819 replies to this topic

[Cementi]

You are correct Quicksliver which is why I think a power draw system might work better.

The issue with gauss in MWO is that when you destroy the gun it does not always blow up. In table top if you crit the gun it blew up every time. Not sure what the percentage of ammo/gauss expolsion (before people say I am an idiot yes I know gauss ammo does not explode but the gun does I am talking about other ammo types as well) but I cannot recall the last time I died due to an ammo/gauss explosion. Especially with the adjusted structure values. Makes it even harder to be destroyed by an ammo explosion. TT was to harsh, MWO is too forgiving in this case.

[invernomuto]

However, according to latest Russ tweets we could test the power drain system on PST.
Right now we are discussing pro and cons of a system that we do not know anything about...

[Satan n stuff]

Appogee, on 14 June 2016 - 11:15 AM, said:

I'd love to see delayed convergence. However, that would require the server to track multiple reticles per Mech, and quadruple the network traffic. So it will never happen.

That's quite the exaggeration, it needs to track exactly two extra variables per mech ( assuming PGI disabled or got rid of them at some point ) and doesn't require any additional linetraces. It would however require a periodic update ( outside of shooting ) to make sure no-one is cheating. That's not all that much considering we've already got those for the positions and orientations of all mechs.
Is it a significant increase in network traffic? Yes. Will it quadruple the network traffic? Hell no, it won't even get close to doubling it.
Verifying the data server side is a fairly simple operation too, effectively negligible as compared to the entire rest of the simulation.

Edited by Satan n stuff, 16 June 2016 - 04:26 AM.

[Quicksilver Aberration]

Cementi, on 14 June 2016 - 12:29 PM, said:

The issue with gauss in MWO is that when you destroy the gun it does not always blow up. In table top if you crit the gun it blew up every time.

Gauss is at 90% chance of explosion in MWO, ammo is much less at 10% (thank god) last I knew.

[Felicitatem Parco]

PGI killed delayed convergence when HSR was introduced. Convergence has to be server-side calculated to prevent cheaters from using a client-side hack to mKe their weapons instantly converge. Server-side convergence HSR checks require as much HSR network traffic as if you are firing an always-on beam weapon at a target whenever you simply point your reticule at it...

[Quicksilver Aberration]

Prosperity Park, on 14 June 2016 - 12:48 PM, said:

Server-side convergence HSR checks require as much HSR network traffic as if you are firing an always-on beam weapon at a target whenever you simply point your reticule at it...

No it wouldn't it would require more information (whether it be sent in by clients or just something extra stored server side), regardless of whether convergence was delayed or lock based. Either way, both require extra calculations, and the game's tick rate is already on the low side.

Edited by Quicksilver Kalasa, 14 June 2016 - 01:08 PM.

[Pjwned]

Mystere, on 14 June 2016 - 10:59 AM, said:

Well, if they're only paying for potato-level servers, then yes HSR and other netcode will be affected.

How would it be affected with that solution though? When you get a target lock then convergence would work just like it does now, and presumably without a target lock then you would just see weapons converged at their optimal range or have them all converge at 2500m or whatever.

What measurable effect would that have on HSR and in what cases?

Edited by Pjwned, 14 June 2016 - 01:15 PM.

[Quicksilver Aberration]

Pjwned, on 14 June 2016 - 01:15 PM, said:

How would it be affected with that solution though? When you get a target lock then convergence would work just like it does now, and presumably without a target lock then you would just see weapons converged at their optimal range or have them all converge at 2500m or whatever.

What measurable effect would that have on HSR and in what cases?

Extra stored data, it would have to at least know the position of the target you have locked for each tick and some method to know if you even had a target locked. Which isn't a whole lot, the big thing is the increased complexity of calculations.

Edited by Quicksilver Kalasa, 14 June 2016 - 01:23 PM.

[Clownwarlord]

In my opinion a heat penalty for going off the scale should hit you like a sledge hammer, otherwise why have a scale? Currently though our heat penalty is more like smothering you with a pillow. You can struggle with it, even catch your breath, and some times die because of it.

[Mystere]

Pjwned, on 14 June 2016 - 01:15 PM, said:

How would it be affected with that solution though? When you get a target lock then convergence would work just like it does now, and presumably without a target lock then you would just see weapons converged at their optimal range or have them all converge at 2500m or whatever.

What measurable effect would that have on HSR and in what cases?

When I wrote that, I was thinking of all weapons firing straight in the absence of a lock. That I am assuming will require more data and/or calculations than when there is a convergence point. I can't really say exactly without knowing how PGI is doing it now.

Quicksilver Kalasa, on 14 June 2016 - 01:23 PM, said:

Extra stored data, it would have to at least know the position of the target you have locked for each tick and some method to know if you even had a target locked. Which isn't a whole lot, the big thing is the increased complexity of calculations.

That's what server-installed GPUs and CUDA are for.

Edited by Mystere, 14 June 2016 - 01:49 PM.

[Appogee]

Satan n stuff, on 14 June 2016 - 12:43 PM, said:

That's quite the exaggeration, it needs to track exactly two extra variables per mech ( assuming PGI disable or got rid of them at some point ) and doesn't require any additional linetraces. It would however require a periodic update ( outside of shooting ) to make sure no-one is cheating. That's not all that much considering we've already got those for the positions and orientations of all mechs.
Is it a significant increase in network traffic? Yes. Will it quadruple the network traffic? Hell no, it won't even get close to doubling it.

CT+LT+RT+RA+LA+HD = 6 potential reticles to track and converge. I assumed you'd have the arms converged already and Hd+CT converged. So that's 4 reticles to track.

So, however much network traffic it takes to track one magically insta-converged reticle, it would take 4X that to track 4 non-converged reticles.

Of course, a Mech that doesn't have weapons mounted in four unique locations would have fewer to track.

[Gas Guzzler]

Cementi, on 14 June 2016 - 12:20 PM, said:

Actually I was recently given the Blood of Kerensky series to read by a friend. (still working on the third book) During Phelan Kell's trial to become a warrior it does indeed mention that the gauss rifles in his opponent take up so much power that the computer has to compensate with the cycling of the other weapons.

So yes in "lore" power draw is indeed referenced. Table top it was not however unless they implement a true heat scale (my preference) something needs to be done and at least a power draw system is talked about in the books. Far better than ghost heat in my opinion as if done correctly it will be able to affect all weapons equally.

That kind of sounds like fluff some authors decided to throw in. Fine, so some BattleTech books vaguely mention it, that isn't evidence to support that it should be in game. Frankly, it simply very misguided, and wont even accomplish what they are trying to do with it.

[Pjwned]

Quicksilver Kalasa, on 14 June 2016 - 01:23 PM, said:

Extra stored data, it would have to at least know the position of the target you have locked for each tick

Pretty sure HSR already does that, so...

Quote

and some method to know if you even had a target locked. Which isn't a whole lot, the big thing is the increased complexity of calculations.

I don't see how simply checking for a target lock would have a measurable impact, but I guess it depends on what would happen in the case of losing a lock (for whatever reason) in the middle of firing your lasers.

Mystere, on 14 June 2016 - 01:46 PM, said:

When I wrote that, I was thinking of all weapons firing straight in the absence of a lock. That I am assuming will require more data and/or calculations than when there is a convergence point. I can't really say exactly without knowing how PGI is doing it now.

I see your point then, but having a defined convergence point without a lock would work fine and be way less of a hassle.

[Chuanhao]

Xetelian, on 13 June 2016 - 06:45 PM, said:

Where people used to carry 5 cMPL they'll be packing a UAC20 and 3 erML on their stormcrows
Where stalkers used to carry 6 LL they'll switch to a couple LPL and SRMs

I actually thought that would be a great condition to be in with more mix weapon config than just one single weapon type.

I will also suspend judgement until the details come in.

But there had better be sufficient explanation and test cases explaining how the mechanic will work for the different types of load outs, and sufficient trials.

[Quicksilver Aberration]

Pjwned, on 14 June 2016 - 02:00 PM, said:

Pretty sure HSR already does that, so...

Why? That's extra state/data that isn't necessary.

Pjwned, on 14 June 2016 - 02:00 PM, said:

I don't see how simply checking for a target lock would have a measurable impact, but I guess it depends on what would happen in the case of losing a lock (for whatever reason) in the middle of firing your lasers.

More like anytime during that HSR threshold, but your example is a perfect case for why that is actually important. Remember, there is always a disconnect between what you see and what the server sees, so situations like that would be a problem for snapshots near the time you locked a target.

Edited by Quicksilver Kalasa, 14 June 2016 - 02:31 PM.

[Razorfish]

Currently this game is messed up, it’s almost not worth playing.

Let’s see how this works before we shoot it down.

They could accidentally do something cool.

[Gas Guzzler]

Razorfish, on 14 June 2016 - 02:30 PM, said:

Currently this game is messed up, it’s almost not worth playing.

Let’s see how this works before we shoot it down.

They could accidentally do something cool.

The only thing that makes it not worth playing is lack of meaningful content. The gameplay itself is good, but we need something more interesting to do. Maps, game modes, more IW type consumables, etc.

[LorDGuilhotinA]

Nightshade24, on 13 June 2016 - 10:28 PM, said:

let me help you with that:

http://farpnut.net/2...ine-heat-scale/
Youtube video: https://www.youtube....?v=wNHoe6D19N8. "Mechwarrior Online Engine Capacity Concept "
(yes: this is the most popular demanded feature for MW: O when it comes from ghost heat: besides "keep ghost heat!!" and "remove ghost heat!!". This was a feature the community repeatedly asked and we are getting it due to that. Welcome to the whole "but I didn't know about it!" and the "they should listen to the majority! We're the majority!" -also has unknown idea)

YES, YES YES
i Want exactly this
This makes sense, few days before i hear about "Power Draw" system i was thinking just about something like this, if the engines generates the power to move, the power for the weapons has to come from this same source, this source can only generates a certain amout of power for weapons as it can generate power to the mech to move. The bigger this energy source is, the faster the mech can move, the more energy it can deliver to the weapon system.
Also, from this point, i can imagine a way to implement another way to use this power draw system, like sensors, i could be able to use a more powerfull "sensor system" that gives me a higher sensor range/ faster information gathering (... ), this more powerfull sensor system uses more energy from the energy source and leaves with less energy to the weapon system.
Or, i could be able to equip a less powerfull sensor system, wich consumes less power from the source leaving this power for the weapons system. (Infotech system anyone???)

Another thing, fixed heat cap makes no sense at all. Adding more heatsink should always increase the heat cap, thats how it work in real life, lets get an audio amplifier, the more powerfull the amplifier is (giving the same tecnology), the bigger the heatsink must be to handle the heat (or you need to add another cooling system, like a fan).

I do really hope that PGI uses a system like the one described in that video.

thanks Nightshade24 your posts are always worth reading.

[Quicksilver Aberration]

Gas Guzzler, on 14 June 2016 - 02:32 PM, said:

more IW type consumables

Gas Guzzler, on 14 June 2016 - 02:32 PM, said:

IW type consumables

Gas Guzzler, on 14 June 2016 - 02:32 PM, said:

type consumables

Gas Guzzler, on 14 June 2016 - 02:32 PM, said:

consumables

Bad Gas (lol), bad!

Don't suggest more consumables, they are still dumb

[Gas Guzzler]

Quicksilver Kalasa, on 14 June 2016 - 02:35 PM, said:

Bad Gas (lol), bad!

Don't suggest more consumables, they are still dumb

But they could be used for cool stuff... I do think coolant flushing should come with your mech, but there are some neat IW bits that you could do with consumables that you couldn't do with anything else.

What's your beef with consumables?