81 replies to this topic

[Strum Wealh]

MrMadguy, on 29 February 2016 - 01:00 AM, said:

Yeah, Spiders and Commandos - are two, that definitely need to be rescaled up, as long,a as other Light 'Mechs. But PGI isn't even talking about rescaling some 'Mech up - only about rescaling worst ones down. What we really need - is size/wight normalization!

Actually, they are going to do the weight-to-size normalization.

As an example, Russ has specifically mentioned that the Crab is actually too small, and could/would need to be scaled up by ~5% to bring it into line with their standardized volumetric scaling (stated at 1:53:53 to 1:54:04 of the Jan. 30th Town Hall).
IMO, the same stance would likely apply to many of the Light 'Mechs, and it is (IMO) not-unlikely that many of those would be scaled up as well, probably significantly so.
Russ also stated that they would be looking at a tolerance level of (possibly) within ±1-2% of the "correct" volume for a given weight increment, and that he expects upwards of half of the chassis then in-game (at the time, 64 chassis) needing to be rescaled.

Admittedly, I haven't had a chance to go through the most-recent Town Hall, so any related information from that one (with time stamps!) would be appreciated.

Edited by Strum Wealh, 29 February 2016 - 08:11 AM.

[CDLord HHGD]

QuulDrah, on 21 January 2016 - 08:09 AM, said:

Only if they have the same density...

Gotta pick a constant somewhere.....

[MrMadguy]

Strum Wealh, on 29 February 2016 - 08:08 AM, said:

Actually, they are going to do the weight-to-size normalization.

As an example, Russ has specifically mentioned that the Crab is actually too small, and could/would need to be scaled up by ~5% to bring it into line with their standardized volumetric scaling (stated at 1:53:53 to 1:54:04 of the Jan. 30th Town Hall).
IMO, the same stance would likely apply to many of the Light 'Mechs, and it is (IMO) not-unlikely that many of those would be scaled up as well, probably significantly so.
Russ also stated that they would be looking at a tolerance level of (possibly) within ±1-2% of the "correct" volume for a given weight increment, and that he expects upwards of half of the chassis then in-game (at the time, 64 chassis) needing to be rescaled.

Admittedly, I haven't had a chance to go through the most-recent Town Hall, so any related information from that one (with time stamps!) would be appreciated.

Would be good, if game's whole life time offenders would finally be fixed, but I really doubt, that it will happen. Lights are untouchable.

[mogs01gt]

Why is this thread not stickied yet!!!!

[zagibu]

I've updated the zip with the profiles and data of the Archer. It's even a bit smaller than his latest brothers, sitting at position 17/61 in the list, with PPTA values of 97.55% and 95.28% compared to mean and median, respectively.

[Gryphorim]

The data shown here is interesting, it even seems to come close to the values PGI use, but this measure of PPTA is a measure of area in front and side projections, apparently ignoring top projection entirely, whereas PGI are using a constant value for density and scaling volume accordingly. Is this data shown here going to be updated to use volume for calculation? Or at least include top projection for it's math?

[zagibu]

I used to think that about this method as well, and actually tried to find ways to do volume comparisons. There is a way of doing it, by importing the model into zbrush and approximating a new quickmesh to it that would give pretty good results, and I even started doing it, but then I realized something.

It's true that volume is a much better value than this PPTA to compare general mech "size". But this general mech "size" also doesn't matter much at all. It's actually less important than the profiles, because the profiles are what players are shooting at. And it's mostly front and to a slightly lesser degree, side profiles, because common engagement ranges leave only little room for vertical angle variance, which is why top profile doesn't matter much at all.

Now I'm not saying that the profiles are actually a good method of comparing mechs, because hitbox size and shape is completely lost in them (as it is in a global volume measurement), which means that two mechs with the same PPTA values can still have completely different survivabilities, even when they should take about the same amount of hits, because WHERE you take the hits is very important. Is it an arm? Or your CT? The profiles can't tell you that easily.

So ideally, we could make projected profiles of a mech's hitbox mesh (with each hitbox in a different color) from about 6-12 different angles around the Z axis and count visible pixels of all the hitboxes and THEN calculate the same PPTA values for all hitboxes. Heffay was working on extracting the hitboxes from the files, and although it looked promising at first, it turned out to be some other data IIRC.

So, of all available methods of comparing a "size" of mechs that's actually somewhat relevant to gameplay, I still think the profile area normalized over tonnage is the best one.

Edited by zagibu, 16 March 2016 - 03:12 PM.

[zagibu]

BTW: Omitting the top profile is the reason why the crab has such a small PPTA value. It has quite a large "top down" profile. But since you rarely shoot at it, we can safely discard it without distorting the data much.

Maybe it does matter more than I give it credit, because of LRMs. I guess we could factor it in with a lower weight than the other two profiles, but I doubt it would change things much. There aren't that many mechs with an unfortunate top down profile. But maybe i'll give it a try some day anyway.

Edited by zagibu, 16 March 2016 - 03:22 PM.

[Gryphorim]

It seems as though PGI are going for a "x tonnes and a fixed density value = given volume" which is scientifically accurate, whereas your method prefers gameplay impact to volumetrically accurate scale. Whilst neither method is wrong per se, I personally prefer PGI's method.

As for the impact of top-down profile, it makes a big difference to how much you can tank LRMs, and there are quite a few mechs that have big top-down profiles. Many of the long-bodied designs such as Crab, Marauder, Stalker, KingCrab, etc.

[zagibu]

Gryphorim, on 16 March 2016 - 04:30 PM, said:

It seems as though PGI are going for a "x tonnes and a fixed density value = given volume" which is scientifically accurate, whereas your method prefers gameplay impact to volumetrically accurate scale. Whilst neither method is wrong per se, I personally prefer PGI's method.

I would prefer it, too, if correct size had any meaning in the game. But it doesn't. I guess it's nice for the people who like "realism" in their games, but otherwise it's completely wasted effort.

[mogs01gt]

bump to keep this on the first page!

[m]

Spr1ggan, on 18 November 2015 - 06:39 AM, said:

The scale in game is weird dude, iirc there were threads a while back showing the issues of fitting someone in the commando cockpit. That's why i brought it up.

*edit* Found what i was looking for.
http://imgur.com/a/bvxiD#16

Look at the Commando one, would be pretty damn cramped, no?

This is what you are referring to:



Many people have tried laying claim as being the originator of the video (even in the international section) in these forums and on youtube. This is the original source.

(here are some others from the same user)







As for volumetric scaling (just skimming the comments and adding my two cents), I am pretty sure the reason for that method is to stay consistent with the volumetric attributes in the mech models. I say this because I remember some time ago, during the launch party and even afterward, that there was an issue discovered with the ever so popular "zombie mech" centurion. One of the main reasons the mech couldn't be destroyed for so long by laser or missile fire up close is because that weekend's patch changed it's attributes by accident from volumetric.

Also, here's a different take on damage modelling. User Escef posted an interesting discussion video some time ago as RNG being a factor. Keep in mind the video was made in June 2015:



I won't hedge on RNG being a primary factor in the game in regards to damage, but I do believe the video represents a variation of PPTA accurately, in that explosive forces are first forced away from the center and torsos (more densely packed pixels) to the extremities of a mechs arms and legs (least densely packed pixels), rather than further damaging what is already damaged.

Some of this is why I believe systems should start to be affected within the mech and its operation before parts are fully destroyed, in more importantly the head and it's critical components.

Edited by m, 18 March 2016 - 10:01 AM.

[zagibu]

I checked out the Kodiak, and it is quite a bit oversized. It sits at position 54 of 62 with around 108% of average PPTA value.

I've also compiled the complete list of all mechs so far (except the IIC mechs, which should be very similar to their IS counterparts). Will be interesting to see how it looks after the big rescale.

[Seal Farmer]

zagibu, on 23 May 2016 - 03:13 PM, said:

I checked out the Kodiak, and it is quite a bit oversized. It sits at position 54 of 62 with around 108% of average PPTA value.

I've also compiled the complete list of all mechs so far (except the IIC mechs, which should be very similar to their IS counterparts). Will be interesting to see how it looks after the big rescale.

Weird that they just released the Kodiak and its over scaled to 108%.
Wouldnt they have done the Volumetric sizing on it too out of the gate, wouldn't that be easier then have to rescale it again.

[Andi Nagasia]

Seal Farmer, on 23 May 2016 - 03:46 PM, said:

Weird that they just released the Kodiak and its over scaled to 108%.
Wouldnt they have done the Volumetric sizing on it too out of the gate, wouldn't that be easier then have to rescale it again.

the Kodiak isnt over Scaled, its 100% Perfact in the Way of Scaling,
it came Straight from Russ over Twitter, next Month Everyone will see how it fits in,
remember Every Mech is being rescaled, but the Kodiak Russ says was release Pre Scaled,

[ProfessorD]

Andi Nagasia, on 23 May 2016 - 04:12 PM, said:

the Kodiak isnt over Scaled, its 100% Perfact in the Way of Scaling,
it came Straight from Russ over Twitter, next Month Everyone will see how it fits in,
remember Every Mech is being rescaled, but the Kodiak Russ says was release Pre Scaled,

Uhh, source?*


*Being exactly as serious as Andi was...

[Andi Nagasia]

ProfessorD, on 23 May 2016 - 04:48 PM, said:

Uhh, source?*


*Being exactly as serious as Andi was...

Of Course, was actually ganna look for it but forgot Sorry Guys,

Here it is,

Twitter said:

Derrick ‏@Wintersdark May 12
@russ_bullock Question (Not loaded I swear, no follow-up): Is the KDK scaled correctly at release, or will it change?

Russ Bullock ‏@russ_bullock
@Wintersdark correctly

Edited by Andi Nagasia, 23 May 2016 - 08:53 PM.

[zagibu]

Andi Nagasia, on 23 May 2016 - 04:12 PM, said:

the Kodiak isnt over Scaled, its 100% Perfact in the Way of Scaling,
it came Straight from Russ over Twitter, next Month Everyone will see how it fits in,
remember Every Mech is being rescaled, but the Kodiak Russ says was release Pre Scaled,

Yeah, but they do volumetric scaling, and I am actually measuring what surface area the mech is presenting from the front and from the side. You know, what actually matters when you try to shoot at them.

[kapusta11]

You're not accounting for the actual hitboxes, only silhouette. And even If you would there's a big difference between narrow and wide hitboxes, how fast mech can move and torso twist, whether it has high mounted hardpoints or not, how fast it accelerates and stops, does it have protruding belly or head, what about shield arms, you know, stuff that makes a difference. Do you count arm and leg surface as equally important as CT and ST area? What kind of conclusion are we supposed to make based on your data? That Crabs are too small? That's why I barely see them in game? Or that Quickdraw with high mounted hardpoints is the third worst mech? Or that BK and Archers are pretty close?

[Tarogato]

kapusta11, on 24 May 2016 - 02:13 AM, said:

You're not accounting for the actual hitboxes, only silhouette. And even If you would there's a big difference between narrow and wide hitboxes, how fast mech can move and torso twist, whether it has high mounted hardpoints or not, how fast it accelerates and stops, does it have protruding belly or head, what about shield arms, you know, stuff that makes a difference. Do you count arm and leg surface as equally important as CT and ST area? What kind of conclusion are we supposed to make based on your data? That Crabs are too small? That's why I barely see them in game? Or that Quickdraw with high mounted hardpoints is the third worst mech? Or that BK and Archers are pretty close?

When I was making and presenting that graph that Tennex copied, people argued the same thing. "But profile doesn't matter, obviously... because look at how meta the huge Quickdraw is! Did you even take into account shield arms vs gun arms? Or just legs for when you're brawling? Or how the hitboxes are? You didn't even account for top-down profile, where the King Crab is huge!"

I mean, what do you expect? Do you want somebody to concoct a formula that takes all of these things into consideration and poops out a number? And how long are you prepared to argue the details and weightings of various aspects because you think they went about it the wrong way? It's all way too much effort. What we should really be concerned with is, "Are the mechs believably to scale with one another - do they look their weight? Can we corroborate this with measurable numbers?" Can you use scale as a balancing mechanism? Sure, I believe you can (ie. Crab being tiny to make up for how useless it is and Kodiak being huge to make up for how good it is, or making the Catapult smaller so that it performs better), and it works the other way around as well (Dragon too big, so you quirk the heck out of it (oh wait... PGI backed up on that one)).

So yeah, just worry about the simple bit, "Is it believable to scale with the other mechs?". I think PGI's volumetric rescale will address that 100%, but at the same time I think it's a bit extreme because some mechs are fine while being scaled slightly incorrectly (do we really need to make all the 30-35-ton light mechs as big as the Kitfox and Adder? I don't think so, but PGI apparently does.)