57 replies to this topic

[Y E O N N E]

Class, dismissed!

[ROSS-128]

Yep, that exactly. Bad geometry with spread is still bad geometry. And when you further consider scale, such as how you can probably fit an entire Locust inside some assault mechs' torsos, well the implications there could be quite bad.

As far as arguing from tabletop behavior... I think trying to replicate tabletop accuracy in a shooter is a pipe dream. In tabletop you can find yourself with a 50% chance or worse of missing the entire mech from punching distance, because apparently everyone in the Battletech universe is cross-eyed and has Parkinson's. Must be where Darth Helmet found his gunners.

[Tarogato]

ROSS-128, on 09 January 2018 - 06:20 PM, said:

Yep, that exactly. Bad geometry with spread is still bad geometry. And when you further consider scale, such as how you can probably fit an entire Locust inside some assault mechs' torsos, well the implications there could be quite bad.

As far as arguing from tabletop behavior... I think trying to replicate tabletop accuracy in a shooter is a pipe dream. In tabletop you can find yourself with a 50% chance or worse of missing the entire mech from punching distance, because apparently everyone in the Battletech universe is cross-eyed and has Parkinson's. Must be where Darth Helmet found his gunners.

Yadda-yadda something about aiming weapons manually is nearly impossible and it's all done by the mechs' computers but all mechs have some degree of built-in electric counter-measures that work against computer tracking so weapon accuracy is shite all over and weapon ranges are listed based on maintaining computer tracking in a battle rather than the weapons' actual physical effective ranges. Something to that effect, I think.

[Tombstoner]

ROSS-128, on 09 January 2018 - 06:20 PM, said:

Yep, that exactly. Bad geometry with spread is still bad geometry. And when you further consider scale, such as how you can probably fit an entire Locust inside some assault mechs' torsos, well the implications there could be quite bad.

As far as arguing from tabletop behavior... I think trying to replicate tabletop accuracy in a shooter is a pipe dream. In tabletop you can find yourself with a 50% chance or worse of missing the entire mech from punching distance, because apparently everyone in the Battletech universe is cross-eyed and has Parkinson's. Must be where Darth Helmet found his gunners.

The only way to compensate for bad geometry is increaced durability. Speed/geometry interacts affecting survivability. The issue has been and always will be 6 weapons hitting the same location turning 6x medium lasers into a single medium laser 30. They kind a tried this with quirks with some sucess but reached to far and tried to make all mechs perform the same. this failed IMO because hard point type, number and location matter.... a lot. Far more then in TT. The TT to FPS port was fubar from day one.

[justcallme A S H]

Yeonne Greene, on 09 January 2018 - 06:11 PM, said:

Class, dismissed!

Well given CK16 is never coming back

As

Usual...

[SeventhSL]

Tarogato, on 09 January 2018 - 05:58 PM, said:

Listen to what you're saying, and think about it in terms of what actually does happen in the game

Thanks for the diagrams and the meaningful discussion.

I drew the image below to scale and then measured the areas using www.sketchandcalc.com. For ease, I expressed the CT areas as a percentage of the original circle. E.g. A 50 diameter circle fits entirely inside of a CT with a width of 50 so the CT area is 100% of the original circle.



The interesting part is the change in percentage as the circle grows larger. As you can see, with the small circle the wider CT takes 100% while the narrower CT takes 60%. That is a 40% difference. When the circle is bigger the wider CT takes 60% and the narrower CT takes 30%. That is a 30% difference.

The narrower CT will always have an advantage over the wider one but as accuracy decreases and the circle gets larger the advantage decreases. Of course there will be a point at which this trend swings the other way again.

If my calcs are wrong please let me know and I'll confirm. I care more about knowing the answer than I do about being right or wrong.

Edited by SeventhSL, 10 January 2018 - 04:34 AM.

[ROSS-128]

But by the time the circle is large enough to do that on the largest mechs, what does that mean for everything else? And since cone of fire is, well, a cone, what does it mean for effective weapon ranges? People might just close distance until their weapons are effectively pinpoint again. How would it affect weapons whose main selling point is range, like Gauss, PPCs, and ERLLs? They'd either remain pinpoint and be go-to weapons for that reason (more so than the ERLL already is) or their range would exist only on paper, and they'd lose their role.

[Sigmar Sich]

What if cone of fire would not be constant, but tended to converge to a pinpoint shot? Move fast - cone is bigger. Shoot small weapon - cone is slightly bigger, shoot big weapon - cone is much bigger. Stand still or move slowly - cone getting smaller. Lock target - cone getting smaller faster. Got targeting info - cone goes pinpoint even faster. Overheat - cone convergence is reduced. Take crit to gyro - get bigger convergence penalties for movement.
I think adaptive cone of fire would be way to go. Though i fear it is too late for this. Only chance of such major change is if MWO loses so much players, so PGI would have nothing to lose.

[Kaeb Odellas]

Mole, on 08 January 2018 - 11:00 AM, said:

Hasn't PGI stated multiple times that a convergance system is beyond their coding capabilities? I'm pretty sure they said that. It's pathetic, but it's true.

I don't think that's quite true. I believe what they said was that they couldn't get non-instant convergence to play nicely with HSR. I think that was their reasoning for dropping inverse kinematics as well.

[Tarogato]

SeventhSL, on 10 January 2018 - 04:24 AM, said:

I drew the image below to scale and then measured the areas using www.sketchandcalc.com. For ease, I expressed the CT areas as a percentage of the original circle. E.g. A 50 diameter circle fits entirely inside of a CT with a width of 50 so the CT area is 100% of the original circle.

[neato visual]

The interesting part is the change in percentage as the circle grows larger. As you can see, with the small circle the wider CT takes 100% while the narrower CT takes 60%. That is a 40% difference. When the circle is bigger the wider CT takes 60% and the narrower CT takes 30%. That is a 30% difference.

The narrower CT will always have an advantage over the wider one but as accuracy decreases and the circle gets larger the advantage decreases. Of course there will be a point at which this trend swings the other way again.

If my calcs are wrong please let me know and I'll confirm. I care more about knowing the answer than I do about being right or wrong.

I like the visualisation!

You got me second guessing myself, so I decided to try it figure this out for realsies, since I've just been going off intuition.

I took your idea of taking a circle with radius of a given weapon spread amount, and trimming it to the limits of a given width (representing how most hitboxes in MWO are like vertical columns).

Thank you wikipedia:




So I can just take the area of the circle (which represents weapon spread), and remove two times the area of that segment 'A', according to the measurement 'd' which represents the width of a mech hitbox. Then I just compare the remaining area to the original area as a percentage, and I can extrapolate that for all 1:1 combinations thusly:



(sorry for low resolution) This should show what I mean by how mechs with good geometry benefit foremost. Let's extract an example:



Let's take the middle rectangle, which is for a mech with the arbitrary hitbox width of "7". (I'm going to eliminate some potential confusion by removing percents from the weapon damage dealt. Let's just pretend that 100% means 100 damage for now, otherwise we'll have to talk about percents of percents, which is irritating.) At the left with a weapon radius of "21", the weapon deals 41.6 damage on target. If we increase the spread by a radius of 5, to a final radius of "26", the weapon deals 33.9 damage, which is 18.5% less (a difference of 7.7 damage). Let's lay the numbers out for all three rectangles:


Mech with thin hitboxes (top rectangle)
- Hitbox = 3 width
- From 21 radius = 18.1 damage
- ... to 26 radius = 14.5 damage
- Difference of 3.6 damage (a 19.9% improvement)

Mech with average hitboxes (middle rectangle)
- Hitbox = 7 width
- From 21 radius = 41.6 damage
- ... to 26 radius = 33.9 damage
- Difference of 7.7 damage (a 18.5% improvement)

Mech with wide hitboxes (bottom rectangle)
- Hitbox = 11 width
- From 21 radius = 63.5 damage
- ... to 26 radius = 52.2 damage
- Difference of 11.3 damage (a 17.8% improvement)


As you can see, as we go from thinner smaller hitboxes to larger hitboxes, while the difference in applied damage grows, the actual percentage improvement diminishes. That is to say... mechs with already "good" hitboxes, benefit slightly more from increased weapon spread compared to mechs with "bad" hitboxes.


In your example with the diagrams you made, you take a CT width of "50". The small concentrated circle deals 100% damage, while the big circle deals 60% damage, which is 40% less (or 40% 'better') With the CT of width only "25", the small circle deals 60% damage, and the big circle deals 30% damage, which is 50% less (or 50% 'better'). The smaller CT in your example still got the better end of the bargain.

[The Lighthouse]

SeventhSL, on 10 January 2018 - 04:24 AM, said:

Thanks for the diagrams and the meaningful discussion.

I drew the image below to scale and then measured the areas using www.sketchandcalc.com. For ease, I expressed the CT areas as a percentage of the original circle. E.g. A 50 diameter circle fits entirely inside of a CT with a width of 50 so the CT area is 100% of the original circle.



The interesting part is the change in percentage as the circle grows larger. As you can see, with the small circle the wider CT takes 100% while the narrower CT takes 60%. That is a 40% difference. When the circle is bigger the wider CT takes 60% and the narrower CT takes 30%. That is a 30% difference.

The narrower CT will always have an advantage over the wider one but as accuracy decreases and the circle gets larger the advantage decreases. Of course there will be a point at which this trend swings the other way again.

If my calcs are wrong please let me know and I'll confirm. I care more about knowing the answer than I do about being right or wrong.

Taro has already explained to you, but to say simply, you got yourself into percentage fallacy, a.k.a "is it 50% better or 33.3% better?" Problem. Basically, you are confused with "50% better or 50% smaller". Those two things are not same at all.

This is extremely common mistake, and as a person who mods game called Stellaris, I do deal with these problem all the time (basically the game's developer, Paradox, mixes these two, causing a lot of errors and wrong numbers.)

Edited by The Lighthouse, 10 January 2018 - 09:13 PM.

[SeventhSL]

@Tarogato thanks so much for going to the trouble to explain this. I concede that you are in fact correct and that an increase in spread will not help bad geometry in regards to taking proportionally less damage to oversidzed hitboxes.

Just to be clear, I still believe that adding an accuracy mechanic would be a huge plus as trade offs for things like alpha size, speed and poor hard point placement but it won't help hitboxes.

[justcallme A S H]

All this hitbox stuff is well and good.

But there is still one burning question left....

Yeonne Greene, on 07 January 2018 - 10:20 PM, said:

Now, the real question is whether or not CK16 will come back to this thread to defend his position, or if he'll just abandon his thread as is his usual modus operandi as a Hot-Take Discussion bomber.

Where is he????


I for one am still waiting for the response to all of this.

Edited by justcallme A S H, 14 January 2018 - 03:40 PM.

[The Lighthouse]

justcallme A S H, on 14 January 2018 - 03:40 PM, said:

All this hitbox stuff is well and good.

But there is still one burning question left....





Where is he????


I for one am still waiting for the response to all of this.

You are wasting your time.

[Tarogato]

SeventhSL, on 14 January 2018 - 03:22 PM, said:

Just to be clear, I still believe that adding an accuracy mechanic would be a huge plus as trade offs for things like alpha size, speed and poor hard point placement but it won't help hitboxes.

I agree. I don't want to add just a generic spread or RNG cone to weapons, because that doesn't really have *meaningful* impact, and like I said has the effect of slightly benefiting already well-off mechs. But I kinda yearn for some sort of accuracy mechanic... like delayed convergence (which PGI couldn't get to play nice with their lag-compensation).

[Yosharian]

FupDup, on 07 January 2018 - 08:54 PM, said:

In a world of reduced or remove convergence, having bad geometry will still statistically increase the chances of certain body parts getting hit.

Was gonna post exactly this

logic fail

[Lily from animove]

FupDup, on 07 January 2018 - 08:54 PM, said:

In a world of reduced or remove convergence, having bad geometry will still statistically increase the chances of certain body parts getting hit.

this, dunno how one can make a post like Op yet not see this issues, because a SCR vs a Hunchback already defines how broken no PP convergence is. Good aim can at least counter good hitboxes, but spread + good hitboxes = godlike no matter what skill you have.

[Agent1190]

My math hurts.