88 replies to this topic

[Widowmaker1981]

Otto Cannon, on 10 March 2015 - 09:29 PM, said:

And nothing of value was lost.

And something of great value was gained. Because PGI couldn't give the Firemoth, they had to give us the Cheater, which is actually good.

[Widowmaker1981]

Macster16, on 10 March 2015 - 11:18 PM, said:

Don't care about the Flea.

The Firemoth though I hope they do something about to get in game. Get rid of MASC if you have to and just cram in another 2 DHS or something for the stock build or whatever - lore be damned! The Firemoth has a special place in my heart from MW2 being the game's "entry" mech.

PS. Off topic, but I do remember stuffing around with the Firemoth in MW2 by freeing up as much tonnage as possible and then filling it up with as many JJs as it could fit and then using the forward thrust (MW2 had lateral thrusting on JJs) to go over 1000km/h. I remember outrunning enemy LRMs and lasers as they tried to shoot my Firemoth as it flew right past them .

The Firemoth would do 180 without MASC. Our survey says: NOT GONNA HAPPEN.

[Macster16]

Widowmaker1981, on 11 March 2015 - 01:41 AM, said:

The Firemoth would do 180 without MASC. Our survey says: NOT GONNA HAPPEN.

Hmm you're right. Forgot to account for speed tweak which would bump it to 178.2km/h from the base 162km/h which certainly exceeds the supposed Cryengine limit. Only way then to get it in game would be to remove MASC completely and reduce engine to atleast 190 (from 200) which matches the Locust engine cap. Dunno if PGI would be willing to diverge this much from the original source, though it would make for a decent 20T light as now it would have more podspace for more pew pew from the less than 5 tonnes of space it would have with a 200XL engine and MASC that it comes with....

[Widowmaker1981]

Macster16, on 11 March 2015 - 02:39 AM, said:

Hmm you're right. Forgot to account for speed tweak which would bump it to 178.2km/h from the base 162km/h which certainly exceeds the supposed Cryengine limit. Only way then to get it in game would be to remove MASC completely and reduce engine to atleast 190 (from 200) which matches the Locust engine cap. Dunno if PGI would be willing to diverge this much from the original source, though it would make for a decent 20T light as now it would have more podspace for more pew pew from the less than 5 tonnes of space it would have with a 200XL engine and MASC that it comes with....

At that point they could just completely invent a clan 20t light that doesnt look like a poo throwing monkey though...

[MeiSooHaityu]

stjobe, on 10 March 2015 - 10:57 PM, said:

In short, it's PGI's crappy code that's to blame.

A lot of net code woes is heavy server side verification. They may be able to streamline their code, but it will always fluctuate with ping and connection quality. Those are things PGI can't directly control.

I will say this though. The Cryengine has some of the most atrocious net coding I have ever seen. I think any company that uses Cry engine is best to rewrite the net code portion of the engine (like PGI did and had to). I swear it feels like everything is handled client side and the server is cool with whatever happens.

I remember buying Crysis and going online to play. About a month after that game came out, it was so hacked the online portion was completely broken. Its fun when infinite ammo nuke tanks are cruising the map at the very start of the match.

Maybe MWO's net code isn't the BEST, but a mentally challenged chimp could program better net code than Crytek.

[FupDup]

Whatzituyah, on 10 March 2015 - 08:57 PM, said:

But this also leads to other mechs that can go over 170 kph not making it in the game Firemoth for instance.

The Firecrotch is a a Myst Lynx with its tonnage reduced to 20 while retaining the same physical size as the Lynx...including those big burly "look ma, no hands!" arms. And since those arms are mounted over the head, they can be hit from almost any angle and are nearly impossible to shield from the red team. Its guns would get torn off real fast. Being 20 tons is in general never a good thing.

In exchange, it gets to run around in circles at warp speed....


Yeah I think I'd rather have the Arctic Cheetah, because it will actually be a very good light unless PGI somehow intentionally messes it up (which would be difficult to do).

[Dawnstealer]

I'm not sure exactly where they said it, but PGI/Russ have said that the release of the Locust replaced the Flea. Considering that the Flea was originally a 15 ton mech when it first appeared in Battle Technologies magazine, it kind of makes sense that the Locust would be the preferred "superlight" mech.

I know people were pushing for this mech as much as we were pushing for the Urbie, but the Locust and the Flea really are the same mech. The only thing missing is a MASC, and that technology hasn't been implemented yet.

And this doesn't mean that PGI won't ever release the Flea, just that the Locust fills that niche very well.

[Mawai]

stjobe, on 10 March 2015 - 10:57 PM, said:

FTFY.

It's not CryEngine's fault, CryEngine doesn't have a speed limit of 170 kph.

MWO perhaps does, but if so it's due to PGI adding server-side authorization and HSR to the engine.

In short, it's PGI's crappy code that's to blame.

Don't get it do you? Do you actually code anything?

There is a fundamental design choice

1) Server authoritative which minimizes hacks and cheats

2) Client authoritative with a constant development war between the game developer and hackers to stop cheating ... which the hackers are always ahead in. The list of possible hacks when the client decides what happens is ridiculously long.

What is the consequence of this design decision?

The server has to determine where mechs are located and resolve weapon fire on an approximate representation of the battlefield on the server based on the latest client data available. It uses host state rewind to determine where the target should have been on the client when the client decided to fire. It then determines whether the shot hit or not.

So what happens with fast mechs? They move. With a 250ms ping this is 1/2 of a second round trip lag to the server. Add another player with a 250ms ping and it is 0.5 seconds one way from one player to the other without taking into account processing overhead ... just network latency ... over which PGI has absolutely NO control ... their code has to deal with it and that is all.

So in 0.5 seconds a 170kph mech will move about 23.5m. Say 12m in 250ms. If the target changes direction ... it will take 250ms for that information to reach the server and 500ms to reach the other client. If the target mech isn't close to where the firing client was drawing it (it could be up to 24m displaced when the information finally gets to the firing client) ... you get rubber banding .. or mech displacement since the target has to suddenly jump to where is should be from where the client has been drawing it. This problem only gets worse the faster the mechs go.

I think the flea with MASC is something like 240kph. The problem should get linearly worse with speed ... so I don't think 170 is a hard cap ... it is a working number where the balance between hit registration issues and "lag" shield is balanced with the desire to be able to operate fast mechs.

Anyway, it may be that PGI could improve the code ... I have no idea what it looks like... but the fundamental issue derives from the server side authoritative design decision (which I agree with) and the physics of light speed limitations and network latency ... which no amount of coding will ever make perfect.

Edited by Mawai, 11 March 2015 - 05:21 AM.

[CDLord HHGD]

I think it's super LAME that they are not going to move forward with the Flea. And it's not the engine's fault, more coding than anything. I do have a solution though...

Reduce, in-engine, the relative speeds of all mechs so that the MASC enhanced speed is tolerated like the current 171 kph speed.

[Corbon Zackery]

From what we seen with Wave 3 M.A.S.C is going to be implemented in some way for the Shadow Cat. That brings the Flea back into play.

Russ did say that the Flea was dead on arrival but that might change as the Flea is the one promise that PGI has not kept.

We asked in the past to release the one version that has no m.a.s.c now to fulfill the promise.

So its a interesting situation.

[CDLord HHGD]

Mayhap I'll start a thread to petition for the Flea like I did for the CN9-AH. Though the CPLT is still waiting for reversals of the tragedy done to it's ears....

[Logan Hawke]

Whatzituyah, on 10 March 2015 - 11:08 PM, said:

I am begining to think you all think PGI did it we don't really know that even though they made bad mistakes before.

Well, it's not that we think PGI did it on person (well, except some of the butthurt brigade perhaps), it's more that PGI are... not very good at this and are learning as they go along, using bandaids instead of the in depth surgery things need. If they could somehow transfer over to an easier to work with engine such as UE4, that'd be great, but the chances of that happening are pretty much nonexistant. So we're stuck with the mistakes the made at the beginning that can't be removed without serious problems and tons of work being needed.

[mike29tw]

The notion that CryEngine has a speed limit is really entertaining.

So what happens when you go over the CryEngine's speed limit?

Does CryTek send you a ticket?

[Catra Lanis]

Do we need mechs that bounce and warp around literally like a flea? The twitch crowd might think so, personally I just think it destroys immersion looking at a 20-25 ton warmachine running around like a chipmunk on crack, lore or not.

[Mawai]

cdlord, on 11 March 2015 - 06:33 AM, said:

I think it's super LAME that they are not going to move forward with the Flea. And it's not the engine's fault, more coding than anything. I do have a solution though...

Reduce, in-engine, the relative speeds of all mechs so that the MASC enhanced speed is tolerated like the current 171 kph speed.

Huh?

So you are suggesting slowing every mech down and then just giving it a different number? If the mechs cover the same distance relative to their size in a given time frame then you run into this issue no matter what number you assign to the speed..

Mechs have a specific size. Terrain has a specific size. In meters. Speed is how fast the object moves relative to its size and the terrain size. If a mech is 10 meters tall and the mech travels its height in one second then it is moving 10m/s,

I agree that they would have fewer issues with latency if they slowed down every mech in the game ... but then every mech would move slower. A Jenner would not do 150 .. it would be doing 130 ... even if you decide to call it 150.

[Heffay]

DivideByZer0, on 10 March 2015 - 11:56 PM, said:

*edit* maybe the answer you're looking for, is, SC uses a really streamlined netcode for an alpha, plus a predictive interpolative system that's practically seamless (as best I can tell). Oh, and it doesn't crash.

Star Citizen is still client side. I'm sure they'll get some cheat detection in soon, but right now it's nothing they can release as a finished product.

Not to mention they have less than half the number of simultaneous connections per match, so the corresponding network traffic is a lot less. And there is still rubber banding going on. They have a great demo, but still a ton of issues to solve.

[Mawai]

Heffay, on 11 March 2015 - 07:36 AM, said:

Star Citizen is still client side. I'm sure they'll get some cheat detection in soon, but right now it's nothing they can release as a finished product.

Not to mention they have less than half the number of simultaneous connections per match, so the corresponding network traffic is a lot less. And there is still rubber banding going on. They have a great demo, but still a ton of issues to solve.

I also think you need to compare funding and team size if you try to look at PGI and RSI. PGI has about 8 engineers to do everything ... any idea how large the RSI development team is?

[Heffay]

Mawai, on 11 March 2015 - 07:52 AM, said:

I also think you need to compare funding and team size if you try to look at PGI and RSI. PGI has about 8 engineers to do everything ... any idea how large the RSI development team is?

Last I saw was over 250, including contractors/outsourced people working on the game.

[Almond Brown]

stjobe, on 10 March 2015 - 10:57 PM, said:

FTFY.

It's not CryEngine's fault, CryEngine doesn't have a speed limit of 170 kph.

MWO perhaps does, but if so it's due to PGI adding server-side authorization and HSR to the engine.

In short, it's PGI's crappy code that's to blame.

PGI's decision to go with "server-side authorization" was a good, no great decision. I have played HACKER plagued games. They all suck large balls. MWO currently does not.

Perhaps it is time to leave that "Island".

[Almond Brown]

DivideByZer0, on 10 March 2015 - 11:13 PM, said:

I don't necessarily want to be negative, but... err....I have yet to see evidence of the contrary. Star Citizen has ships moving 1,000 km/h (300 m/s) in cryengine in an alpha.. for example..
I play MWO because it's fun, and I enjoy it. I support PGI in hopes of improvement, like you.

Then why are you here raging on PGI when you could be flying around in the SC Sim? Think of the angst you would save yourself, and everyone else here, not having to deal with such apparent incompetence...

Edited by Almond Brown, 11 March 2015 - 08:16 AM.