53 replies to this topic

[0rionsbane]

its a nice layout, mostly weighted in favor of offensive power.
a few ones i disagree with because though the numbers are there overall the mech itself is missing something to make them line up. ex the on1-VA its got nice quirks and its hitboxes are better than ever. but whats missing is how much heat that mech produces with the meta load out and how it has no - heat quirks so it cant preform as well as it would on paper.
or the dragon , specifically its lesser quirked brothers, are rated quite low but are actually more flexible than the over quirked ac 5 dragon. a dragon with guass and 2 ll's is as effective, and never did need the quirks.
so basically because it would take a while to add in more variables like heat scale, effective weapon ranges, durability (especially for easily ct'd or legged mechs) its never going to be accurate. its just a reference point on mostly offensive capability disregarding heat, and defense disregarding speed, and assuming you dont allow the opponent to hit your massive ct or weapon arm on some of these mechs.

[Duke Nedo]

0rionsbane, on 14 April 2015 - 04:48 PM, said:

its a nice layout, mostly weighted in favor of offensive power.
a few ones i disagree with because though the numbers are there overall the mech itself is missing something to make them line up. ex the on1-VA its got nice quirks and its hitboxes are better than ever. but whats missing is how much heat that mech produces with the meta load out and how it has no - heat quirks so it cant preform as well as it would on paper.
or the dragon , specifically its lesser quirked brothers, are rated quite low but are actually more flexible than the over quirked ac 5 dragon. a dragon with guass and 2 ll's is as effective, and never did need the quirks.
so basically because it would take a while to add in more variables like heat scale, effective weapon ranges, durability (especially for easily ct'd or legged mechs) its never going to be accurate. its just a reference point on mostly offensive capability disregarding heat, and defense disregarding speed, and assuming you dont allow the opponent to hit your massive ct or weapon arm on some of these mechs.

Thanks!

Aye, the heat handling is a difficult one, especially for IS where you can adjust heat by reducing engine size or going xl in addition to reducing the number of hot guns. For clams it is to some extent taken into account in the Hardpoint capacity category since that also includes available podspace. I tried to consider this to some extent when giving scores to the lighter mechs by rewarding a higher score to mechs with many E hardpoints (allowing boating of small lazors), and a low score for having a single M, or 1-3 B (hard to use efficiently). In general though I don't think there is a way to evaluate the heat of various potential builds and implement it in a consistent way, have to draw the line somewhere...

There are a few variants here that I also disagree with, but I think that is a good thing. That means that it's not 100% my opinion here, but that there is a "model" behind it based on independently rated characteristics that sorts the mechs into Tiers that in my opinion looks intuitively rather good. More good than bad atleast!

Edited by Duke Nedo, 14 April 2015 - 10:44 PM.

[Vandul]

FupDup, on 13 April 2015 - 12:35 PM, said:

His graph is counter-intuitive in the way it represents effectiveness. In this graph, the lower bar is actually better because it's closer to Tier 1. The higher bars are worse tiers like T4 or T5. It's kinda wonky.

Ah, that makes more sense now. I was like, when did the Awesome become so Awesome?

[That Dawg]

I just went out and bought AWS-8V, DRG-5N, VIN-1B, LCT1V based solely on this science.
Will my ELO skyrocket now?

[Duke Nedo]

If you read the graphs as high==good you're in for a surprise...

[That Dawg]

Duke Nedo, on 17 April 2015 - 03:26 AM, said:

If you read the graphs as high==good you're in for a surprise...

wait...WHAT??? WHO makes graphs like that unless they are trying to sell CRAP to a trusting and unassuming audience??!?!!?

[Duke Nedo]

Don't blame me, it's PGI started this whole Tier 1 mech business...

[kapusta11]

My two cents regarding tier calculation algorithm:

Off. meta score should be based on the amount of damage a mech can deal to a single component when "trading" damage. I'd call it Effective Damage.

Def. meta score should be based on the ability to avoid receiving damage to single component (CT or ST if XL). Key variables are (the numbers after them are weight coefficients which can be tweaked, they are just rough estimation)

• CT armor + structure or ST armor + structure when carrying IS XL engine, hitting anything else is basically a miss as it won’t kill the opponent.
  
• Acceleration/deceleration/torso twist rating (since they tied to engine rating and thus movement speed I'll rate it around that), 1 - below 60 kph, 1.1 - 60-65 kph, 1.2 - 65-70kph, 1.3 - 70-80 kph, 1.4 - 80-90 kph, 1.5 - 90-100 kph, 1.6 - 100-120 kph, 1.8 - 120-140 kph, 2 - 140 kph+
  
• Mech's hitboxes size (based on square surface alone). 1 - huge, easy to hit hitboxes (Atlas, Awesome, Battlemaster, Banshee), 1.1 slightly smaller hitboxes (smaller assaults like Highlander, Zeus and oversized heavies like Orion or Catapult), 1.2 - most heavies and Stalker, 1.3 small heavies (Jagermech with his narrow hitboxes and oversized mediums), 1.4 mediums with small or narrow hitboxes (Shadowhawk), 1.5 small mediums like Cicada, Stormcrow (because of really narrow hitboxes) and oversized lights (only jenner comes to mind but for a good reason - huge CT, Commando has pretty big rear CT too), 1.6 - lights with really small hitboxes like Firestarter, Spider etc.
  
• JJs (help spread damage, but only against lasers which are the meta on the other hand). 1 - no JJs, 1.33 - with them.
  
• Hitbox Layout (determines how well a mech can protect CT/ST when torso twisting). 1 - nothing special, 1.33 – huge shield arms or shield side.
  
• Hardpoint Layout (high mounted hardpoints is more of a deffensive parameter rather than offensive, as it allows you to expose yourself less and thus prevents you from taking damage). 1 - requires to expose yourself completely to do full damage (normal Victor variants and their widely placed arms for example), 1.33 only upper/left/right half. Might not be included in HP calculation for fast lights.
  
• Engagement range (depends on weapon's range) 1 - facehug range, 1.1 – 200-300m range, 1.2 - 300-600m, 1.3 600m+.

Armor is the base value and the variables mentioned above multiply mech's defensive capabilities rather than beigh additive with each other. This one I'd call Effective HP (or armor).

Mech Tier = Effective damage x Effective HP. Higher means better, final results should be inverted and brought to 1-5 scale, but I have no time to calculate this stuff. The model is rough but should be pretty accurate inside of particular weight class. Another formula might be: Mech Tier = Average Effective HP (across all mechs) / Effective damage x Effective HP / Average Effective damage, basically how fast you kill multiplied by how fast you die.

Main source of errors:

• Wrong estimation on how particular variable affects Effective HP (wrong weight coefficients and their linearity)
  
• Possible unaccounted correlation between said coefficients like laser burn duration and mech’s speed or hardpoint layout and torso twist speed (one is useless without the other and thus should probably be treated as a whole)

Notes:

Spoiler

That’s just my opinion but ECM is overrated and is easily countered by UAV and NARC …and PPCs had they not been as bad as they are now and someone actually used them. And against LRMs cover > ECM. I.d give it no more than 10-15%, thus 1.1-1.15 modifier

Edited by kapusta11, 17 April 2015 - 09:01 AM.

[Duke Nedo]

kapusta11, on 17 April 2015 - 07:32 AM, said:

My two cents regarding tier calculation algorithm:

Off. meta score should be based on the amount of damage a mech can deal to a single component when "trading" damage. I'd call it Effective Damage.

Def. meta score should be based on the ability to avoid receiving damage to single component (CT or ST if XL). Key variables are (the numbers after them are weight coefficients which can be tweaked, they are just rough estimation)

• CT armor + structure or ST armor + structure when carrying IS XL engine, hitting anything else is basically a miss as it won’t kill the opponent.
  
• Acceleration/deceleration/torso twist rating (since they tied to engine rating and thus movement speed I'll rate it around that), 1 - below 60 kph, 1.1 - 60-65 kph, 1.2 - 65-70kph, 1.3 - 70-80 kph, 1.4 - 80-90 kph, 1.5 - 90-100 kph, 1.6 - 100-120 kph, 1.8 - 120-140 kph, 2 - 140 kph+
  
• Mech's hitboxes size (based on square surface alone). 1 - huge, easy to hit hitboxes (Atlas, Awesome, Battlemaster, Banshee), 1.1 slightly smaller hitboxes (smaller assaults like Highlander, Zeus and oversized heavies like Orion or Catapult), 1.2 - most heavies and Stalker, 1.3 small heavies (Jagermech with his narrow hitboxes and oversized mediums), 1.4 mediums with small or narrow hitboxes (Shadowhawk), 1.5 small mediums like Cicada, Stormcrow (because of really narrow hitboxes) and oversized lights (only jenner comes to mind but for a good reason - huge CT, Commando has pretty big rear CT too), 1.6 - lights with really small hitboxes like Firestarter, Spider etc.
  
• JJs (help spread damage, but only against lasers which are the meta on the other hand). 1 - no JJs, 1.33 - with them.
  
• Hitbox Layout (determines how well a mech can protect CT/ST when torso twisting). 1 - nothing special, 1.33 – huge shield arms or shield side.
  
• Hardpoint Layout (high mounted hardpoints is more of a deffensive parameter rather than offensive, as it allows you to expose yourself less and thus prevents you from taking damage). 1 - requires to expose yourself completely to do full damage (normal Victor variants and their widely placed arms for example), 1.33 only upper/left/right half. Might not be included in HP calculation for fast lights.
  
• Engagement range (depends on weapon's range) 1 - facehug range, 1.1 – 200-300m range, 1.2 - 300-600m, 1.3 600m+.

Armor is the base value and the variables mentioned above multiply mech's defensive capabilities rather than beigh additive with each other. This one I'd call Effective HP (or armor).


Mech Tier = Effective damage x Effective HP. Higher means better, final results should be inverted and brought to 1-5 scale, but I have no time to calculate this stuff. The model is rough but should be pretty accurate inside of particular weight class. Another formula might be: Mech Tier = Average Effective HP (across all mechs) / Effective damage x Effective HP / Average Effective damage, basically how fast you kill multiplied by how fast you die.

Main source of errors:

• Wrong estimation on how particular variable affects Effective HP (wrong weight coefficients and their linearity)
  
• Possible unaccounted correlation between said coefficients like laser burn duration and mech’s speed or hardpoint layout and torso twist speed (one is useless without the other and thus should probably be treated as a whole)

Notes:

Spoiler

That’s just my opinion but ECM is overrated and is easily countered by UAV and NARC …and PPCs had they not been as bad as they are now and someone actually used them. And against LRMs cover > ECM. I.d give it no more than 10-15%, thus 1.1-1.15 modifier

Thanks for the input!

I don't disagree with anything you wrote, it's just on a totally different level of detail than I attempted and then it starts to get really difficult to implement without creating ambiguities. Like you touched upon there, hitboxes for example can be good^TM for various reasons. While XL-friendly hitboxes increase the survivability for some mechs, the oppsite is true for others.... for example Stalkers who have great durability because it's low engine cap makes it always run a STD engine and is extremely XL-unfriendly, making the CT very slim... some are just over all small, or fast, and some have very well balanced hitboxes allowing them to instead spread damage efficiently.

That's why I went with more bundled stats just giving a subjective score. Mainly I believe that hardpoints and hitboxes are the important ones, I broke down hardpoints into capacity and locality, but refrained from breaking down hitboxes further.

The benefit is that by breaking it down a little bit into subparameters one can be a little more systematic and still make judgement calls and take all the things you mention above into account when setting the score. The drawback is that there is room for bias. That's why I made an attempt to get some experienced guys give their scores in this format but so far I am not being run down with requests for a form to fill....

[Telmasa]

Just a passing idea - for the graph, why not make the bars top-down rather than the orthodox bottom-up?

i.e. flip green to red http://i.stack.imgur.com/uERc9.png

vertical side, 0 through 5 tier going from top to bottom, while preserving your left-to-right horizontal side

That'd probably make it a bit more noob-friendly. At the same time you could revert back to stronger mechs first, weakest ones last.

Edited by Telmasa, 17 April 2015 - 03:10 PM.

[Duke Nedo]

Perhaps this could be a basis for discussing the upcoming PGI BV balance pass? (necro justification)

As I understood it PGI are doing something along these lines, but more advanced and probably using game metrics as input... so they have a good chance to get better quality of input, but what I am curious if they intend to attempt this without any subjective input at all?

As I remember this, it would be really hard to catch a some things by objective numbers that can be extracted in a simple way, like:

- Heat efficiency. This will depend on lots of factors. XL engine vs geometry, Hardpoint distribution vs viable builds, # truedubs etc. Not straightforward to say if a mech is heat efficient or not without deciding which build you are going for...

- Hardpoints. The number and placement of hardpoints is one thing... but how viable are available asymmetric builds actually? The type of hardpoints vs the weight of the mech. How many of the hardpoints are placed high, does that reach the critical mass vs chassi? For example, for a heavy it can be enough with 2x B, for a light that would be useless, while it would need a critical mass of 3-4E clan mech and 5-6E for an IS mech....

I am sure there are plenty plenty more of those.

Any speculations?



PS. now that we have wave 3 we can see that they performed pretty close to what was predicted, no? Exe may be off because MASC was completely unknown but anyways.

[Lily from animove]

Quote

- Heat efficiency. This will depend on lots of factors. XL engine vs geometry, Hardpoint distribution vs viable builds, # truedubs etc. Not straightforward to say if a mech is heat efficient or not without deciding which build you are going for...

heat efficieny, how do you calculate this, i saw already tons of people speaing about heat efficiency, not having any idea what it truly means.

actually, heat efficiency would mean, how much damage a mech can generate out of its heat. Also how muh damage can be sustained. heatgeneration)

Also, this should not take geometry and stuff into account, geometry and hardpoint locations shuld be under an own judged topic: time to exposure. Time a mech needs to expose the weapons a builds want to fixe and to hide. Since this defines the opponents ability to retailate.
because this is where hardpoint location, geometry accel/decc and speed come together to a game relevant value.
This si very important form scientific aspect, because soem emchs cna expose, fire and retreat as target below a humans average reaction time, So when someone relaises this mech, twists to it, and fires, he may nto even have a chance anymore to hit this target. That is a very important fact form a scientifc point of view.

Edited by Lily from animove, 14 August 2015 - 05:02 AM.

[Duke Nedo]

Lily from animove, on 14 August 2015 - 04:58 AM, said:

heat efficieny, how do you calculate this, i saw already tons of people speaing about heat efficiency, not having any idea what it truly means.

actually, heat efficiency would mean, how much damage a mech can generate out of its heat. Also how muh damage can be sustained. heatgeneration)

Also, this should not take geometry and stuff into account, geometry and hardpoint locations shuld be under an own judged topic: time to exposure. Time a mech needs to expose the weapons a builds want to fixe and to hide. Since this defines the opponents ability to retailate.
because this is where hardpoint location, geometry accel/decc and speed come together to a game relevant value.
This si very important form scientific aspect, because soem emchs cna expose, fire and retreat as target below a humans average reaction time, So when someone relaises this mech, twists to it, and fires, he may nto even have a chance anymore to hit this target. That is a very important fact form a scientifc point of view.

Yeah, defiantly very complex. Some things like geometry, hardpoints and hitboxes will affect several performance metrics...

To make things worse, different 'rules' to calculate bv will have to apply differently to different chassi, even to different variants in some cases.

Some examples I can think of:

Is 2 high E hardpoints good on a light? Yes for a Raven. No for a mist lynx. The lynx have little podspace so it can't run 2 big guns. The Raven can... And the Raven variants look different for hardpoints but if they only use those 2 E, then the 4x and 2x are identical. How can the game know?

Is a slim ct hitbox good? Yes for the stalker, not so much for the awesome.

Add to that also locked equipment, the exe is a good example. Does it have good hardpoints? Can't make up my mind...

Then we have critical mass of podspace... The summoner would have had great hardpoints for dual gauss without locked jj, but now it can't use them like that and have to rely on low slung arms...

Etc etc

[Duke Nedo]

Quickly just removed the quirks in that model to get the "unquirked" tier list based on the input for hardpoints, geometry and stuff that I'd assume would be part of PGIs upcoming internal BV. Just for kicks, will be fun to see how this compares to PGIs final BV. This one uses lots of subjective input for good or bad...