Jump to content

Alternative Modeling Of Masc, So I Can Have My Dadgum Flea!


18 replies to this topic

#1 CocoaJin

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,607 posts
  • LocationLos Angeles, CA

Posted 11 September 2014 - 10:36 AM

I'd like to discuss alternative methods for modeling MASC. Ultimately, these alternative methods would be assumed to be temporary until the applicable restrictions of the game engine with respect to max speed limits could be resolved.

1) As a Damage modifier:
The idea here is to have the player activate MASC to provide a damage resistance buff, not an increase in speed, for the activation duration. Activation would/could/should require the mech be at a minimum speed to be activated...perhaps a minimum of 75% throttle and/or some minimum absolute speed for the chassis or class.
Successful activation of MASC would cause all damage received to be reduced by some %, plus the appropriate increase in heat production while MASC is active(I assume this means prolonged use could cause a mech to over-heat, automatically toggling MASC off). Perhaps as a counter-balance, all weapons fired by the user mech while MASC is active would also reduce own damage output by some %...I'd suggest somewhere between 50-67% of the % reduction for the user's damage resistance buff(so if MASC reduced damage received by 10%, the user's weapon damage output would reduce by 5 to about 7%). These mechanics are designed to emulate the difficulty in both being hit and successfully hitting others at the higher speeds MASC would have provided.

These modifiers could be further modified by modules, perhaps even by seperate modules.

In addition to the obvious buff for successful activation, their needs to be a mechanic for when MASC fails. I was thinking of a damage resistance debuff that increases damage received for a fixed duration of time when it fails. The debuff could be 1.5-2x that of the normal damage resistance buff when MASC operates successfully...so a MASC failure would result in an increase in damage received by 15-20%. This could be in lieu or in addition to the appropriate leg damage sustained as part of a MASC failure. A MASC failure would cease to develop additional heat, even during the failure duration.

2) As a fixed acceleration and max sustained speed modifier:
Basically, find some averaged speed for a mech over a prolonged period in which it is assumed the mech is traveling at max speed, and considering failure chance rates, find the averaged speed based on repeatedly using MASC, plus a MASC failure. What ever distance covered for the time period used, make that averaged speed the mech's new sustained max speed when equipped with MASC...no temporary speed boost.

In addition, provide an appropriate increase in standard acceleration for the mech.

This method would require no activation mechanic for the MASC equipped mech, because all benefits and drawbacks are averaged into the "new" performance specs for the mech.

Modules could possibly be used to modify these values.

3) Re-scale the mechs and applicable system models:
Re-scale so mechs and systems travel/operate at higher scale speeds without the possibility to exceed the absolute speed limit of the engine. This re-scale would essentially make all mechs, and possibly even all other 3D models on the map such as trees and buildings smaller. We'd also need to adjust mech turn and rotation rates to complete the scale down...I believe it would require reducing turn and torso rotation rates by approximatly the same multiple of mech and 3D object scale. It would reduce weapon ranges and projectile speeds by the same multiple...it might also be justifiable to adjust fire rates also.

In the end, the goal of re-scaling is to make everything feel the same, but allow for what feels like faster mech max speeds. In addition, maps would now feel bigger without actually being bigger.

Please discuss.

Edited by CocoaJin, 11 September 2014 - 10:50 AM.


#2 Scratx

    Member

  • PipPipPipPipPipPipPipPipPip
  • Survivor
  • Survivor
  • 3,283 posts

Posted 11 September 2014 - 10:54 AM

1 ) Umm... you suggest MASC to work like a weird sort of damage reduction technology instead of a speed booster. This is ridiculous, sorry.

2 ) This is nicer but on the other hand not really. What it boils down to in the end is a tonnage/crit-space trade-off between going bigger engine and save crits or go smaller engine and spend crits. And given you can go max engine and still put MASC on for additional mojo? Yeah, how does this help with the implementational roadblocks PGI has ahead of them to get MASC in?

3 ) You fail to understand the issues. Re-scaling does nothing. If going 100kph for one second moves you 1.0 units of space in the game engine, making you go 10 units or 0.1 units instead does nothing. You're still going at 100kph.

It's not the absolute speed by the floating point numbers that gives trouble, it's the relative speed. It's the fact that when you're going at 250kph, a miss on predicting where you are and/or going is going to land you a LOT further away from where you really are supposed to be than if you were going at 50kph. The network code has to deal with this crap. Right now it's been pushed to ~170-180kph that we know of... we're not allowed to go any further than 171kph because the network code will likely break if we do. PGI's afraid of trying to tackle this because it's a monster of a problem and they have a lot more things to deal with that are, frankly, more important.

We'll get MASC eventually, in whatever form.

#3 xMintaka

    Member

  • PipPipPipPipPipPipPip
  • The Nimble
  • The Nimble
  • 882 posts

Posted 11 September 2014 - 11:12 AM

View PostScratx, on 11 September 2014 - 10:54 AM, said:

We'll get MASC eventually, in whatever form.


PGI's track record has killed most people's optimism. Glad to see there's still one delusional person out there.

#4 Scratx

    Member

  • PipPipPipPipPipPipPipPipPip
  • Survivor
  • Survivor
  • 3,283 posts

Posted 11 September 2014 - 11:16 AM

View PostLunatech, on 11 September 2014 - 11:12 AM, said:


PGI's track record has killed most people's optimism. Glad to see there's still one delusional person out there.


Because all the progress forward in 2014 never happened. Got it.

#5 Ghogiel

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • CS 2021 Gold Champ
  • CS 2021 Gold Champ
  • 6,852 posts

Posted 11 September 2014 - 11:26 AM

I think PGI and the player base can come to a compromise on MASC.

Disregarding the actual implementation and mechanics on how it actually would work in MWO, the main issue stopping MASC seems to be the mechs going +150kph with it. So let's just not do that specifically then.

How about we just don't put any mechs in the game with MASC that push those speeds. There are numerous mechs that could in my mind be considered for addition to the game without breaking the speed barrier.

Let's forget the Flea, Firemoth etc. And just agree for now that getting a Shadowcat, Executioner etc.

Why can't we put MASC in as long as we don't let it break stuff. Clan mechs are easy since we can't change engines. We pretty much just have to limit light mechs with it is all.

#6 MAXrobo

    Member

  • PipPipPipPipPip
  • Bridesmaid
  • Bridesmaid
  • 143 posts
  • Locationmiddle of nowhere

Posted 11 September 2014 - 11:58 AM

I would like to see MASC as an increase in agility. Have it so that when it is activated, it gives 75% more acceleration, deceleration and turn speed for a short burst. obviously the exact numbers could be tweaked. but a significant increase to agility would be useful on any mech regardless of weight class and still acts close to how it does in canon without causing the game any problems.

#7 Tombstoner

    Member

  • PipPipPipPipPipPipPipPipPip
  • Bridesmaid
  • Bridesmaid
  • 2,193 posts

Posted 11 September 2014 - 12:04 PM

View PostScratx, on 11 September 2014 - 10:54 AM, said:

1 ) Umm... you suggest MASC to work like a weird sort of damage reduction technology instead of a speed booster. This is ridiculous, sorry.


Faster moving targets should be harder to hit. simply extrapolate the TT hit modifiers and you see that moving faster makes you harder to hit = more misses and that reduces damage. since that cant be done in the cry engine skill based tarting environment of MWO, then a flat damage resistence to emulate speed related misses is resonable to consider.

#8 Ph30nix

    Member

  • PipPipPipPipPipPipPipPipPip
  • 2,444 posts

Posted 11 September 2014 - 12:17 PM

they never said NO masc ever, they just said no masc at start of CW.

now frankly given their abilities shown to date i don't think they will ever be able to get masc working how it does in lore.

I think best we can hope for is the mechs equipped with it get their engine treated like its X ratings higher. (they could even make it something all(some) clan mechs can upgrade too)
so you could throw on onto a Kitfox and have its speed be the same as if it actually had an XL 210 instead of XL180
actually id rather see adder get it than kitfox (if it happened)

but back to your flea, it would just be able to have a higher then normal top speed for its engine rating compared to other mechs with the same weight/engine. Probably easiest way for them to do it.

#9 General Taskeen

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • 5,737 posts
  • LocationCircinus

Posted 11 September 2014 - 12:33 PM

It can't be that hard to make MASC seeing as MW:LL implemented it just fine, and they programmed it to be non-complicated (Hold button-> Go Fast+Heat Buildup)... for free. They even have a MASC Mechs that exceed 150 km/h.

#10 Xarian

    Member

  • PipPipPipPipPipPipPip
  • Wrath
  • 997 posts

Posted 11 September 2014 - 12:39 PM

Problem
PGI is currently at the maximum speed allowed by their network protocol.

Cause of the Problem
When you move, the game engine has to communicate with the server in order to tell it that you moved. How fast your computer thinks you are moving and how fast the server thinks you are moving depends on your acceleration and, likely, speed. When you have low amounts of acceleration/speed, the difference between where the server thinks you are and where the client thinks you are is very small - this looks like you are moving properly. You will only see rubberbanding if your connection is having problems.

When you are at high amounts of acceleration/speed, the client moves your mech to a certain location, then receives a new location from the server. If they aren't the same, you will observe rubberbanding when your PC changes your location to where the server just said it was.

Discussion
CryEngine was developed for man-scaled critters. When it communicates positions, it uses a fairly high level of accuracy in order to guarantee that weapons are accurate and such. When PGI made MWO, it made mechs to full scale - that is, a human in Crysis would look tiny compared to a Mech in MWO. The speeds involved are likewise much higher. This level of accuracy uses a lot of bandwidth, much of which is being wasted in measuring distances that are not important on the scale of MWO.

Possible Solutions
There are several possible solutions, but in the end they all boil down to the same thing: reducing the accuracy of the positions communicated between the client and server.

1. Decrease the scale of everything in the game by a significant factor (suggested 1/4x scale), then change all the distance readouts to compensate. The game would look the same, but it would think that you are moving more slowly - 160 kph in the new scale would behave exactly the same as 40 kph in the old scale.

2. Decrease the accuracy of the information communicated between the server and the client. Instead of using 64-bit numbers, use 32-bit numbers; instead of using single-sized 32-bit numbers, use 16-bit numbers. This might kill accuracy too much, however: an even better solution would be to encode your packets so that you transmit 28-bit numbers and decode them back into 32 bits after receiving the information (essentially losing 4 bits of accuracy, which for our purposes wouldn't be noticeable at all).

Comparison of Methods
Every decrease in scale by a factor of 2 is the equivalent of decreasing numerical accuracy by 1 bit. That is, if you decrease the size of everything to 1/256x scale, that's the same as going from 32-bit numbers to 16-bit numbers. A 1/16x rescale would have the same weapon accuracy as using 28-bit numbers instead of 32-bit numbers (the game most likely uses 32 bits, because that's standard). (For non-programmer-oriented folks: note that this has nothing to do with 32-bit vs 64-bit processors whatsoever)

Difficulty
Decreasing scale is not, by itself, conceptually or programmatically difficult. Programmatically, you would have to change the numeric readouts to compensate for the new factors - 1/4x scale, for example, would require you to multiply everything by 4, and you'd have to add some coding into the weapon configuration subroutines to adapt to the new scale values. That's not the hard part. It would require redoing every 3d mesh in the game; you could write a script to do it for you, but that takes time and there are a lot of models.

Encoding values to use non-32-bit data is programmatically not difficult but it is not trivial, either. The upshot of this method is that it doesn't require altering anything other than the transmission protocol.

Other effects and other unintended consequences
These methods would hurt weapon accuracy slightly due to issues with rounding. However, that would be significantly less than the current effects of network latency.

Using these methods, network behavior would improve significantly. Not only would you no longer rubberband when you are moving, but your connection would be better overall. Smaller packets = better connection, hands down. This would further help with desync issues.

Conclusion
Of these two methods, re-coding the transmission protocol to use lower numeric accuracy would have significant positive effects on the game and essentially no drawback except for the cost in development time. It would also have additional positive effects that are currently being pursued by the development team (namely issues with desyncs and latency).

#11 Nathan Xain

    Member

  • PipPipPip
  • The Ironclad
  • The Ironclad
  • 58 posts

Posted 11 September 2014 - 12:51 PM

I like the concept someone else had of it being a temporary agility boost. Perhaps with heat build up while activated.
So, for a short time you can turn, accelerate, decelerate, twist and suchlike much faster than even mastered stats (but functioning in a similar way to them) would let the mech, but not for long, and not while alpha striking constantly.

I see it being useful in a scenario such as scouting- activate it to come as close as a mech can to diving for cover.

Or something. Basically, I want it in the game somehow so I can run around in a Shadow Cat.

#12 Tarogato

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Civil Servant
  • Civil Servant
  • 6,558 posts
  • LocationUSA

Posted 11 September 2014 - 01:08 PM

View PostXarian, on 11 September 2014 - 12:39 PM, said:

long post full of decent points


Ever since I learned that they made MWO full scale in CryEngine, I immediately thought it was stupid. For all we care an Atlas is 2 meters tall in the game engine - it's the environment and behaviour that makes it feel like a huge stompy robot. I urge the devs to experiment - try the scaling idea. Cut the games scale in half and make a public test of it. Even if it takes a lot of monotonous work to change all the relevant coding, I think it's worth a shot. And heck, if we reduce the scale so much that it makes pinpoint targeting more difficult, then hey, we've solved two issues at once - weapon convergence.

#13 CocoaJin

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,607 posts
  • LocationLos Angeles, CA

Posted 11 September 2014 - 01:46 PM

View PostMAXrobo, on 11 September 2014 - 11:58 AM, said:

I would like to see MASC as an increase in agility. Have it so that when it is activated, it gives 75% more acceleration, deceleration and turn speed for a short burst. obviously the exact numbers could be tweaked. but a significant increase to agility would be useful on any mech regardless of weight class and still acts close to how it does in canon without causing the game any problems.


I could along with an agility boost, but I'd have an issue with a mech being able to utilize both increased acceleration and turning still the same time...I fear it'll make some light mechs move like squirrels. Since speeding up, slowing down and turning are all accelerations, we should see total acceleration bonus shared as a sum of accelerations.

In other words, if we get a 10% boost in acceleration as part of the agility bonus, we should see straight line speed increases using the full 10% for a 10% in acceleration....or a slow turning like pivot using the full 10% for a 10% faster turn...or a running turn using a sliding scale that provides a 3-7% increase in turn and 7-3% increase in speed increase respectively.

#14 Xarian

    Member

  • PipPipPipPipPipPipPip
  • Wrath
  • 997 posts

Posted 11 September 2014 - 02:15 PM

View PostTarogato, on 11 September 2014 - 01:08 PM, said:

Ever since I learned that they made MWO full scale in CryEngine, I immediately thought it was stupid. For all we care an Atlas is 2 meters tall in the game engine - it's the environment and behaviour that makes it feel like a huge stompy robot. I urge the devs to experiment - try the scaling idea. Cut the games scale in half and make a public test of it. Even if it takes a lot of monotonous work to change all the relevant coding, I think it's worth a shot. And heck, if we reduce the scale so much that it makes pinpoint targeting more difficult, then hey, we've solved two issues at once - weapon convergence.

Changing scale also requires you to change all physical constants, such as gravity, to fit the new scale system. Using lossy compression on the transmitted numbers, however, doesn't require changing anything except for a their coding and decoding methods.

#15 Funkadelic Mayhem

    Member

  • PipPipPipPipPipPipPipPip
  • Knight Errant
  • 1,811 posts
  • LocationOrokin Void

Posted 11 September 2014 - 02:28 PM

I want my Flea! If you cant get MASC working give the Fleas Jump Jets instead. Fleas are jumpers not runners.

#16 Naduk

    Member

  • PipPipPipPipPipPipPipPip
  • Overlord
  • Overlord
  • 1,575 posts
  • LocationAustralia

Posted 11 September 2014 - 02:47 PM

why not just make MASC about speed boost rather than max speed ?

would it not be just as powerful to apply a 45%-60% increase to a mechs launch speed ?
but generates heat while you have it on and move around ? (gota remember to turn it off folks)

this way a flea could have a max speed of 170Kph like everyone else, but get there A LOT faster
thus making their MASC feel valuable ?

#17 CocoaJin

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,607 posts
  • LocationLos Angeles, CA

Posted 11 September 2014 - 03:21 PM

View PostNaduk, on 11 September 2014 - 02:47 PM, said:

why not just make MASC about speed boost rather than max speed ?

would it not be just as powerful to apply a 45%-60% increase to a mechs launch speed ?
but generates heat while you have it on and move around ? (gota remember to turn it off folks)

this way a flea could have a max speed of 170Kph like everyone else, but get there A LOT faster
thus making their MASC feel valuable ?


I think we have to be careful about giving too much of an acceleration boost...high max speed isn't anything too eye-brow raising, one can intuitivly accept that given enough time and performance, you can get just about anything going fast enough. But too much acceleration will immediately raise red flags, even if you can't quite put your finger on why(like when a movie tries to speed up the film to simulate high speeds...it's not for the same reason, but because your brain picks up on the inconsistencies). You don't want the mechs to look squirrelly and weightless/massless in their movements.

One of the key issues with high acceleration is that it's not just about mech performance, it's about how much performance it's applying at once/per interval. In order to cram as much performance into an interval of application not only requires the mech to be able to deliver it, but of the mech's structure and the environment to accept and recieve it. You have to have an environment and a mech's structural capacity willing too cooperate with the mech's desire to deliver that much performance at a time. There are limits to how much is believable.



#18 CocoaJin

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,607 posts
  • LocationLos Angeles, CA

Posted 11 September 2014 - 07:30 PM

Anybody else able to chime in regarding the validity of changing the scale of the game to resolve the speed limit cap? Right now it's 1-1 for and against. We need a tie breaker by someone else who sounds like they know what they are talking about :)

#19 CDLord HHGD

    Member

  • PipPipPipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 9,190 posts
  • Location"You're not comp if you're not stock."

Posted 11 September 2014 - 08:07 PM

View PostLunatech, on 11 September 2014 - 11:12 AM, said:


PGI's track record has killed most people's optimism. Glad to see there's still one delusional person out there.

TWO!





1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users