Jump to content

"targeting Computer" Cylinder Of Fire Aiming Mechanic - Video + Demo


132 replies to this topic

#21 Nerdboard

    Member

  • PipPipPipPipPipPip
  • Shredder
  • Shredder
  • 226 posts

Posted 09 February 2016 - 08:26 AM

No clue how that would feel ingame... but I like the idea especially since steady aim and high accuracy is something which differs a lot between this game and battletech books or tabletop games.

Also, nice demonstration.

Edited by Nerdboard, 09 February 2016 - 08:28 AM.


#22 Greyhart

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • Bad Company
  • 894 posts
  • LocationUK

Posted 09 February 2016 - 08:34 AM

Kudos to the OP.

I think it's a good idea and deals with the issues of pin point damage.

As for coding, well no one here knows but cone of fire is already in the game for MG so the basics are there.

#23 Lostdragon

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 2,711 posts
  • LocationAlabama

Posted 09 February 2016 - 08:42 AM

View PostGreyhart, on 09 February 2016 - 08:34 AM, said:

Kudos to the OP.

I think it's a good idea and deals with the issues of pin point damage.

As for coding, well no one here knows but cone of fire is already in the game for MG so the basics are there.


Not to mention missile weapons and LBX ACs... They have quirks to tighten their spread so PGI has the framework for many types of weapons to use a COF that can be adjusted. MGs are hitscan like lasers but SRMs should not be that much different than ballistics/PPCs.

#24 Sader325

    Member

  • PipPipPipPipPipPipPipPip
  • 1,181 posts

Posted 09 February 2016 - 09:12 AM

We have lasers IRL that can shoot down missiles kilometers away.

This video doesn't make any sense.

Come of fire with cannon weapons while shooting, sure between recoil and stuff that makes sense. However giving lasers what amounts to recoil literally makes no sense at all. All I'm seeing in your video is a mech malfunctioning that, not something that anyone would put up with in the field.

In short, I knew this was a dumb sounding idea and the video makes it look even worse.

#25 Lostdragon

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 2,711 posts
  • LocationAlabama

Posted 09 February 2016 - 09:28 AM

View PostSader325, on 09 February 2016 - 09:12 AM, said:

We have lasers IRL that can shoot down missiles kilometers away.

This video doesn't make any sense.

Come of fire with cannon weapons while shooting, sure between recoil and stuff that makes sense. However giving lasers what amounts to recoil literally makes no sense at all. All I'm seeing in your video is a mech malfunctioning that, not something that anyone would put up with in the field.

In short, I knew this was a dumb sounding idea and the video makes it look even worse.


Battletech weapons are supposed to be inaccurate as shown in the video.

#26 p4r4g0n

    Member

  • PipPipPipPipPipPipPipPip
  • Knight Errant
  • 1,511 posts
  • LocationMalaysia

Posted 09 February 2016 - 09:57 AM

IIRC, PGI stated that they were running into bandwidth limitations between server and clients with the current amount of data that is being transferred while a match is running.

If this is correct, how would the proposal in the OP impact this if at all? I am assuming that you will want to maintain having all this run on the server rather than client side?

#27 Rhialto

    Member

  • PipPipPipPipPipPipPipPipPip
  • Philanthropist
  • 2,084 posts
  • Twitter: Link
  • LocationQuébec, QC - CANADA

Posted 09 February 2016 - 09:57 AM

It makes no sense to have lasers that move or deviate like this...

#28 Oderint dum Metuant

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 4,758 posts
  • LocationUnited Kingdom

Posted 09 February 2016 - 10:25 AM

Homeless bills idea is still the best idea to date

#29 Champion of Khorne Lord of Blood

    Member

  • PipPipPipPipPipPipPipPipPip
  • Shredder
  • Shredder
  • 4,806 posts

Posted 09 February 2016 - 12:35 PM

View PostTexAce, on 09 February 2016 - 06:52 AM, said:

Thanks a lot for making that video and demo!



Who is saying the mech is not shutting down when your heat exceeds? Its nowhere stated like that. People these days....


I'd like to enlighten you that I implied that people wouldn't shut down due to them no longer hitting ghost heat. Right now if you tried to fire 12 clan small pulse lasers you'd generate over 76 heat, without the ghost heat you'd generate 36, a much more manageable damage.

Seeing as its cone of fire all you'd see is people face hugging with masses of small weapons such as those and pulling out close range alpha strikes of upwards of 72 damage. Right now if someone tries that they fry for a moment.

But replying to this further will do just about as much as replying to a troll thread, you guys are stuck in your opinion.

#30 nehebkau

    Member

  • PipPipPipPipPipPipPipPipPip
  • 3,386 posts
  • LocationIn a water-rights dispute with a Beaver

Posted 09 February 2016 - 12:49 PM

View PostTexAce, on 09 February 2016 - 08:14 AM, said:

Love it how someone finally posts a video of a working CoF as Bill intended and all the complainers can say now is "it takes too much time to code"

Seriously guys you are starting to crack me up. You are running out of arguments.



We are just giving you reality -- and an applet isn't an in-game fire control system. I can write an app that simulates a SCADA system in an afternoon -- It would take me a year (or more) to do a real one.


To the person that did the video: Excellent job in showing what you mean. I like the idea but as the above poster said, you would simply see a face-hugging match. Get close enough with SRMs and they mostly hit the same spot -- as it stands now -- and it would be no different with your method I believe.

Edited by nehebkau, 09 February 2016 - 01:28 PM.


#31 no one

    Member

  • PipPipPipPipPipPipPip
  • Bad Company
  • 533 posts

Posted 09 February 2016 - 02:37 PM

View PostTroutmonkey, on 09 February 2016 - 06:24 AM, said:

(and losing convergence is really bad, as per my last demo).


If I may, that's only true if you completely remove convergence on the weapons. There's also the option of adding some factor of TCL to the focal range when you alpha, thereby putting the weapon focus some defined x*tcl dozen or hundred meters behind the target. That way you don't get complete divergence or randomized fire. Just a consideration. The videos are lovely, is the source code available anywhere?

Edited by no one, 09 February 2016 - 02:39 PM.


#32 Troutmonkey

    Member

  • PipPipPipPipPipPipPipPipPip
  • Moderate Giver
  • Moderate Giver
  • 3,776 posts
  • LocationAdelaide, Australia

Posted 09 February 2016 - 02:56 PM

View PostTarogato, on 09 February 2016 - 06:46 AM, said:

Reduce the size of the cone by about 35% (maybe 40%) and that's similar to what I'd like to see in MWO. OR---- alternative, have the spread be determined by your rangefinder. So that you get the same spread on the same size target regardless of your distance from it.

Then take a cone about 10% of that size and apply it to movement (the faster you go, closer to 10% of your already demonstrated spread)

The spread is already calculated this way. It takes the position under the cursor and adds the spread to that. If the spread is say 1m, it will be 1m whether the target is 100m or 600m.

As for movement, no. I refuse to penalize movement in anyway shape or form because movement isn't an issue that needs addressing. I know it's common place in many other games but hurting movement in MWO will encourage peak and shoot and we do not want that.

#33 Troutmonkey

    Member

  • PipPipPipPipPipPipPipPipPip
  • Moderate Giver
  • Moderate Giver
  • 3,776 posts
  • LocationAdelaide, Australia

Posted 09 February 2016 - 03:07 PM

View Postno one, on 09 February 2016 - 02:37 PM, said:


If I may, that's only true if you completely remove convergence on the weapons. There's also the option of adding some factor of TCL to the focal range when you alpha, thereby putting the weapon focus some defined x*tcl dozen or hundred meters behind the target. That way you don't get complete divergence or randomized fire. Just a consideration.

When I was writing the fixed and delayed convergence I realised that these methods both require extra raycast and some more steps. It's possible to add them to this project, but I don't think they will add much compared to the extra overhead they require. What I'll likely do is just reduce the amount of spread, as the demo has a fairly large one. The other part is that the randomized focus points are in a sphere around the target, which means for larger radius it may focus on a closer point and appear less accurate than if I'd used a circle for randomisation. Sphere was just easier for this demo

View Postno one, on 09 February 2016 - 02:37 PM, said:

The videos are lovely, is the source code available anywhere?

I have the source for the previous demo up, but I've yet to commit these changes. I should be able to add it after work today. Keep in mind that I normally write much neater code with better project setups. This was just a quick and dirt solution ;)

#34 Troutmonkey

    Member

  • PipPipPipPipPipPipPipPipPip
  • Moderate Giver
  • Moderate Giver
  • 3,776 posts
  • LocationAdelaide, Australia

Posted 09 February 2016 - 03:11 PM

View PostDakota1000, on 09 February 2016 - 12:35 PM, said:


I'd like to enlighten you that I implied that people wouldn't shut down due to them no longer hitting ghost heat. Right now if you tried to fire 12 clan small pulse lasers you'd generate over 76 heat, without the ghost heat you'd generate 36, a much more manageable damage.

Seeing as its cone of fire all you'd see is people face hugging with masses of small weapons such as those and pulling out close range alpha strikes of upwards of 72 damage. Right now if someone tries that they fry for a moment.

But replying to this further will do just about as much as replying to a troll thread, you guys are stuck in your opinion.

Yes, but under the current system, they'd jump out from behind a bush, fire of 72 damage to a single pixel, instantly blow off your side torso and kill you if your were running an XL. Then they'd shut down.
Under this CoF system they'd jump out, fire off their 72 point alpha, have it spread across your arm, ST, CT and the dirt beside you but not shutdown. They'd still be running powered up, but they'd have dealt far less meaningful damage and still be running pretty hot.

#35 Troutmonkey

    Member

  • PipPipPipPipPipPipPipPipPip
  • Moderate Giver
  • Moderate Giver
  • 3,776 posts
  • LocationAdelaide, Australia

Posted 09 February 2016 - 03:21 PM

View PosttortuousGoddess, on 09 February 2016 - 07:05 AM, said:

An interesting idea, however it seems likely to me that PGI probably wouldn't find the daunting amount of work to complete it acceptable. Adding in an entirely new stat is hard enough, but that on top of adding necessary components to the UI, coding in all the new mechanics, and the extensive public testing it would require? That's a pretty serious amount of resources and labor that PGI would likely rather put toward the new engine. Perhaps if you simplified it, but as is there is just too much work required.

Yes I agree there is a lot of work to do testing wise, but adding in the mechanics and UI is rather simple as I have just displayed (Server code may require more work, I'm not sure what their netcode is like). A simpler system would be nice, but we already have a "simpler" system called Ghost Heat and it's too easy to work around it. This system would be impossible to bypass with different combinations of weapon systems. It's a future proof system that can be used to address most balance problems for the forseable future.
Spoiler


If I were to balance it I would start off and say that alpha of "~25" are okay, and that anything above that would need some spread. Then you would set the Targeting Computers limit to 25, and then set each weapons CPU to the same as it's damage value. Then you would work out the finer details from there. Reduce the CPU for LBX weapons and SRMs, increase Gauss and PPC CPU slightly.

After that point you would take it to the test server and gather metrics. You can check if the maximum cone is too large or small. If any weapons are too inaccurate/powerful you can decrease/increase their CPU. If any mechs are too heavily effected you could increase their total CPU limit.

Then, you can start to use the system to add value to other less used in game items like the Command Console and the Clan Targeting Computers. These items could improve the CPU cap at the expense of tonnage and slots in addition to their other benefits.

Edited by Troutmonkey, 09 February 2016 - 03:27 PM.


#36 keith

    Member

  • PipPipPipPipPipPipPipPip
  • Legendary Founder
  • Legendary Founder
  • 1,272 posts

Posted 09 February 2016 - 03:33 PM

View PostRhialto, on 09 February 2016 - 09:57 AM, said:

It makes no sense to have lasers that move or deviate like this...


and it makes no sense to have mechs on the battlefield instead of tanks. hey this is a game right:P the vid looks spot on for something that would work

#37 Troutmonkey

    Member

  • PipPipPipPipPipPipPipPipPip
  • Moderate Giver
  • Moderate Giver
  • 3,776 posts
  • LocationAdelaide, Australia

Posted 09 February 2016 - 03:34 PM

View PostSader325, on 09 February 2016 - 09:12 AM, said:

We have lasers IRL that can shoot down missiles kilometers away.

This video doesn't make any sense.


Balance / Gameplay trumps realism. If this game were realistic it would not feature giant robots who are physically impossible. They would sink into the ground and fall over due to weight distribution.

#38 White Bear 84

    Member

  • PipPipPipPipPipPipPipPipPip
  • Ace Of Spades
  • Ace Of Spades
  • 3,857 posts

Posted 09 February 2016 - 03:59 PM

View PostQuaamik, on 09 February 2016 - 08:09 AM, said:

It is too much coding and testing, especially when there are simpler ways already in game to limit the high pinpoint alphas. Weapon charge up (see Gause) Projectile velocity (true, gave us laser vomit, but fixed most of the rest) Ghost heat


Charge up - effectively ruined the gauss as a weapon, in addition to its glass cannon status.
P-Vel - Kinda makes sense, on the fence on this one.
Ghost Heat - Quite understated as the most despised mechanic in game.

Simplest is not always best.. ;)

#39 Grimlox

    Member

  • PipPipPipPipPipPipPip
  • The 1 Percent
  • The 1 Percent
  • 511 posts
  • LocationBC, Canada

Posted 09 February 2016 - 04:15 PM

I like the idea. I think it might be more useful to have it based on the amount of damage as well as the number of weapons so a 3 LPL build would suffer as much as 6 mlas. Also probably would require some kind of quirk to soften the blow to some mechs like the nova for example... not entirely but to just soften the blow a bit.

#40 Sadist Cain

    Member

  • PipPipPipPipPipPipPip
  • Philanthropist
  • 605 posts

Posted 09 February 2016 - 04:22 PM

love it, would make things much more interesting





6 user(s) are reading this topic

0 members, 6 guests, 0 anonymous users