69 replies to this topic

[Nightbird]

So, why should we care about physics and mech agility?

1. It might be a fun discussion, if you don't love mechs nor enjoy discussing them then you should probably stop reading here.
2. There is some common sense involved. For example when we see a cheetah next to a giraffe, we automatically can feel which is more agile. If this expectation is proven wrong, as it often does in MWO, we instinctively feel a wrongness.
3. Would gameplay balance and mech "feel" be improved if physics were considered? Read to the end and decide for yourself


I wanted to keep this discussion simple and testable, so instead of math, we'll use some tests that can be done at the gym and illustrated with stick figures. To do the tests by yourself, take some dumbbells around 30% of your weight total (if you weigh 200lbs, 2x35lb dumbbells, if you're 150lbs, 2x25 lb dumbbells, etc), but be safe, don't overdo it, and don't hit anyone around you.

The dumbbells will represent either the center of mass for the weapons equipped (high or low) on a mech for humanoid shaped mechs or the head and rear for long marauder shaped mechs.

Mech Pitch Angle



For A, B, C don't push your rear back to counter balance. Try to control effort, working harder just for one move defeats the purpose of the test.

A: To represent low slung mechs like the Atlas, Cataphract, Highlander put weights at waist and lean forward. Use a mirror if needed. You'll find your balance to be excellent and it is easy to pitch forward 90 degrees or more

B: To represent shoulder mounted mechs like the Blood Asp, Corsair, Battlemaster, put weights over the shoulders by your head and lean forward, you'll find that your balance to be precarious and your pitch greatly limited.

C: To represent chest weapon mechs like the Warhammer, Victor, Cyclops, hold weights at chest height, it feels about halfway between A and B

D: To represent a long mech like a Marauder (IIC), Mad Cat II, Raven, hold with arms fully extended. Move arms up and down without tilting torso to represent the pitch. This rest is common sense, as by having a counter-balance, balance is not really affected by pitch angle with the only limit being physical obstruction.

E: There is no E depicted, which is half way between D and C to represent squat mechs like the Timberwolf, Direwolf, Champion, but pitch should still be good due to low center of gravity and some counter-balancing.

D>A=E>C>B

How it compares to MWO today:
Mechs with low slung weapons are usually not rewarded with better pitch range. For example, the Atlas and Firestarter have some of the worst pitch ranges in game. Mechs with high mounted weapons are usually not punished with less pitch range. For example, the Blood Asp and Corsair and Blackjack. (Notably, the KDK-3 is the only one with the right pitch penalty) Squat and long counter-balanced mechs are also not rewarded with better pitch, for example the Marauder, II, IIC. The Nightstar seems to be one mech that is long and have improved pitch. If for technical reasons pitch is limited to 15-35 degrees, then it could be distributed among mechs in a way that makes more sense.



Mech Pitch Speed

Using the same diagrams.

For A, B, C above, having a better pitch range due to lower center of gravity also carries over to pitch speed. It's fast to bow with A and delicate and slow with B. Pitch down is faster than pitch up since down is gravity assisted.

For D, E there is no gravity assistance due to the counter-balance, so pitch up and down is the same speed. Also, by having mass further away from the center, it takes a little more leverage to pitch compared to humanoid mechs.

A=E>C=D>B


How it compares to MWO today:
Pitch speed seems to be rolled into one with torso yaw speed, see below.


Mech Yaw Angle

There isn't the same physical balance considerations as pitch here, so PGI is free to assign anything they see fit. Though, for a declared walking tank game, I am against mechs that twist less than 90 degrees for obvious reasons.

Mech Yaw Speed



Try to control effort, working harder just for one move defeats the purpose of the test.

For A, B, C twist left and right, and the first thing you notice is that the speed is about the same. This makes sense as the axis of rotation is vertical and the center of mass stays close to the axis of rotation even as it moves up or down.

For D, and implied E (squat mechs), the mass is further away from the axis of rotation and therefore it takes a lot more leverage to twist, making the torso twist movement slower.

A=B=C>E>D

How it compares to MWO today:
Yaw speed and pitch speed being the same is a problem, since the order from best to worst is VERY DIFFERENT. Looking at just Yaw, sometimes the relative benefit is not right, for example the KDK should twist better than the Direwolf, it doesn't. Sometimes, the scale of the difference is not sufficient. Ingame the Atlas twists 15% faster than the Marauder II, but from testing in the gym, maybe 50% to 100% faster would be feel more 'right' and this should be applied to all humanoid versus squat mech (twist 15-30% slower) versus long mech (twist 33-50% slower).

This makes sense not only because of physics in the gym, but also from a mech design perspective. Humanoid mechs are designed to twist to avoid damage, as they shoot from the front but have the slimmest profile from the side to mitigate damage, so they should be turning 90 degrees often. For long mechs, the slimmest profile is already the front, and mitigating damage is done by wiggling the nose, so a full 90 degree turn is rarely needed.


Mech Accel, Decel, Turn



Test both running from a stop, seeing how long it takes to reach a comfortable speed, and running in a circle and seeing how small a circle you can make. Try to control effort, working harder just for one move defeats the purpose of the test.

I think most can guess how this turns out, the higher the center of mass, the slower accel, decel and turn.

A>C>B

D is actually hard to test, since the center of mass representation is different - around our chest - but it is comparable to where the center of mass is in A. Squat mechs (E) also center around the same place as for C.

A=D>C=E>B

The difference between A and B is very significant, with B being half as fast at accelerating and turning as A.

How it compares to MWO today:
Here are some of the mechs with wrong relative accel/decel/turn speed. The Direwolf has a lower center of gravity than the Kodiak yet it is slower. The Blood Asp has a higher center of gravity than the Mad Cat II yet has the same agility. The Mauler and Cyclops have similar centers of gravity yet the Cyclops is almost twice as agile. The grasshopper has a higher center of gravity than the Cataphrat yet is also 50% more agile than it. I will add that for many mechs, the difference is in the right direction, but the scale of the difference is far less than what is experienced in test. (magnitude feels wrong)



So after such a long post, am I asking PGI to implement it? No, we all know PGI isn't updating MWO anymore. I don't even think we should only follow RL physics, just that it is a good starting point and tweaks can follow. Of note is that high weapon mounts have natural penalties due to the high center of mass, something PGI rarely uses to balance mechs. There is also the difference between humanoid and plane shaped mechs that don't really get fleshed out ingame.

I do hope for some of these comparisons, you felt "Oh! That's why that mech feels off" and that is the case with many mechs in MWO and the point of this thread.


Thanks for reading.

Edited by Nightbird, 23 January 2020 - 05:30 PM.

[LordNothing]

there are some areas in physics i think would make the game better

1. conserve momentum ffs, especially while jump jetting.
2. recoil on all weapons that should have recoil (in cases where this is also a nerf, appropriate buffage may be neccisary)
3. non-instantaneous convergence (i know there are technical reasons why we dont have this, but rotating a couple 5 ton laser systems or a 15 ton gauss rifle should not be instant).

Edited by LordNothing, 23 January 2020 - 01:34 PM.

[Gristle Missile]

Game already breaks so many laws of physics and reasoning

Seen some information that the PSI required from the legs to move a 100t mech up to speed would cause it to crush through the ground and get them stuck... and a mechs surface area would prevent the speeds you see in battletech due to wind resistance

[Nightbird]

Gristle Missile, on 23 January 2020 - 01:49 PM, said:

Game already breaks so many laws of physics and reasoning

Seen some information that the PSI required from the legs to move a 100t mech up to speed would cause it to crush through the ground and get them stuck... and a mechs surface area would prevent the speeds you see in battletech due to wind resistance

Link? We have tanks that are 100 tons and go 65kph (only 40 mph), and even if you shrink the tracks by 75% to the size of feet, it would damage concrete but wouldn't go "through" it or the ground for that matter. The mechs also don't use pneumatic actuators (where PSI comes from) to move but rather they use electroactive muscle tissue. This already exists IRL but obviously not to the degree needed for mechs https://en.wikipedia...active_polymers

Keep in mind dinosaurs existed in the past and mechs are made of stronger materials than flesh and bones.

The Argentinosaurus is 100 tons on 4 legs, so 50 tons on two legs.

Edited by Nightbird, 23 January 2020 - 02:03 PM.

[MisterSomaru]

Nightbird, on 23 January 2020 - 12:22 PM, said:

So, why should we care about physics and mech agility?

1. It might be a fun discussion, if you don't love mechs nor enjoy discussing them then you should probably stop reading here.
2. There is some common sense involved. For example when we see a cheetah next to a giraffe, we automatically can feel which is more agile. If this expectation is proven wrong, as it often does in MWO, we instinctively feel a wrongness.
3. Would gameplay balance and mech "feel" be improved if physics were considered? Read to the end and decide for yourself


I wanted to keep this discussion simple and testable, so instead of math, we'll use some tests that can be done at the gym and illustrated with stick figures. To do the tests by yourself, take some dumbbells around 30% of your weight total (if you weigh 200lbs, 2x35lb dumbbells, if you're 150lbs, 2x25 lb dumbbells, etc), but be safe, don't overdo it, and don't hit anyone around you.

The dumbbells will represent either the center of mass for the weapons equipped (high or low) on a mech for humanoid shaped mechs or the head and rear for long marauder shaped mechs.

Mech Pitch Angle



For A, B, C don't push your rear back to counter balance. Try to control effort, working harder just for one move defeats the purpose of the test.

A: To represent low slung mechs like the Atlas, Cataphract, Highlander put weights at waist and lean forward. Use a mirror if needed. You'll find your balance to be excellent and it is easy to pitch forward 90 degrees or more

B: To represent shoulder mounted mechs like the Blood Asp, Corsair, Battlemaster, put weights over the shoulders by your head and lean forward, you'll find that your balance to be precarious and your pitch greatly limited.

C: To represent chest weapon mechs like the Warhammer, Victor, Cyclops, hold weights at chest height, it feels about halfway between A and B

D: To represent a long mech like a Marauder (IIC), Mad Cat II, Raven, hold with arms fully extended. Move arms up and down without tilting torso to represent the pitch. This rest is common sense, as by having a counter-balance, balance is not really affected by pitch angle with the only limit being physical obstruction.

E: There is no E depicted, which is half way between D and C to represent squat mechs like the Timberwolf, Direwolf, Champion, but pitch should still be good due to low center of gravity and some counter-balancing.

D>A=E>B>C

How it compares to MWO today:
Mechs with low slung weapons are usually not rewarded with better pitch range. For example, the Atlas and Firestarter have some of the worst pitch ranges in game. Mechs with high mounted weapons are usually not punished with less pitch range. For example, the Blood Asp and Corsair and Blackjack. (Notably, the KDK-3 is the only one with the right pitch penalty) Squat and long counter-balanced mechs are also not rewarded with better pitch, for example the Marauder, II, IIC. The Nightstar seems to be one mech that is long and have improved pitch. If for technical reasons pitch is limited to 15-35 degrees, then it could be distributed among mechs in a way that makes more sense.



Mech Pitch Speed

Using the same diagrams.

For A, B, C above, having a better pitch range due to lower center of gravity also carries over to pitch speed. It's fast to bow with A and delicate and slow with B. Pitch down is faster than pitch up since down is gravity assisted.

For D, E there is no gravity assistance due to the counter-balance, so pitch up and down is the same speed. Also, by having mass further away from the center, it takes a little more leverage to pitch compared to humanoid mechs.

A=E>C=D>B


How it compares to MWO today:
Pitch speed seems to be rolled into one with torso yaw speed, see below.


Mech Yaw Angle

There isn't the same physical balance considerations as pitch here, so PGI is free to assign anything they see fit. Though, for a declared walking tank game, I am against mechs that twist less than 90 degrees for obvious reasons.

Mech Yaw Speed



Try to control effort, working harder just for one move defeats the purpose of the test.

For A, B, C twist left and right, and the first thing you notice is that the speed is about the same. This makes sense as the axis of rotation is vertical and the center of mass stays close to the axis of rotation even as it moves up or down.

For D, and implied E (squat mechs), the mass is further away from the axis of rotation and therefore it takes a lot more leverage to twist, making the torso twist movement slower.

A=B=C>E>D

How it compares to MWO today:
Yaw speed and pitch speed being the same is a problem, since the order from best to worst is VERY DIFFERENT. Looking at just Yaw, sometimes the relative benefit is not right, for example the KDK should twist better than the Direwolf, it doesn't. Sometimes, the scale of the difference is not sufficient. Ingame the Atlas twists 15% faster than the Marauder II, but from testing in the gym, maybe 50% to 100% faster would be feel more 'right' and this should be applied to all humanoid versus squat mech (twist 15-30% slower) versus long mech (twist 33-50% slower).

This makes sense not only because of physics in the gym, but also from a mech design perspective. Humanoid mechs are designed to twist to avoid damage, as they shoot from the front but have the slimmest profile from the side to mitigate damage, so they should be turning 90 degrees often. For long mechs, the slimmest profile is already the front, and mitigating damage is done by wiggling the nose, so a full 90 degree turn is rarely needed.


Mech Accel, Decel, Turn



Test both running from a stop, seeing how long it takes to reach a comfortable speed, and running in a circle and seeing how small a circle you can make. Try to control effort, working harder just for one move defeats the purpose of the test.

I think most can guess how this turns out, the higher the center of mass, the slower accel, decel and turn.

A>C>B

D is actually hard to test, since the center of mass representation is different - around our chest - but it is comparable to where the center of mass is in A. Squat mechs (E) also center around the same place as for C.

A=D>C=E>B

The difference between A and B is very significant, with B being half as fast at accelerating and turning as A.

How it compares to MWO today:
Here are some of the mechs with wrong relative accel/decel/turn speed. The Direwolf has a lower center of gravity than the Kodiak yet it is slower. The Blood Asp has a higher center of gravity than the Mad Cat II yet has the same agility. The Mauler and Cyclops have similar centers of gravity yet the Cyclops is almost twice as agile. The grasshopper has a higher center of gravity than the Cataphrat yet is also 50% more agile than it. I will add that for many mechs, the difference is in the right direction, but the scale of the difference is far less than what is experienced in test. (magnitude feels wrong)



So after such a long post, am I asking PGI to implement it? No, we all know PGI isn't updating MWO anymore. I don't even think we should only follow RL physics, just that it is a good starting point and tweaks can follow. Of note is that high weapon mounts have natural penalties due to the high center of mass, something PGI rarely uses to balance mechs. There is also the difference between humanoid and plane shaped mechs that don't really get fleshed out ingame.

I do hope for some of these comparisons, you felt "Oh! That's why that mech feels off" and that is the case with many mechs in MWO and the point of this thread.


Thanks for reading.

A lot of those mechs with low slung weapons or body shapes tend to be underperformers as well. better agility would go a long way to making them a bit more competitive in game as actual viable alternatives.
I'm not saying we should nerf the weapons with more high mounts, but rather give buffs to the mechs with primarily low mounts to compensate.

[Nightbird]

MrSomaru, on 23 January 2020 - 02:08 PM, said:

A lot of those mechs with low slung weapons or body shapes tend to be underperformers as well. better agility would go a long way to making them a bit more competitive in game as actual viable alternatives.
I'm not saying we should nerf the weapons with more high mounts, but rather give buffs to the mechs with primarily low mounts to compensate.

I don't think anything will be buffed or nerfed, MWO dead after all. I'm just noting that if mechs went through a physics common sense evaluation before initial release, mechs would feel more natural and also be better balanced in terms of weapon mount location to agility relationship.

But damn... quoting that entire first post lol

Edited by Nightbird, 23 January 2020 - 02:12 PM.

[Red Potato Standing By]

Just the physics of them stopping suddenly or quickly would probably cause them to fall down. Specially the high mounted ones

[Nightbird]

Red Potato Standing By, on 23 January 2020 - 02:20 PM, said:

Just the physics of them stopping suddenly or quickly would probably cause them to fall down. Specially the high mounted ones

That's in the accel and decel tests in the OP, you can't stop sharply at all, but have to do it gradually. This is something not implemented ingame for top heavy mechs.

[DaZur]

I would love this kind of detail... Quite honestly bordering on simulation level would make me giddy.

That said, the vast majority of the player base that is interested in MWO would not have the patience to learn these intricacies and would drop it like a hot potato.

This is why high fidelity flight simulations like Falcon 3.0 and the DCS series is such a backwater niche now days...

It's kinda sad if you ask me...

[Red Potato Standing By]

Nightbird, on 23 January 2020 - 02:24 PM, said:

That's in the accel and decel tests in the OP, you can't stop sharply at all, but have to do it gradually. This is something not implemented ingame for top heavy mechs.

Well it seems I read about as well as I play...lol

[YueFei]

Agreed with all of this. For even more fun and to create opportunities for higher skill ceiling with mech maneuvering, dynamically-adjusting accel/decel and asymmetric turn rates based on your twist/lean angle would also be awesome!

Want to accelerate forward a bit faster? Lean forward first before pressing 'W'.

Want to turn to the right faster? Twist to the right and lean forward.

etc...

[LordNothing]

id love to see living legends style jumpjet control. angle of the jets is controlled by torso pitch. nose up go backwards, nose downe go forward, level to hover or go up (depends on jj config, but there would be a ground effect phase where you get a bit more up and then lose lift). we already have the yaw control through the leg controls. kind of like a helicopter. living legends also had roll control but it was kind of hard to use, i think we would do better without it though. when i fly my rc heli (from before all that drone crap) a little nose down means a lot of forward acceleration. it would go a long way towards the kind of jj control you need to get the long horizontal jumps described by lore, without the scourge of poptarting.

Edited by LordNothing, 23 January 2020 - 09:13 PM.

[PhoenixFire55]

All I'm gonna say is that mechs have gyros for a reason.

[General Solo]

More Gyro stuff

[Dimento Graven]

LordNothing, on 23 January 2020 - 01:30 PM, said:

...

3. non-instantaneous convergence (i know there are technical reasons why we dont have this, but rotating a couple 5 ton laser systems or a 15 ton gauss rifle should not be instant).

In my observations it's not actually "instantaneous", it's just very, VERY fast.

But opinions vary I guess...

As to the OP - I think someone will have mentioned it by now, giant multi tonned gyros are what help keep these giant beasts upright, that and (according to lore) the pilot's sense of balance imparted through the nuerohelmet.

In fact, it's intimated in the fiction at least, that the better the pilot's on sense of balance, the more the potential for 'mech nimbleness.

That said, I believe that most assaults need about 20% more pitch and yaw range and speed. The engine desync years past was one of the larger and more continuous moments of "disruption of suspension of disbelief" that occurs in game.

The other is the incredibly physics breaking motion of the smaller, faster lights, making incredible 180 degree (or even more severe) turns at top speed without turning the pilots into puddles of goo.

Certain lights are overly nimble, and most assaults/heavies are not nimble enough.

Edited by Dimento Graven, 24 January 2020 - 06:09 AM.

[_Magno_]

LordNothing, on 23 January 2020 - 01:30 PM, said:

there are some areas in physics i think would make the game better

1. conserve momentum ffs, especially while jump jetting.
2. recoil on all weapons that should have recoil (in cases where this is also a nerf, appropriate buffage may be neccisary)
3. non-instantaneous convergence (i know there are technical reasons why we dont have this, but rotating a couple 5 ton laser systems or a 15 ton gauss rifle should not be instant).

Most mechs, especially the heavier you get, have a brake pad after you stop throttle. So momentum is definitely conserved.
While jump jetting, you get stuck in the original vector launch path, so momentum again conserved.

What momentum are you talking about? Besides never falling down.


Recoil physics are definitely tame.
Angular momentum physics are more pronounced on MW5.

[Nightbird]

PhoenixFire55, on 23 January 2020 - 11:57 PM, said:

All I'm gonna say is that mechs have gyros for a reason.

OZHomerOZ, on 24 January 2020 - 12:45 AM, said:

More Gyro stuff

Yes and No.

Yes, Gyros are cool. No, Gyros cannot be used to maintain a greater pitch.

Gyros can momentarily be used to prevent a fall by gripping a fast spinning flywheel and feeding off it to give force in the opposite direction to the fall. That moment can be used for the mech's other systems to reorient the mech to a position of balance. However, to maintain a pitch, you will have to be constantly offsetting gravity. After 1 to 3 seconds, the flywheel will stop spinning and you'll fall.

Edited by Nightbird, 24 January 2020 - 07:21 AM.

[Dimento Graven]

The theory of the gyros in 'mechs is that they are kept spinning, somehow, by being fed power from the fusion engine in the 'mech.

Perhaps this is yet another instance of 'Stackpoling' but, it's good enough for me and maintaining my 'suspension of disbelief'.

[Nightbird]

Dimento Graven, on 24 January 2020 - 07:20 AM, said:

The theory of the gyros in 'mechs is that they are kept spinning, somehow, by being fed power from the fusion engine in the 'mech.

Perhaps this is yet another instance of 'Stackpoling' but, it's good enough for me and maintaining my 'suspension of disbelief'.

Well no... gyros are always kept in a close to zero friction environment, it'll keep spinning without needing additional energy. The issue is that to use the gyro to stabilize the mech, you have to 'grip' it. Gripping it will very quickly reduce the spin to 0, and that spin energy is transferred as a torque force to the outside of the gyro (onto the mech).

If nothing else, just understand that gyros only have a few seconds of use at most. After the mech regains balance, the gyro can be spun up again to store more energy. Once the spin is restored, it can be used again. While spinning the gyro up, it will apply a force on the mech opposite to gripping it, so it can only be done when the mech is upright and balanced.

Edited by Nightbird, 24 January 2020 - 07:29 AM.

[Dimento Graven]

Nightbird, on 24 January 2020 - 07:27 AM, said:

Well no... gyros are always kept in a close to zero friction environment, it'll keep spinning without needing additional energy. The issue is that to use the gyro to stabilize the mech, you have to 'grip' it. Gripping it will very quickly reduce the spin to 0, and that spin energy is transferred a torque force to the outside of the gyro (onto the mech).

If nothing else, just understand that gyros only have a few seconds of use at most. After the mech regains balance, the gyro can be spun up again to store more energy. After those few second, the gyro runs dry.

Right, but energy has to be applied to spin the gyros up initially, and there has to be some methodology of keeping them spinning once spun up, because after all, 'close to zero friction' is not 'zero friction', so obviously they have to be fed energy once they slow down enough.

So again, according to BattleTech lore, gyros are more or less constantly being fed energy to keep them 'spun up' for usage.

It's why gryo hits were tracked, and part of the reason why getting too hot would affect movement, and why piloting skill rolls would get hit with modifiers and such.

Again, maybe not in our 'current' reality and implementation of gyros is this possible, but hell... 30 years ago when BattleTech was being dreamed up, a lot of the stuff the lore is based on wasn't 'real', yet now, now we have our militaries equipping offensive laser and magnetic rail guns on planes and ships, we have electro responsive materials that are not too far off from being actual 'myomer cables', and the like.

Was commenting on the stream how back in the 30's the "two-way radio wrist watch" of the **** Tracy comic strip was a total "********" unrealistic SCI-FI concept, and now we have smart watches that do THAT VERY THING, and an F-ton more.

In this particular case, the lore allows for it and it's not too stupidly reality breaking for me to accept.