09:11 AM, said:

When an assault does exactly that or immediately starts backing up in reverse to the direction my light is going is a good indicator that assault has some experience with killing lights. That's when the smart(hopefully) light pilot runs off to find a new target or attack angle.

It also means you nullified that assault though as you can get back to the fight far more easily.

[Kiiyor]

Here, let's throw some more fuel on this raging bonfire, with some extremely flammable SCIENCE.



Not sure how much it will add to the light usage arguments here, but it's always nice to have some stats to twist to your advantage.

I've scraped some numbers from the leaderboards for each class. I sorted by matches played, and collected data from everyone that had played 100+ games, so that general performance would be more stable with less chance of any statistical outliers influencing the totals. I could probably go as low as 50 matches to avoid outliers, but I couldn't be bothered with all the copy/pasting. Some API's would be nice. I've also compared total overall performance against the top 50 for each class.

Not sure what the point is, since people rarely read past page 1 in a thread unless there's people quotin' them with fightin' words, but here goes:

Ooh, also, bear in mind that this contains data from the group and solo queues, as there's no way to seperate them on the leaderboards - and those queues play very differently.

Match numbers:



Heavies have the most representation, no surprise there. Lights are the few and the proud, it would appear.

Match score:



Lower damage, lower match score for the lights. Please take note of the scale; the numbers are (sort of) close, but the scale has been adjusted to show the difference with more clarity.

KD/R:

No statistical analysis in MWO would be complete without some highly subjective e-peen measuring.


KILL SECURED.

Very, very interesting. Those top light pilots certainly are top dogs.

The numbers above also don't show standard deviations, because I couldn't be bothered making a chart. The data I have shows that the gap between the top and bottom light pilots is far greater than those for other classes. Higher skill cap I guess.

Here's the numbers:

Standard deviations
(the higher the number, the higher the variance):

Lights:
Match score: 66.89
KD/R: 1.59

Mediums:
Match score: 65.14
KD/R: 0.87

Heavies:
Match score: 64.82
KD/R: 0.89

Assaults:
Match score: 72.05
KD/R: 0.93

Anecdotally, my gut is telling me that the kill numbers are higher because lights are usually the last to die in most matches. As for the disparity in kills/match score, i'd say it's because lights need to get close and make their damage count. Assaults and heavies do the armor stripping work, lights move in to slip a knife in the spine, and move on. Lights be nimble, lights be slick. Lights can kill and reposition quick.

Thoughts?

[HGAK47]

I think a lot of this has to do with latency and bad aim because many times I have been in a light and someone with good aim has smashed a leg or something off in a single easy volley.

I for one have no problem targeting light mechs even though my aim leave a bit to be desired. I will however say that I find PPC'ing lights at close range can be a pain.

Honestly go and use the gauss, AC/20, AC/10 hell even the PPC, that will teach you how to aim. "Easymode" lasers should be giving you no problems to stay on target.

And LOL at mechs that dont fit arm weapons and then moan that they dont have the same level of coverage when shooting. That is the price you pay for torso mounted weapons. Learn the advantages and disadvantages of torso vs arm weapons.

If you have arms make use of the damn things, peek around sharp corners and fire down from high places when possible (or up).

Edited by HGAK47, 13 June 2017 - 02:59 AM.

[Vordhosbn11]

Spr1ggan, on 12 June 2017 - 02:36 PM, said:

They do have a huge advantage, most assault pilots outside of comp are complete potato though. I've seen Kaffe kill/cripple a whole light wolfpack with his Dire.

Now THAT is something I'd like to see!

[UnofficialOperator]

Kiiyor, on 13 June 2017 - 02:49 AM, said:

Anecdotally, my gut is telling me that the kill numbers are higher because lights are usually the last to die in most matches. As for the disparity in kills/match score, i'd say it's because lights need to get close and make their damage count. Assaults and heavies do the armor stripping work, lights move in to slip a knife in the spine, and move on. Lights be nimble, lights be slick. Lights can kill and reposition quick.

Definitely agree with this.

Could you do 1 for WLR? Very interesting work.

Epeen: Just got ace of spades with artic cheetah

Edited by UnofficialOperator, 13 June 2017 - 04:44 AM.

[Kiiyor]

UnofficialOperator, on 13 June 2017 - 04:44 AM, said:

Definitely agree with this.

Could you do 1 for WLR? Very interesting work.

Epeen: Just got ace of spades with artic cheetah

I'll probably knock one up for most of the major stats tomorrow. It's too late here in the land down under at the moment.

I wish we could separate the data into individual mechs. I used to do it with my old SCIENCE stuff from donated EOM screens, but it's tedious as hell.

And congrats on the ace of spades It doesn't matter how much the planets align to give you kills, you don't get that without some great mech murder skills.

[Catra Lanis]

Decrease the speed and acc/deacc a little. I suspect the problem is that it's to fast for the netcode to handle, especially as it can start/stop almost instantaneously. I run a Raven that tops out at 146 and it's fast enough to get me out of most situations. Then a Locust topping out at the same speed should tbe able to do it too since it's a lot smaller.

Edited by Catra Lanis, 13 June 2017 - 05:18 AM.

[Jman5]

Kiiyor, on 13 June 2017 - 02:49 AM, said:

Here, let's throw some more fuel on this raging bonfire, with some extremely flammable SCIENCE.

...

Anecdotally, my gut is telling me that the kill numbers are higher because lights are usually the last to die in most matches. As for the disparity in kills/match score, i'd say it's because lights need to get close and make their damage count. Assaults and heavies do the armor stripping work, lights move in to slip a knife in the spine, and move on. Lights be nimble, lights be slick. Lights can kill and reposition quick.

Thoughts?

Is this Season 11 data?

Speaking of anecdotes, I went and took a look at my own stats. Since I'm not the kind of person who runs off and hides when he loses, By subtracting my deaths by my losses and dividing that by my wins, I could get a pretty close approximation to how often I die in games I win. (Deaths - Losses)/Wins = Percent of Deaths Per Win. For me, Lights were my best category meaning I died the least in games that we went on to win.

However when I looked at Season 10 stats overall my anecdote didn't seem pan out for the larger playerbase. They were virtually identical across weight classes (~38%). Then I decided to look at the top 100 players by matchscore in each category with at least 50 games played. Again, the numbers were very similar across the board, albeit much lower (~18%).

I would think that if Lights were dying last, which is what I always presumed, they would have significantly lower deaths per win. They would be less likely to be part of the 38% who typically die on the winning team. However they don't appear to be which is strange.

Some other stats:

Kills Per match and WLR
Assault: 0.76 WLR:1.03
Heavy: 0.69 WLR:1.02
Medium: 0.63 WLR:1.00
Light: 0.59 WLR:1.00

KPM (top 100)
A: 1.84 WLR:2.12
H: 1.66 WLR:2.37
M: 1.47 WLR:1.93
L: 1.59 WLR:1.99

By the way the Win loss ratio for top 100 heavies is interesting. My guess is that the nature of tonnage restrictions in the group queue means that heavy-centric groups perform best. Or best groups tend to play heavies. Another interesting point is that Lights surpass Mediums within the top 100 for Kills per match and WLR, but not in the overall playerbase.

Average Matchscore for top 100 players in each category mirrors the larger playerbase where Assaults are the highest, then heavies, then mediums and lights are the lowest.

Edited by Jman5, 13 June 2017 - 12:02 PM.

[Weeny Machine]

@Kiiyor: Nice statistics. Thanks for sharing. It actually seems to fit the amount of lights I see per average in matches.

The heavy mech statistics do not surprise me either. Heavies are still too agile as are some of the assaults. It is next to impossible to stay out of their firing arc because they turn like crazy

Catra Lanis, on 13 June 2017 - 05:18 AM, said:

Decrease the speed and acc/deacc a little.

Are you joking? That's the only thing that keeps lights alive. Heck, a WLF-1 turns like a truck and his accel is so bad that I just was wtfbbqed by a Battlemaster because I reversed and tried to get away from him. Amusingly, it felt like in a roadrunner cartoon. A second nothing happened than I started to move...and a second later I was dead. One of the fastest mechs with 156 km/h accelerates like a schoolbus. Good job, PGI!

What is worse, though, I ate only 2 alphas in that match. One from a heavy and the second from said Battlemaster. Note, the first one was at full speed of 156 km/h.

The last sentence is also directed at the "hitreg problem" crowd. The mechs which have dire hitreg problems are locust and commando. Things are happening there which cannot be explained with just "having a bad aim".
The other mech is the Javelin. They tank more than an Atlas. I gave up shooting their torsi and simply leg them. They die 4 times as fast then. The ACH shares to some degree this problem but its arms and legs pop off pretty easily. Torsi are a different matter, though.

The rest of the mechs, especially the 35t mechs, are pretty fragile in the current meta

Edited by Bush Hopper, 13 June 2017 - 01:04 PM.

[Fox With A Shotgun]

Bush Hopper, on 13 June 2017 - 12:56 PM, said:

@Kiiyor: Nice statistics. Thanks for sharing. It actually seems to fit the amount of lights I see per average in matches.

The heavy mech statistics do not surprise me either. Heavies are still too agile as are some of the assaults. It is next to impossible to stay out of their firing arc because they turn like crazy




Are you joking? That's the only thing that keeps lights alive. Heck, a WLF-1 turns like a truck and his accel is so bad that I just was wtfbbqed by a Battlemaster because I reversed and tried to get away from him. Amusingly, it felt like in a roadrunner cartoon. A second nothing happened than I started to move...and a second later I was dead. One of the fastest mechs with 156 km/h accelerates like a schoolbus. Good job, PGI!

What is worse, though, I ate only 2 alphas in that match. One from a heavy and the second from said Battlemaster. Note, the first one was at full speed of 156 km/h.

The last sentence is also directed at the "hitreg problem" crowd. The mechs which have dire hitreg problems are locust and commando. Things are happening there which cannot be explained with just "having a bad aim".
The other mech is the Javelin. They tank more than an Atlas. I gave up shooting their torsi and simply leg them. They die 4 times as fast then. The ACH shares to some degree this problem but its arms and legs pop off pretty easily. Torsi are a different matter, though.

The rest of the mechs, especially the 35t mechs, are pretty fragile in the current meta

Yup. The only ones with real issues are those which are scaled far too small for what they are (LCT, COM) and the ones with loosely grouped hitboxes with lots of gaps (SDR, ACH). These are throwing hitreg out the window, particularly when they move at higher speeds.

35-ton mechs have neither the agility nor size to throw off hitreg too much. Even 30 or even 25 tonners that are blob-shaped (KFX, ADR, MLX) are very easy to hit with both lasers and projectiles. It's just four specific mechs that really mess around with HSR thanks to either their geometry or size, combined with their high agility.

[UnofficialOperator]

We need to nerf lights more

[The6thMessenger]

I ******* hate Locusts, Commandos -- those small fast lights. While large mechs does all the work, these opportunistic ***** are just gonna ruin your game.

But honestly, I don't think that those fast mechs would be useful at all if anyone can just hit them with ease by being big, considering their paper-thin armor by comparison.

If there's something that should be done, hit-reg should be fixed. Hit-and-run with smart use of cover is one thing, but a locust running like crazy in the open barely getting hit is another.

[Old dirty B]

The6thMessenger, on 16 June 2017 - 12:20 AM, said:

...but a locust running like crazy in the open barely getting hit is another.

Now thats something indeed, a locust running around like crazy in the open barely getting hit, how terrifying that must be! The whole enemy team must have hold hands and prayed to the gods for help to escape and survive from this ferocious evil killing machine.

Locust or any small mech that run / fight like that are either very ineffective (the amount of focussed damage that can be outputted is pretty low), steal / secure kills critically damaged mechs are targeted - they already made big errors and should be punished) or get destroyed with less then 100 damage... You need to worry about those you cant see (and thus cant hit at all) ...

Anyway, i don't understand any of the complaining and wining about light mechs. Im sure any of the complainers have never piloted a locust or small mech for more then 100 games, probably less then 10 or none at all...
The queue for light mechs is often below 10, only a few dedicated, experienced / skilled players still run them and then its not that weird they are doing well.

Drop the Rambo, Leroy or Chicken behaviour and get into a tiny bit of teamwork (cover multiple angles / each other backs), that will render the Locust or small mech useless...

[Shifty McSwift]

Old dirty B, on 16 June 2017 - 01:14 AM, said:

Drop the Rambo, Leroy or Chicken behaviour and get into a tiny bit of teamwork (cover multiple angles / each other backs), that will render the Locust or small mech useless...

Uhm, except seeing a raven run the gauntlet through 8 enemy mechs legs, is pretty common, they can be so heavily shielded by ping/lag etc that they get away with it, rather than die or lose most of themself in the process.

Being able to scout/backstab/outpace/snipe, these are all fine, but being able to tank some 6 dudes by virtue of dodgy registration mechanics from being too fast/small whatever, I do not believe was ever their purpose, yet you see it constantly, even when teamwork/good aim/specific weapons are applied and they are killed off quickly for suicide gauntlet running.

If you want to mention rambos and leeroy jenkins, these guys are it man, in most cases they should never survive getting within 200m of half a dozen mechs by themself, but still you see them every second game or whatnot, trying to be rambo.

[Old dirty B]

Shifty McSwift, on 16 June 2017 - 01:32 AM, said:

Uhm, except seeing a raven run the gauntlet through 8 enemy mechs legs, is pretty common, they can be so heavily shielded by ping/lag etc that they get away with it, rather than die or lose most of themself in the process.

Oddly, i never encounter those "godmode" ravens, usually they go down pretty quickly (taking out their legs) or they aren't effective and hiding most of the time. Of course, there are a few dedicated Raven pilots that do super specially good. But most of the time, those players that all turn around to shoot the "squirrel" (or raven in this case) are - (lets put this friendly) - not very skilled / dont have good aim / dont have good tracking and obviously don't play as a team but as 8 individuals.
Perhaps lag or better said latency is a factor of importance, but thats situational and occasional at best. If this is a structural problem other mechs (also the big ones) benefit from data loss and so on. In other words, with latency problems, any mech can be influenced and "shielded" like that...

Apart from a few situational and occasional occurances, most common cause of these complaints is a skill gap between the light and other pilot. You can check the data up in this thread...

Edited by Old dirty B, 16 June 2017 - 01:52 AM.

[Aramuside]

If your comments about commandoes were true we'd have been using them heavily pre skill tree. Strangely we weren't - statistically they were barely used and strangely none of these mechs have suddenly changed geometry.

What did change is we suddenly could add a lot of survival nodes and your easy mode kills were gone.