

Forget Heat Penalties: A Comprehensive Balance Solution To Alphas, Convergence, Poptarts, Boats, And Clans
#441
Posted 03 July 2013 - 06:40 PM
I'm all for the penalties but this would be a great addition too!
#442
Posted 03 July 2013 - 07:42 PM
#443
Posted 04 July 2013 - 08:23 AM
Maverick01, on 03 July 2013 - 06:34 PM, said:
Bill and I talked about that; some older screenshots features a warning message similar to the heat warning. But as TCS dissipates so quickly (and BB is already sloooow, e.g. the infamous "Jump Jet fuel.... at ... twenty-five percent"), this might not be useful.
#444
Posted 04 July 2013 - 08:31 AM
#445
Posted 04 July 2013 - 10:59 AM
Homeless Bill, on 03 July 2013 - 12:41 PM, said:
Just wanted to add that this (gradual ramp up of CoF/Convergence loss at TC>100) is absolutely the way to go and would be freaking perfect.
#446
Posted 04 July 2013 - 04:06 PM
If you implement this system, PGI... you will have not only saved the game from what many would consider an utter disaster... you will have saved your programmers and developers costly labor hours having to implement band-aid solutions, as this system has revealed itself to be self-sustaining for months or years into the future of the lifecycle of this game.
Please make this system part of your 'aggressive series' of changes to weapon systems as we near the launch date, Paul.
#447
Posted 05 July 2013 - 01:32 AM
#448
Posted 05 July 2013 - 06:54 AM
Haji1096, on 20 June 2013 - 04:13 AM, said:
If this solution is not implemented, PGI is basically saying they do not want this game to last.
I do not embellish, because it needs no embellishment, this thread (or at least this topic) is the most important on the forums. After reading through all related threads, this, Docbach's and Jaguars are the most sensible. I pondered on alternative ideas, and came to the same conclusion that convergence is the issue that plagues this game. It's the root cause for the low mech mortality rates and quick matches that last only a third of the allotted time.
#449
Posted 05 July 2013 - 07:20 AM
But PGI already knows this. And if over 200 people think it's a good idea now... think of how many more people who don't read the forums would like it.
This targeting computer system not only works on the fundamental levels of the game, it makes sense in BattleTech canon. Tabletop enthusiasts who likewise want game balance should have a difficult time not accepting this as a plausible solution to problems we all face every day in every match.
This is not a silver bullet solution, and I will not say it is. But it works. It works far better than anything else ever presented. This really is the most important thread on the forums.
That isn't embellishment. That isn't hyperbole. That isn't exaggeration. That isn't emotionally-charged rhetoric.
This is the most important thread on the forums and unless PGI offers a solution of equal or greater value (unlikely), this game will suffer as a result.
And if PGI doesn't think it's a good idea when over 200 of us who were savvy enough to press the 'Like' button on his OP, then I challenge them to debate Homeless Bill and the rest of us on why it isn't a good idea.
Edited by Renaissance, 05 July 2013 - 07:21 AM.
#450
Posted 05 July 2013 - 07:52 AM
Bravo.
#451
Posted 05 July 2013 - 07:52 AM
#452
#453
Posted 05 July 2013 - 10:36 AM
Never mind, I see the posst above me thought the exact same thing. ...voted...
Edited by Nasty McBadman, 05 July 2013 - 10:38 AM.
#454
Posted 05 July 2013 - 10:41 AM
I only wish PGI had the same sense you do.
#455
Posted 05 July 2013 - 11:12 AM
Homeless Bill, on 03 July 2013 - 12:41 PM, said:
As time goes on, I do feel that a gradual loss of convergence between 101 and 150 is probably better than immediate loss of convergence (still with a cone of fire increasing between 101 and 200). It's less jarring, and I don't see it solving the problem any less effectively. It would solve your main complaint of the immediate, boolean punishment for even the slightest error, but it also wouldn't necessitate per-weapon spread stats. Thoughts?
Well, I would just ignore drawing multiple rings and just display how bad your TCL is (like in a bar or indicator), and then based on that, players can get a rough idea of how wide the spread will be by the range of the weapon.
So, lets just say PGI implements this and instead of displaying a true bar or analog value, they just display a colored indicator.
Green < 100 TCL
Green-Yellow 133 >= x >= 100
Yellow 166 >= x > 133
Orange 200 >= x > 166
Red > 200
Green means that your weapons will not produce a CoF.
Green-Yellow means that at optimal range of the weapon, the CoF will be small.
Yellow means that at optimal range of the weapon, the CoF will be medium.
Orange means at optimal range of the weapon, the CoF will be large.
Red means convergence/lock on is lost, the CoF will continue to be large.
***EDIT: The only problem is displaying what weapons would actually cause the TC to go into one of the states listed above. Maybe color code the weapon bar themselves or the weapon group selected when it would go into a specific state? I am going to develop a picture of what I am talking about.***
A small CoF for a weapon would mean that weapons fire is still very likely to hit a Medium sized mech at optimal range and have a reasonable chance of hitting other locations not being aimed at.
A medium CoF for a weapon would mean that weapons fire is likely to hit a Medium sized mech at optimal range and have a good chance of hitting other locations not being aimed at.
A large CoF for a weapon would mean that weapons fire is about change to completely miss a Medium sized mech at optimal range and will most likely hit a location not being aimed at.
That is what I was thinking. The CoF is based off the optimal range of the weapon, thus the accuracy of the weapon. That would also mean that long ranged weaponry will need to be much closer, absolutely, than short ranged weaponry to guarantee hits with a CoF, which I think is a fair trade for the range.
This also will effect brawling effectively because if you happen to not be aiming dead center of a section, then the chances of hitting an adjacent section is pretty good, which can make or break brawling due to being exposed to more weapons fire from other mechs.
Either way, that was my premises behind doing a CoF before losing convergence first.
Edited by Zyllos, 05 July 2013 - 11:16 AM.
#456
Posted 05 July 2013 - 11:41 AM
Hopefully anyone reading the above can understand what I am talking about.
#457
Posted 05 July 2013 - 11:47 AM
JeremyCrow, on 04 July 2013 - 08:31 AM, said:
Fixed =D
Renaissance, on 05 July 2013 - 07:20 AM, said:
But PGI already knows this. And if over 200 people think it's a good idea now... think of how many more people who don't read the forums would like it.
This targeting computer system not only works on the fundamental levels of the game, it makes sense in BattleTech canon. Tabletop enthusiasts who likewise want game balance should have a difficult time not accepting this as a plausible solution to problems we all face every day in every match.
This is not a silver bullet solution, and I will not say it is. But it works. It works far better than anything else ever presented. This really is the most important thread on the forums.
That isn't embellishment. That isn't hyperbole. That isn't exaggeration. That isn't emotionally-charged rhetoric.
This is the most important thread on the forums and unless PGI offers a solution of equal or greater value (unlikely), this game will suffer as a result.
And if PGI doesn't think it's a good idea when over 200 of us who were savvy enough to press the 'Like' button on his OP, then I challenge them to debate Homeless Bill and the rest of us on why it isn't a good idea.
I'll just let you take over from here =P
I appreciate the effusive praise. Honestly, I'm impressed that 200 people even took the time to read it. I never thought it would get this kind of support.
I'd love to see a real response to this thread, but don't hold your breath. I may have a hard-on for debating game design, but most game designers I've worked with get attached to their ideas and take things very personally. I don't see any developer showing up to make the case for two reasons:
- They'd be arguing at a disadvantage. In 23 pages of feedback, no one has been able to poke holes in my proposal's solvency. The only real point of contention they would have is complexity. And I think I'm rather well-prepared for that argument.
- They have the power, and it's a waste of time. My opinion and the opinions of everyone else on the forums are largely irrelevant. They could engage us as equals, but why bother when they're calling the shots? Why defend themselves when it's not required? At worst, it would rile up the community; at best, it's a time-suck.
Roheryn, on 05 July 2013 - 07:52 AM, said:
Bravo.
It took a few months of thinking and tinkering, but at the end, even I was impressed =P
DV McKenna, on 05 July 2013 - 07:52 AM, said:
Thanks =D
If they actually answer my question about release, I'll probably start shamelessly asking for a response in future Ask the Devs.
Sam Donelly, on 05 July 2013 - 10:41 AM, said:
Homeless Bill: scholar, gentlemen, and small-breed, modern-day renaissance millionaire.
Zyllos, on 05 July 2013 - 11:12 AM, said:
The CoF is based off the optimal range of the weapon, thus the accuracy of the weapon. That would also mean that long ranged weaponry will need to be much closer, absolutely, than short ranged weaponry to guarantee hits with a CoF, which I think is a fair trade for the range.
Either way, that was my premises behind doing a CoF before losing convergence first.
I definitely like the cone of fire based off optimal range, but again my big problem is player communication. Sure, they know how bad their TCL is going to be, and anyone that takes the time to get acquainted with the numbers will have no problem understanding. It's new player's I'm worried about.
It's that guy dropping into his first round and watching as two of his weapons take wildly different accuracy penalties. Sure, deductive reasoning should link the number by the weapon and the severity of the accuracy penalty, but it won't. I say it won't because I've seen too many virgin playtests of this game and others. People almost never do what you expect. They always manage to click on the wrong thing or misinterpret information heavily.
Without a way to represent it visually, I see it causing more frustration than even the immediate loss of convergence. Consistency is of higher value to me than easing the players into the penalty.
At the end of the day, our difference is largely a philosophical one, and I'm down with either implementation. I'd be happy if they would just try out any part or variation of this system.
Edited by Homeless Bill, 05 July 2013 - 12:48 PM.
#458
Posted 05 July 2013 - 12:19 PM
#459
Posted 05 July 2013 - 04:15 PM
#460
Posted 05 July 2013 - 05:55 PM
I’m having a hard time finding any fault in this solution. Although I’m terrible when it comes to balance suggestions, I could see this system actually showing quarks in each ‘Mech. For example, the D2j in the Jagermech having a different... I don’t know… TCS cost for certain weapons or something.
Edited by CompproB237, 05 July 2013 - 07:58 PM.
10 user(s) are reading this topic
0 members, 10 guests, 0 anonymous users