30 replies to this topic

[Koniving]

Pinpoint convergence is a problem.

It's been something that's very clear ever since the introduction of host state rewind, where 'Delayed Convergence' was ultimately removed and the "Pinpoint convergence" Elite Skill became a 3,000 XP "tax" with its yield removed.

Back in the day, if I were to shift aim from something 700 meters away to an Atlas 50 meters away, I would have to wait for my weapons to align. If I did not, my arm weapons would hit your arms and my torso weapons would hit your torsos.

If we did the same thing and you were in a Hunchback, my arm weapons would go completely around you and my torso weapons would probably do some glitched out criss-cross X.

What it did was teach a little bit of firing discipline and added an element into the game. It was removed for two reasons. One, it messed with HSR. Your convergence was always shifting. Two, it caused issues for players with high ping.

However PGI's solution of removing it whether temporary or permanent completely sucks.

All weapons converge to hit a specific point in the air, and the weapons appear to adjust course to hit even after being fired. For example if I aim at your arm from 900 meters with AC/2s and the wall beyond you is 1,200 meters, then you twist your arm out of the way in the nick of time... the shots are already fired but instead of the weapons converging where your arm was and then scattering out, they will hit as if they were always aimed at the 1,200 meter wall in a perfect pinpoint spot.

That means at some point when the intended point of convergence failed, the bullets ignored that point and continued on to converge in a pinpoint pile beyond it.

This isn't just pinpoint, it's physics defying pinpoint.

We need another solution.
Seriously a targeting computer shouldn't be giving enhanced range to lasers, should not be making bullets faster. Faster locks? Sure. But since when does a computer pack extra powder into rounds? When does a computer behave as a capacitor to give lasers extra burn?

Now a targeting computer giving faster convergence? Possibly. Giving more options to convergence levels? Definitely.

So what about manual convergence?

Spoiler

Button options to manually establish where we want our convergence. No ping reliance. No HSR issues (as they can detect where we establish it), and we can use intervals of 100 or 200 meters. Tighter intervals for targeting computers and command consoles.

The only pain is manually setting these and knowing to set them. We'd want a simple system like the mouse wheel where we can quickly scroll to the desired range and then start firing. Obviously weapons such as LRMs and streaks when locked would ignore this convergence.

This would actually open us up to more skill based tricks, such as "criss-cross trickshots" where we fire to converge prematurely and the shots just X through to hit two separate targets, and SRM scattering tricks. It'd be neat to say the least.

Or what about the third person targeting reticule?

Spoiler

It's drawn (unlike the first person one) directly to the cockpit of the third person battlemech model which bobs, bounces, side-leans, etc, as the mech moves.

The first person camera? Pfft. No matter what happens, the reticule is straight. It doesn't bounce, it doesn't weave, it just glides on a magical monorail.

Here's some of the B.S. we can get out of it... and a comparison of the third person camera.
The question is... can we get that in first person? Can the first person model have the exact animations of the third person model, with the reticules in first person drawn from either the pilot's head or the cockpit itself.

This would create a reticule that moves with the mech, simulating the accuracy penalties of Battletech for movement. It would give character to different mechs. It would allow mechs to survive longer. It would make mediums and lights more viable, as their smaller frames would be more difficult to hit as crosshairs bob and weave.

Interestingly enough, aiming would still be 100% pinpoint, leaving a 100% skill based aiming system instead of a cone of fire.

If we combine it with a "not-100%-locked" armlock, then the arms would have only a tiny bit of free movement which would in turn scatter shots a bit more. Sometimes this won't matter, but sometimes this is the difference between life and death.

But more so, a targeting computer and Command Console could reduce the range of that 'bob' by different levels of strengths [very, very small values; but these values would be enough to mean something in the long run].

This would also significantly increase the value of a Gyro Stablizer module, because an AC/20 will cause an Atlas's third person model to jerk an entire 82 degrees to the left or right when hit in the shoulder. This in turn would under normal circumstances also jerk the pilot's twist that way... and so a gyro stablizer would reduce that.

We need something. If weapons fired twice in 10 seconds to our double armor it wouldn't matter. But we fire 3 to more than 14 times in 10 seconds.

We also need a continuation of the hitbox overhaul. Jenners are still walking center torsos. I was LRM'ing a Jenner and he just got obliterated CT with no damage to anything but one side torso, moving at 150+ kph. Why are its side torsos as skinny as its arms? Why not do a logical design like run the side torsos up the sides of the torso? Or a Raven-esque hitbox design?

Just some thoughts I've had.

[EvilCow]

Please stop using logic, this is the wrong place.

[Adiuvo]

Do we really need this reimplemented? By nature of having to be always adjustable it's always going to mess with HSR and screw the game for high ping players.

[Agent of Change]

http://mwomercs.com/...25#entry3413825

That is all.

[Mcgral18]

Adiuvo, on 28 July 2014 - 09:11 AM, said:

Do we really need this reimplemented? By nature of having to be always adjustable it's always going to mess with HSR and screw the game for high ping players.

Is fixing the core of weapon balance issues important?

Maybe a little.

Edited by Mcgral18, 28 July 2014 - 09:29 AM.

[Prezimonto]

Agent of Change, on 28 July 2014 - 09:14 AM, said:

http://mwomercs.com/...25#entry3413825

That is all.

We've head threads like this since closed beta. I've started a few myself. Good luck getting them to listen.

[Adiuvo]

Mcgral18, on 28 July 2014 - 09:28 AM, said:

Is fixing the core of weapon balance issues important?

Maybe a little.

Weapon balance between IS mechs is near perfect besides stuff like pulses and small lasers/flamers. Clans muck it up with their lasers but it's not that bad either.

[Mcgral18]

Adiuvo, on 28 July 2014 - 09:39 AM, said:

Weapon balance between IS mechs is near perfect besides stuff like pulses and small lasers/flamers. Clans muck it up with their lasers but it's not that bad either.

Aside from the massive PP FLD alphas which have plagued the game for...over a year now? Whenever HSR was introduced for ballistics.

SRMs are still lackluster for their hardcapped 270M range, heat system still can't really support laserboats with 20+DHS, since you're still being punished for bringing more than 17 DHS with less than 2.0 dissipation per heatsink, unlike when you bring less than 18.


PP FLD is still the most efficient way of killing.

Edited by Mcgral18, 28 July 2014 - 10:39 AM.

[Wingbreaker]

[Reno Blade]

Easier:
All ACs do multi-bullet bursts,
PPCs do arc to 4 additional areas instead of 2.
PPC and Gauss cooldown increase.

FTFY

Edited by Reno Blade, 28 July 2014 - 10:06 AM.

[Adiuvo]

Mcgral18, on 28 July 2014 - 09:44 AM, said:

Aside from the massive PP FLD alphas which have plagued the game for...over a year now? Whenever HSR was introduced for ballistics.

SRMs are still lackluster for their hardcapped 270M range, heat system still can't really support laserboats with 20+DHS, since you're still being punished for bringing more than 17 DHS with less than 2.0 dissipation per heatsinks, unlike when you bring less than 18.


PP FLD is still the most efficient way of killing.

The typical PPFLD alpha is 30-35 damage. You can easily push laser alphas above that, generally in the 50ish range.

SRMs work fine. We bring them to multiple drops on multiple maps and they always work fine, and they kill faster than PPFLD mechs. Gauss/PPC is not sustainable in a brawl and PPC/AC5 is just outperformed in one.

[FupDup]

Koniving, on 28 July 2014 - 08:58 AM, said:

-snip-

We need something. If weapons fired twice in 10 seconds to our double armor it wouldn't matter. But we fire 3 to more than 14 times in 10 seconds.
---

I'm going to comment on this part specifically for now.


Here's a fun experiment to try:
--------------------------------------------------------------------------------------
Try playing a custom match of TT where all of your weapons fired hit the CT with at least 70% accuracy (maybe go all the way up to 100 for experimental purposes). Robots would die a lot faster than if you made guns fire a few more times each turn but each individual shot was subjected to an RNG 2d6 dice roll (meaning each shot hits a random location or even misses entirely). For fairness, don't count critical hits in the experiment because TT's crit system would skew things towards the guns that fire faster rather than the ones with flawless accuracy. Try it at home!
--------------------------------------------------------------------------------------

Convergence is the issue here. Rates of fire sometimes cause some weird things like heat efficiency strangeness, but overall weapons that fire quickly are far easier to counter than weapons that put huge amounts of damage into a single spot instantly. "Damage per 10 seconds" is a fairly meaningless statistic unless you're constantly exposing yourself to enemy fire for literally 10 seconds straight...if you do that, you're probably playing this game with a steering wheel.

If you only stay exposed for ~1-3 seconds, you generally receive less "proportional" damage than TT (i.e. 30 point alpha from a Dragon Slayer translates to a single Gauss Rifle in TT) but pinpoint builds still crush everything. There's a reason that 2 ERPPC/2 Gauss Daishis are feared more than 6 UAC/5 Daishis, and it isn't because of the damage dealt in an arbitrary 10 second window.

Judging by the fearsome-ness of high alpha builds like the Dragon Slayer and the 2 ERPPC/2Gauss Dire Whale, "damage per zero seconds" is far more important and meaningful than "damage per ten seconds" ever will be (because you don't need to stay exposed for 10 full seconds like a dumbass).

Edited by FupDup, 28 July 2014 - 10:49 AM.

[Mcgral18]

Adiuvo, on 28 July 2014 - 10:07 AM, said:

The typical PPFLD alpha is 30-35 damage. You can easily push laser alphas above that, generally in the 50ish range.

SRMs work fine. We bring them to multiple drops on multiple maps and they always work fine, and they kill faster than PPFLD mechs. Gauss/PPC is not sustainable in a brawl and PPC/AC5 is just outperformed in one.

Laser alphas will never be focused on a single location, and their effectiveness is dictated by the target, as you should know.

SRMs are still pretty mediocre. Half weight cSRMs are a bit more worthwhile than the IS ones, but they still need something. Dead-Fire warheads maybe?

They kill faster under 270M, as you say. Completely useless beyond that point, and with their slow travel speed it can be difficult to hit before 200M against fast targets.

I feel it's more the JJ nerf that has affected the PPC combo's more than anything else, along with the Clans which can't mount it very effectively. Hot and ammo starved for the TimberTart, before the JJ nerf. Dire Whale is just as it's always been. Dakka kills them both quickly.


Besides, convergence nerfs would be a big buff to SRMs as well because of their CoF.

[Adiuvo]

Mcgral18, on 28 July 2014 - 10:30 AM, said:

Laser alphas will never be focused on a single location, and their effectiveness is dictated by the target, as you should know.

SRMs are still pretty mediocre. Half weight cSRMs are a bit more worthwhile than the IS ones, but they still need something. Dead-Fire warheads maybe?

They kill faster under 270M, as you say. Completely useless beyond that point, and with their slow travel speed it can be difficult to hit before 200M against fast targets.

I feel it's more the JJ nerf that has affected the PPC combo's more than anything else, along with the Clans which can't mount it very effectively. Hot and ammo starved for the TimberTart, before the JJ nerf. Dire Whale is just as it's always been. Dakka kills them both quickly.


Besides, convergence nerfs would be a big buff to SRMs as well because of their CoF.

The important thing with lasers is that they're hitscan which gives them some benefits in long range. Previously it was completely stupid to build a laser long range deck due to the effectiveness of jumpsniping but that isn't the case anymore. Close range lasers are just better unless you're doing something like all larges. Torso twisting isn't possible from multiple angles - this isn't a 1v1 game. It's supremely effective in the PPC/AC trades we've been doing for ages, but during a brawl the effect is mitigated due to the number of eyes on target. The DPS advantage of lasers can shine in this environment.

SRMs are really, really not mediocre. Their DPS is massive and the spread really is not that big of a concern once you get close enough, which, if properly played, will happen even in the highest level matches.

JJ nerfs have affected the PPC combo, and coupled with the SRM fix is why IS balance is pretty good now.

[Zyllos]

Koniving, on 28 July 2014 - 08:58 AM, said:

...
All weapons converge to hit a specific point in the air, and the weapons appear to adjust course to hit even after being fired. For example if I aim at your arm from 900 meters with AC/2s and the wall beyond you is 1,200 meters, then you twist your arm out of the way in the nick of time... the shots are already fired but instead of the weapons converging where your arm was and then scattering out, they will hit as if they were always aimed at the 1,200 meter wall in a perfect pinpoint spot.
...

I have never seen this before. So your telling me that if I fire two AC/2s at the same time at something 800m out, and once the shells reach 800m, the target isn't there, they will "redefine" their ballistic path to converge onto whatever range the next target was (that be physically another target or terrain)?

I could not see the above happening because that would mean that if your target was shifted to your right (the target's left) just slightly, making the converging point a miss, then the AC/2 would have to recalibrate their trajectory, very possibly into the target's path again for the right AC/2 shell.

BUT, I wholeheartedly agree that convergence is a huge issue with weapons now.

I really think a CoF is truly the only way to fix the problem. Every other way to fix convergence still leaves weapons all hitting a single location in some way (except for making every weapon firing individually).

I would think a good CoF system would be section based (LA, RA, LT, RT, CT, H). This means that all weapons within a specific section would randomize to a single point for each section. So if you fire 5 weapons in the LA and 1 weapon in the RA, you would have 2 randomized points, the 5 in the LA and 1 in the RA.

That would, at it's simplest, would reduce pin point damage unless they are all mounted in a single location.

I would like to take it a step further and make lasers fire in right angles from their mounting points and ballistic weapons have small CoF for themselves, on top of the CoF for a section but that is just me.

[Joseph Mallan]

Koniving, on 28 July 2014 - 08:58 AM, said:

Pinpoint convergence is a problem.

It's been something that's very clear ever since the introduction of host state rewind, where 'Delayed Convergence' was ultimately removed and the "Pinpoint convergence" Elite Skill became a 3,000 XP "tax" with its yield removed.

Back in the day, if I were to shift aim from something 700 meters away to an Atlas 50 meters away, I would have to wait for my weapons to align. If I did not, my arm weapons would hit your arms and my torso weapons would hit your torsos.

If we did the same thing and you were in a Hunchback, my arm weapons would go completely around you and my torso weapons would probably do some glitched out criss-cross X.

What it did was teach a little bit of firing discipline and added an element into the game. It was removed for two reasons. One, it messed with HSR. Your convergence was always shifting. Two, it caused issues for players with high ping.

However PGI's solution of removing it whether temporary or permanent completely sucks.

All weapons converge to hit a specific point in the air, and the weapons appear to adjust course to hit even after being fired. For example if I aim at your arm from 900 meters with AC/2s and the wall beyond you is 1,200 meters, then you twist your arm out of the way in the nick of time... the shots are already fired but instead of the weapons converging where your arm was and then scattering out, they will hit as if they were always aimed at the 1,200 meter wall in a perfect pinpoint spot.

That means at some point when the intended point of convergence failed, the bullets ignored that point and continued on to converge in a pinpoint pile beyond it.

This isn't just pinpoint, it's physics defying pinpoint.

We need another solution.
Seriously a targeting computer shouldn't be giving enhanced range to lasers, should not be making bullets faster. Faster locks? Sure. But since when does a computer pack extra powder into rounds? When does a computer behave as a capacitor to give lasers extra burn?

Now a targeting computer giving faster convergence? Possibly. Giving more options to convergence levels? Definitely.

So what about manual convergence?

Spoiler

Button options to manually establish where we want our convergence. No ping reliance. No HSR issues (as they can detect where we establish it), and we can use intervals of 100 or 200 meters. Tighter intervals for targeting computers and command consoles.

The only pain is manually setting these and knowing to set them. We'd want a simple system like the mouse wheel where we can quickly scroll to the desired range and then start firing. Obviously weapons such as LRMs and streaks when locked would ignore this convergence.

This would actually open us up to more skill based tricks, such as "criss-cross trickshots" where we fire to converge prematurely and the shots just X through to hit two separate targets, and SRM scattering tricks. It'd be neat to say the least.

Or what about the third person targeting reticule?

Spoiler

It's drawn (unlike the first person one) directly to the cockpit of the third person battlemech model which bobs, bounces, side-leans, etc, as the mech moves.

The first person camera? Pfft. No matter what happens, the reticule is straight. It doesn't bounce, it doesn't weave, it just glides on a magical monorail.

Here's some of the B.S. we can get out of it... and a comparison of the third person camera.
The question is... can we get that in first person? Can the first person model have the exact animations of the third person model, with the reticules in first person drawn from either the pilot's head or the cockpit itself.

This would create a reticule that moves with the mech, simulating the accuracy penalties of Battletech for movement. It would give character to different mechs. It would allow mechs to survive longer. It would make mediums and lights more viable, as their smaller frames would be more difficult to hit as crosshairs bob and weave.

Interestingly enough, aiming would still be 100% pinpoint, leaving a 100% skill based aiming system instead of a cone of fire.

If we combine it with a "not-100%-locked" armlock, then the arms would have only a tiny bit of free movement which would in turn scatter shots a bit more. Sometimes this won't matter, but sometimes this is the difference between life and death.

But more so, a targeting computer and Command Console could reduce the range of that 'bob' by different levels of strengths [very, very small values; but these values would be enough to mean something in the long run].

This would also significantly increase the value of a Gyro Stablizer module, because an AC/20 will cause an Atlas's third person model to jerk an entire 82 degrees to the left or right when hit in the shoulder. This in turn would under normal circumstances also jerk the pilot's twist that way... and so a gyro stablizer would reduce that.

We need something. If weapons fired twice in 10 seconds to our double armor it wouldn't matter. But we fire 3 to more than 14 times in 10 seconds.

We also need a continuation of the hitbox overhaul. Jenners are still walking center torsos. I was LRM'ing a Jenner and he just got obliterated CT with no damage to anything but one side torso, moving at 150+ kph. Why are its side torsos as skinny as its arms? Why not do a logical design like run the side torsos up the sides of the torso? Or a Raven-esque hitbox design?

Just some thoughts I've had.

I like how you started out, but you lost me when you said it became a problem with HSR. It was a problem back in Closed Beta, but to many of the players demanded they hit what they aim at... no matter the # of weapons fired. Apparently the armchair soldiers have better aim than a naval broadside.

Heck even GPS GUIDED munitions are only accurate within a +5.5 Meter radius!

[Agent of Change]

This is the most basic, simplest solution i have been able to come up with:

Step 1: Create Variable Convergence set at a specific point prior to the match, if you want to get fancy set arm and torso convergence separately. This is a tried and true method in bot real life and video game mirroring how war machines with multiple fixed fire points create their convergence, ex. Fighter aircraft.

Step 2: Eliminate the "Arm lock" entirely after the first 24 matches, including the quick center functionality. This was intended to help new players and as always happens is being abused by the PPFLD crowd, if they want perfect arm and torso convergence i'll tell them what they tell everyone else, "get good".

Step 3: Leave everything else alone and see what happens on a PTS. Odds are you'll be able to de-nerf several weapons after those two changes, but allow people to test it for a bit on a PTS and actually listen to their feedback.

Step 4: (For PGI) Remember that even though you can't charge people directly for this and that means you'd usually blow this off and release a new skin or something, think about the fact that a solid game that doesn't piss people off will be a earnings booster.

Edited by Agent of Change, 28 July 2014 - 12:17 PM.

[Shae Starfyre]

Agent of Change, on 28 July 2014 - 11:36 AM, said:

...

Step 4: (For PGI) Remember that even those you can't charge people directly for this and that means you'd usually blow this off and release a new skin or something, think about the fact that a solid game that doesn't piss people off will be a earnings booster.

Uh, huh...

MC Only Module: 10,000 MC for a module that turns off Arm Lock and Centering, and allows for no convergence... I'll pay!

Module Name: The Chuck Norris Module.

Edited by Aphoticus, 28 July 2014 - 11:52 AM.

[cSand]

If you guys really think this is that big a deal:


Get rid of arm lock

That convergence efficiency we can unlock can now increase the speed that your arms recentre

boom

"problem" solved

This is supposed to be the year 3000, it's not like we need a 3 man crew in a mech adjusting rangefinders and stuff.

I'm here all day for your consulting needs

Edited by cSand, 28 July 2014 - 11:56 AM.

[Joseph Mallan]

cSand, on 28 July 2014 - 11:52 AM, said:

Skip all your steps except 2.

Get rid of arm lock
boom

"problem" solved

I'm here all day for your consulting needs

You forgot, "Tip your Waitress!"