16 replies to this topic

[Kin3ticX]

This is my third attempt to post material on this. Making a skill tree guide turns into a cluster you know what really fast because of all the different builds and screenshots required. It's no easy task. I feel it's more refined now and I am definitely not skilling 'mechs the same as I was when the skill tree first came out.

The skilltree does sort of have a clear meta, but it's also very subjective at the same time. It makes me sort of dread trying to make a guide for it. Disagreements between doing it this way or that way probably won't make or break your matches, but maybe it can, who knows. Nobody agrees on everything due to it's often subjective nature.

My own rules

1. Firepower, Survival, Operations, and Aux get first dibs over mobility, sensors, and jump jets.
-never use the jump jet tree, not even for a Shadowcat or a Viper

2. Only use mobility and speed tweak on lights and fast mediums. Maybe you might have a rule that you don't want to go under 60kph or something... that's fine.... but most of your mediums, heavys, and assaults will run fine without speed tweak.

3. I tend to go light on consumables, (1x strike, 1x coolshot), so the consumable/aux tree is more a grey area. You will want to think about getting both adv salvo skills even if its just 1 strike. If you like to burn 4+ consumables every drop, by all means go ham on these skill pts.

4. For ECM mechs I get both ECM skills. I generally try to play non-ecm mechs without radar deprivation but sometimes I get a couple token points of radar derp.

5. I tend to use laser duration skills on clan mechs and less often on IS mechs. A lower burntime is superior but if your mech is already hot, maybe its not needed as it also slighty lowers your cooldown rate.

6. Never bother with flamer skills or high explosive skills, both of these I never ever use. (except if using high explosives to take shortcuts as junk points)


If anyone has a question or a tree request, feel free to ask. If you reply, I get an email, so I might catch it .

Edited by Kin3ticX, 22 February 2018 - 12:24 AM.

[Kin3ticX]

Common Cfgs and Pathing

Firepower Tree


This is one of the ways to get max heat gen (14/14) without taking too many shortcuts. Notice on the left I get missile spread as a throwaway point but it saves me 1 skill point for something else in another tree.



Trimming via junk pts
Its possible to sneak a few points back by trimming if you really need the points elsewhere (while still getting 14/14 heat gen)




This one is another setup which gets 11/14 heat gen. I may not use it exactly like this but sometimes I use part of the pathing to save points.




This is another tree I use which is light on investment with max burntime, especially for my shadowcat to free up points for ECM and Mobility, 7/14 heat gen





Common Cfgs for OPS tree

5/5 coolrun

You can also just as easily go down the other side to get the last 2 coolruns.




4/5 coolrun (this one is best bang for buck when you dont have enough points to have everything)




3/5 coolrun: I use this one far LESS often because for 2 more points you can get 4/5 coolrun




Mobility Tree


Mobility tree is simple, it's basically this or nothing. It's only really worth it on mechs 97kph or faster imo, but you could also just as easily only use it on mechs 120kph or faster. These are wasted points on anything slower unless you have some rule that you refuse to go slower than 60kph or something for solo queue purposes. It probably wont save you from a endless nascar in the slowest of slow mechs but it could help your 'mech feel a little less like driving a bus.

You can also go down the other side instead to pick up hardbrake and burst instead of yaw.





Survival Tree

This is what a maxed out surival tree looks like. If you go with a partial survival tree, it should have some left side bias because that gets more armor over structure.


Remember, a partial surival tree should have a left sided bias to grab armor over structure.

Edited by Kin3ticX, 22 February 2018 - 12:01 AM.

[Kin3ticX]

Mech examples

I go pretty light on consumables. If you want to spam Strikes, Coolshots, and UAVs, by all means go ham on it.

WVR-6K w/ 3xLL, 2xERML




Shadowcat w/ 2xHLL, 1xERML and full ECM




Banshee 3M w/ 3xLPL 5xERML




Arctic Cheetah with 3xHML, 6xLMG




Trebuchet 7k w/ 2x HPPC - i picked up 4 velocity skills on the weapons tree. The 5th velocity is a big investment so I skipped it. I also scrounged up some extra cooldown skills.




Mauler MX90 w/ 6xUAC2. I seem to have broken my rule here since when I tree'd this i didnt want to go too much under 60KPH. You could also just not use mobility, thats fine too.



HBR 2xHLL, 4xERML w/ full ECM. Going full ECM makes the points really have to stretch, especially since I wanted nearly max survival. This is why I did some trimming on the Firepower tree. I also wanted the burntime skills but I went with 3/4 because the 4th was too much for the points I had.




Mauler 2xHGauss 4xERML: Again, breaking my rule with agility but only to not move so far under 60KPH for solo queue purposes. You could probably easily put those points elsewhere if you want. I think I put the 1 pt into gauss charge because pgi did something funny with the HGauss chargeup/chargedown. Its optional. After all that, hunt down some more cooldown modules I guess.





Kintaro-20 with 2xMRM-20, 4xML: Here's an MRM example with a mech that has a 5% missile velocity quirk and where I soak up all the spread, velocity, and rack skills I could. I am not claiming this is anything special, it's just an example for MRM based skills.





Warhammer-6R w/ 2xUAC10, 4xERML

Edited by Kin3ticX, 25 February 2018 - 12:54 AM.

[Deathlike]

I'll just say there are better ones out there.

A way of looking at it for me is that if it is worth spending 1 node to get 2 or 3 followup "good nodes", it's worth it.

Here's an example:
https://kitlaan.gitl...mor%20Structure

Take note of what is actually gated and how many "beneficial nodes" are available for structure/armor. Many side nodes are gated poorly and thus are crappy investments.

For agility... while I don't use it generally, but you want to path in such a way...
https://kitlaan.gitl...b6ac8#s=Agility
You want accel+decel nodes as that will allow you to peek better. You do have an option to go for anchor turn over torso twist speed on the lower corners pathing into speed tweak (especially if you're a Light mech - not as useful as a slower mech). Torso pitch is way overrated (useful in theory for mechs like the Atlas/Firestarter, but too expensive/not as useful than other forms of agility).

Firepower is a beast, but heavily dependent on the build itself. There are too many terrible sidegrade nodes (especially stuff on the edges for heat gen) that cost too much. You should be spending almost 30-40 points depending on the build to get them all - assuming you are using something stupidly hot.

Sorry for trying to derail this useful thread though.

[InspectorG]

I like Mobility for slow Assaults in order to get a 90 degree(or as close to) twist for side-pokes. Anni and Dire has to expose full frontal nudity to fire anyhow, this might help get back into cover before the sun sets.

[Deathlike]

Here's a few ideas for common-meta-ish builds...

Pure laservomit:
https://kitlaan.gitl...b6c51#s=Weapons

While I could path to get the leftmost edge's heat gen, you spend at least 2 nodes to do that. Pathing thru velocity isn't useful technically, but 33 nodes consumed isn't too shabby IMO.

PPC heavy builds:
https://kitlaan.gitl...24e0e#s=Weapons

At a nice low of 28 nodes, the last velocity node is gated poorly, so you are almost advised to do something else with your time/life. Note that velocity nodes gate pretty well for optimal trees.

Laservomit+Gauss:
https://kitlaan.gitl...9bc08#s=Weapons

With 34 nodes, you path a lot more thru velocity to get to certain nodes more effectively. You'll probably want to pick up more range nodes to be @ 10% or so (6% is meh) with some cooldown potentially.

PPC+Gauss (more like PPC+projectiles due to ghost heat):
https://kitlaan.gitl...34b3b#s=Weapons

It's barebones @ 32 nodes here. It's still easy enough to enable more cooldown and range nodes as necessary.

Here's laser and missiles:
https://kitlaan.gitl...24e7a#s=Weapons

Do we see a trend here? @ 34 nodes, laser duration is still a must plus pathing thru missile nodes helps make it easy. Velocity is still a good side path if needed. High explosives is overrated, but here's it's only useful for pathing.

Finally, missiles and projectiles (PPC, Gauss, etc.):
https://kitlaan.gitl...9bc74#s=Weapons

At 32 nodes, I'm pretty sure I could do better. In any case, there should be a "heatmap" of nodes taken, but it's probably Lostech @ PGI. Chances are based on your configuration, you will want various other nodes that are in the way.

In summation, you can build a pretty slim firepower tree with added range, cooldown, and velocity nodes as a sidegrade bonus while having most of the core necessary functionality to compete with everyone else. Some of this I didn't come up by myself to begin with (gotta credit people @ NS back when the Skill Tree was a PITA - it still is a PITA today/now). I'm sure it's not new and comp teams are probably using some variance to it, so it's just knowing what path makes the most sense (a node analysis by value-pathing - a Lostech PGI design decision) instead of randomly taking things you don't really need (if your structure tree is maxed out, you're doing it effing wrong).

The other trees don't need that much attention (unless you are totally clueless, in which case... I'm sorry to hear that), but it is what it is based on what is there.

Edited by Deathlike, 18 February 2018 - 07:51 PM.

[Kin3ticX]

I really don't like that kitlaan tool. All the color codes get in the way of actually being able to read it.

[Deathlike]

Kin3ticX, on 18 February 2018 - 08:10 PM, said:

I really don't like that kitlaan tool. All the color codes get in the way of actually being able to read it.

If there was a better one, I'd use it. Tarogato's version is out of date.

I'm sure you can generate ones of the real game anyways (could not use imgur to upload pics for some reason).

[Horseman]

Re: Operations, AFAIK the Improved Gyro nodes reduce shake from all sources (including your own jump jets). Jumper builds might find right side more attractive than the left side as a result.

Kin3ticX, on 18 February 2018 - 02:17 AM, said:

2. Only use mobility/speed tweak on lights and fast mediums. Maybe you have a rule that you dont want to go under 60kph or something... fine whatever.... but most of your mediums, heavys, and assaults will run fine without speed tweak.

3. For most mediums, all heavys, and most assaults.... no agility tree. For example, going from 81kph to 87kph in a HBR doesnt make a difference (nor does going from 75 to 80 in a Warhammer). I know it can be really hard to give up speed tweak, but just let it go. Its a huge point sink for what it is on bigger mechs. If you get in an infinity nascar in a slow mech, a few kph doesnt make much difference.

Small edge cases:

• Negative mobility quirks (eg on some TBR frankenbuilds), Anchor Turn / Torso Speed may be needed to compensate unless you like a Heavy that moves like a Kodiak.
  
• "Turret" configurations with no horizontal arm movement and relatively slow torso rotation such as Jagermech benefit from Torso Speed

Quote

5. For ECM mechs I get both ECM skills. I generally try to play non-ecm mechs without radar deprivation but sometimes I get a couple token points of radar derp.

LRM mechs will want Target Decay instead.

Quote

6. I only use laser duration skills on clan mechs (sometimes) and never on IS mechs. Its also just as easy to not bother with burn time skills at all but I definitely never use them with Inner Sphere. It feels like you need to get all 4 or go home since it only adds up to 10% burn time.

It's 15% for IS and several IS mechs have Laser Duration quirks in the 10% range (DRG-5N has 15%, COM-1B and SDR-5V have 20%). Stacking skill tree duration (max 15% for IS) with them leads to Standard lasers that have durations very close to an unquirked/unskilled Pulse version : https://docs.google....dit?usp=sharing

(EDIT: fixed the link)

Quote

7. Never bother with flamer skills or high explosive skills, both of these I never ever use.

High Explosive is relatively cheap for missile boats (literally just off the usual Cooldown/Heat Gen path to one of the Rack nodes).

Something else to note: deliberately speccing for Missile Spread makes sense when paired with Artemis (and ideally a Missile Spread quirk, which some mechs have), since it should (going by how all other skill bonuses stack with equipment and quirkes) stack additively. It's pointless for MRM builds and going by this also ATMs (they're supposed to have Artemis bonuses "built in", but apparently that is only lock speed and not spread), as well as any mechs that do not have Artemis equipped.

Edited by Horseman, 20 February 2018 - 04:03 PM.

[Kin3ticX]

Deathlike, on 18 February 2018 - 08:18 PM, said:

If there was a better one, I'd use it. Tarogato's version is out of date.

I'm sure you can generate ones of the real game anyways (could not use imgur to upload pics for some reason).

There is a button on the top right that fixes my problem

[Kin3ticX]

Here's a skill tree strategy created by King Harkinian



Order of priorities in how I use skill points, universally, from must-use to left-overs:

• 
  Left side of Survival tree... if you do this, you're only missing 2 Armor SPs from the right side. This is beyond mandatory, if someone's build doesn't have this, and it isn't rocking inherent armor quirks and massive armor numb ers to begin with, I consider it strictly inferior. Why lower your own TT[b]K'd.
  
• 
  Deep to 4/4 Laser Duraction, OR deep to 5/5 Velocity in Firepower tree. Firing lasers or peeps without these is just strictly... inferior, again.
  
• 
  Heat Gen and Cooldown quirks (all of them, which I cut corners through grabbing **** like LBX and Missile SPs to save a couple points). A full Firepower tree goes to 42 or 43 for me.
  
• 
  Left side of Sensors tree for 1 Seismic and 3 Radar Derp.

now it gets down to the littler things...

• 
  Right-of-center down Mobility tree for a few Hard Brake, Speed Tweak, Anchor Turn... this is low prio because I only do it when I want to take a mech to a specific baseline speed, but tonnage is getting tight (ie, want to take a mech to 100 kph, but bumping up engine would cost too much tonnage and eat into my wep build, so I can get the KPH up with Tweak... honestly the way I see things I do this on either really tonnage-starved Lights [like say dual PPC Panther] or surprisingly on really large mechs where my speed baseline and max DHS become optimal together]). Another reason is if a mech is slightly lower agility in general than what I would prefer, but these points make it 'comfy' in terms of mobility, which isn't often since I limit what mechs I play to generally being agile ones.
  
• 
  Right side of Survival, to grab the last 2 Armor SPs and fill out the Structure. If a mech has little structure this gets moved up for sure and becomes automatic. Full Survival tree is, I wanna say, 32 points, missing the 2 AMS and 1 -Crit SP.
  
• 
  Right side of Sensor tree for the 2nd Seismic. This one might as well be automatic, just because it tends to be where I only have a few points left to go for something, so I figure, hey.
  
• 
  And then typically the 4th Radar Derp right below that 2nd Seismic
  
• 
  Left-of-center down Mobility tree, if I can't get to my baseline optimally as stated above and need the other STweaks to get the speed baseline right, this is a tough one because it takes full Survival tree off the table.
  
• 
  If I was doing a full survival tree and no mobility (likely), and I don't want that 4th Seismic, I'll grab the first couple Accel/Decel quirks in the Mobility tree.
  
• 
  If I'm doing a ballistic build (me lol omg) then I gotta haz those +Magazine quirks in Firepower and will limit the Sensors.

Yes there's a method to the madness.

[Deathlike]

Kin3ticX, on 19 February 2018 - 12:24 PM, said:

There is a button on the top right that fixes my problem

1st world problems.

Kin3ticX, on 19 February 2018 - 04:34 PM, said:

• 
  Deep to 4/4 Laser Duraction, OR deep to 5/5 Velocity in Firepower tree. Firing lasers or peeps without these is just strictly... inferior, again.

Laservomit requires all 4 nodes to be effective (effectively mandatory). I think you can get away with only 4 velocity nodes since the 5th one (bottom most) is deep in the wrong place gated by less useful nodes (only good if Gauss+PPC were a thing, PPC+ACs may be OK). The difference between 12% and 15% is meh

Quote

• 
  Left side of Sensors tree for 1 Seismic and 3 Radar Derp.

100m of Seismic is really weak on its own, and means facehugging range. At that point, you are better off giving Seismic up.

Quote

• 
  
  Right side of Survival, to grab the last 2 Armor SPs and fill out the Structure. If a mech has little structure this gets moved up for sure and becomes automatic. Full Survival tree is, I wanna say, 32 points, missing the 2 AMS and 1 -Crit SP.

I don't buy into that on principle. Node gating on the price you pay can be used for so many other things easily. 21 points on Structure IMO is already pretty efficient for what you're paying for it.

There's a lot of stuff I don't buy into though, but the most notable ones are the ones you really pay for.

[Deathlike]

I would like to point out something specific (from reading the Reddit thread).

Reinforced Casing is next to useless at its current level. Think of it in terms of it being like 1% chance to evade or dodge a hit. In this case, it's crit chance. In order for it to be useful, it would have to be like 3 to 4% chance per node (you will pick up like 4 of those nodes in my tree) to have some impact (there are 8 of them max). When you get shot with a crit seeking weapon like MGs, you're barely randomly suppressing any of the MG damage since the chance to avoid said crit damage is abysmally low (I'm sure someone can do the math on that - 4% on my tree or the 8% max is a joke).

Math, PGI doesn't.

[Horseman]

Horseman, on 19 February 2018 - 07:43 AM, said:

It's 15% for IS and several IS mechs have Laser Duration quirks in the 10% range (DRG-5N has 15%, COM-1B and SDR-5V have 20%). Stacking skill tree duration (max 15% for IS) with them leads to Standard lasers that have durations very close to an unquirked/unskilled Pulse version : https://docs.google....dit?usp=sharing

After doing some extra math (edited the spreadsheet to account for this)... turns out that for IS mechs each Duration node is basically an extra Cooldown node even if you ignore the damage delivery aspect. About 1.2-1.3 cooldown nodes for LL, ERLL and SL, and worst case 0.875 for ERML, MPL, LPL,

Edited by Horseman, 20 February 2018 - 04:03 PM.

[Kin3ticX]

Horseman, on 20 February 2018 - 05:56 AM, said:

After doing some extra math (edited the spreadsheet to account for this)... turns out that for IS mechs each Duration node is basically an extra Cooldown node even if you ignore the damage delivery aspect. About 1.2-1.3 cooldown nodes for LL, ERLL and SL, and worst case 0.875 for ERML, MPL, LPL,

Since the skill tree is kinda subjective, I think burntime is a good place for disagreement.

My las vom Banshee for example is already pretty hot without the burntime skill. It doesn't need a faster cooldown. Under the strictest sense, sure, better burntime is always superior for any laser. However, i'd rather have the 4 points plus ALL the points it takes to get the 4th burntime skill elsewhere sometimes. I tend to use 3/4 or 4/4 burntime combined with HLL and ERML since those burntimes could use it the most.

[Kin3ticX]

Deathlike, on 19 February 2018 - 06:31 PM, said:

I would like to point out something specific (from reading the Reddit thread).

Reinforced Casing is next to useless at its current level. Think of it in terms of it being like 1% chance to evade or dodge a hit. In this case, it's crit chance. In order for it to be useful, it would have to be like 3 to 4% chance per node (you will pick up like 4 of those nodes in my tree) to have some impact (there are 8 of them max). When you get shot with a crit seeking weapon like MGs, you're barely randomly suppressing any of the MG damage since the chance to avoid said crit damage is abysmally low (I'm sure someone can do the math on that - 4% on my tree or the 8% max is a joke).

Math, PGI doesn't.

Yeah the 8% is so low we have no way of knowing if it works. It could be broken for all we know. I could see them buffing it by 3 or 4 times, but at .5% at a time this could take forever assuming they were planning it.

[Horseman]

Kin3ticX, on 20 February 2018 - 12:24 PM, said:

Since the skill tree is kinda subjective, I think burntime is a good place for disagreement.
My las vom Banshee for example is already pretty hot without the burntime skill. It doesn't need a faster cooldown. Under the strictest sense, sure, better burntime is always superior for any laser. However, i'd rather have the 4 points plus ALL the points it takes to get the 4th burntime skill elsewhere sometimes. I tend to use 3/4 or 4/4 burntime combined with HLL and ERML since those burntimes could use it the most.

This isn't exactly the takeaway I meant. Look at it from a min-maxing perspective.

• If you want more laser cooldown on a build where you maxed the hardpoints already and your heat management is still fine, then Laser Duration is the way to push in a couple nodes' worth - it doesn't hurt that two nodes are right next to Heat Gen nodes you'll probably grab anyway.
  
• If you're optimizing your skill tree on a laservomit build and can swap a Cooldown node for Laser Duration, for many commonly boated laser types (see ref) Duration has the same net effect with an additional benefit - and for a few is just plain better. IS LL boats could actually trade 4 Cooldown nodes for 3 Duration nodes and have the same net cooldown, effectively saving a node you could use elsewhere.

Incidentally, spreadsheet updated with Clan weapons and relative impact of a duration node on cooldowns compared to a cooldown node: https://docs.google....dit?usp=sharing

Edited by Horseman, 20 February 2018 - 10:54 PM.