26 replies to this topic

[Excalibaard]

So we're 2 weeks into the skill tree and we've gotten more accustomed to it. It's good for the most part, but there is also some bad and some ugly that I'd like to see sorted out (sorted from easiest to hardest fix):

- UI: Unclear respec payments

It's very impractical that you can't see which node you have selected for respec with XP and which with GXP. If you've bought skillpoints and run out of one of the types of XP for the respec, you have to basically take away every node and build it from the ground up. Clicking only half of a node without direct previous indication, and the skill tree moving when you click and accidentally drag just a little bit is also not improving the user friendliness of speccing skills.
Color the respecced nodes in the color of the type of experience they take (or remove respec costs as there's no explicit need or request for an XPsink).

- UI: Inconsistent/confusing naming

Skill trees are named differently between the tabs and the actual skill tree itself (Mobility is called both Agility and Mobility). Also, some of the skills are named impractically. People familiar with the old system will know what Kinetic Burst means, but just naming it Acceleration would be much more straightforward and there's little to no flavour attached to these names.

- UI: Huge unnecessary space usage

The Skill Tree consists of so many elements that on most regular 1080p monitors you'll have to zoom out very far to get an overview of your assigned skills. Players at a lower resolution (due to performance issues) experience even bigger problems with the inefficient usage of space. Instead of having everything in one wide area, allow for a different way to organise these separate trees, akin to the compact Layout:Column style for example.

- Functionality: Networking

It would be nice to be able to go 'up' the skill tree as well, It looks like a network and there seems to be no specific advantage to lower nodes than higher nodes in most cases except Speed Tweak. It could help to work around unwanted skill nodes like Shock Absorbance and Speed Retention that are terrible nodes to take.

- UI: New User Experience in Firepower Tree

For mixed builds (which is very common for people buying/building their first mech) the firepower tree is just a nightmare to decide which nodes are best for you, resulting in a lot of wasted skill points early on, which is a negative experience for new players. This could be fixed by suggesting nodes automatically based on current loadout (add in a weighing system), or user selected classes with predefined node suggestions. (credits: Insanity09 on classes)

- Functionality: RNG-enhancing skills lack feedback. (credits: Trev Firestorm)

Things such as the UAC Jam Chance reduction give very little response for the skill points invested, because you don't notice their underlying dice manipulation at all. Instead of the UAC jam chance, a UAC jam duration reduction would feel much more rewarding to spec into (This argument also partially applies to High Explosive Missiles)

- Functionality: Not enough skills for non-weapon, non-consumable equipment.

Some parts that I want to spec into are missing. increased ECM counter duration for PPCs and TAG target painting retention would be very useful for scouts, also an increased MASC duration/threshold would be nice, because there are ECM and jumpjet enhancements, why not also for other mech-specific 'abilities' like MASC.

- Balance: ECM and Radar Deprivation - Unspecced vs specced

The differences between unskilled and skilled ECM and Radar derp are too large, making these nodes almost mandatory or and the most affected weapon system - LRMs - either reign supreme, or are completely useless.
Bringing these two extremes closer together will make this controversial weapon system more rewarding to use and counter in all tier levels, and can be balanced easier. As example values, never allow for 100% radar derp, but only go to around 75%, and have ECM 50% effective without skills and each node adds 10% instead of 22,5%)

- Balance: redefine Operations and Jumpjets.

The Operations Tree and Jumpjet tree need some work, as they are now they're too weak.
Jumpjet just needs a little numbers buff, too small for the skillpoints invested and maybe heat shielding is useless as jumpjet and movement heat have almost no impact compared to weapons on most builds. This leaves the tree in particular very little nodes, which could be put in a different tree on the side, just like ECM.
Operations is a let-down because the strong parts (HC, CR) are bogged down by skills nobody is interested in. Speed Retention is extremely niche and unimpactful. When legged, you die soon after, 30kph or 50kph. Improved Gyro's might be nice for assaults, if the shake was the problem and not the vision-obscuring explosions of shake-inducing ballistics. Hill Climb is very dependent on map and may be suited better in the mobility tree.
Instead of these niche/out of place upgrades, I suggest to make this a tree for all non-weapon, non-consumable "Equipment". That's NARC, ECM, MASC, TAG, TC, AMS, Heat Sinks (and maybe also combine Jumpjets) in one tree. That also means moving some out of place AMS from structure, and ECM from Sensors to this tree.
Gyro's and retention can be dependent on Mech weight and no one will miss them, as their modules weren't used generally either.

- Functionality: Multiple starting points Firepower due to tree size.

With the great flexibility of the skill tree, it's weird that I can only start from Range in the firepower tree. Since it's a special tree in it's own right due to the much larger size, at least let us start at different locations in the tree with Heat Gen, Cooldown or Range, so not every build gets clogged with range nodes we might not even want and to account for the fact that this tree requires much more nodes invested to get to a certain side.

Two major problems with current tree design:

- Optimal Pathing (copy&paste)

Paths in the skill tree are basically interchangable between mechs. As predicted, instead of allowing extremely precise optimization, the optimal paths for certain bonuses have already been found. MetaMechs has written an article/guide for every playstyle in the game, and the 'incremental choices' are bogged down to taking 1 of 2-3 paths. Having to click 20 nodes down a tree instead of just 1 to 'choose your path' is very inefficient.
Unless you're planning a proper redesign (see end of the post), it'd be really nice to have/save certain presets that I can save in order to copy them to most of my mechs.

- Unrewarding Grind (lack of milestones)

Grinding out XP for mechs is infinitely less rewarding. With the subgoals of Basicing or Eliting a Mech, there were incentives in the form of milestones to try and get that XP. With the incremental nodes, there's no incentive to grind for 'that 0.75% cooldown node'. Also adding in the C-Bill cost of skill points, makes you less eager to buy them. Most people are always saving up for at least 1 mech/loadout, others will have enough C-bills that the skill point bill-sink hardly affects them.
A big drive for playing your non-standard mechs is removed with the tiny incremental enhancements, which leads me to suggest a slight rework of the system into larger, more impactful nodes.
In order to have meaningful progression for the grind, make nodes stronger and less and/or permanently unlock nodes for all variants of a chassis as a limited form of permanent progression.

In the spoiler I've written some changes that I'd like to see to mitigate these larger problems, while including the smaller changes I talked about earlier into a slight system rework. It still uses the regular skill points, but is easier to spec into.

Spoiler

Most of the 'choices' in these trees could also have been done without a 'skill maze'. The initial argument for the skill tree to allow for personal customization turns out to be highly incomprehensive to new players and for experiences players there are just 2 or 3 optimal paths to take.
This could be mitigated with a change into less nodes with more power. This also means the differences between certain choices are more noticable, and that gaining enough XP/Bills for a new node is more rewarding.

For example here I've adjusted the current basic skill trees into just a few 'super nodes' that are easy to comprehend: You can take any amount in any tier, but you can only take from the next tier after getting at least 1 point in the previous tier.
Let's say each of these super nodes costs around 7 skillpoints in the current economy system, so you don't have to redo the economy around skillpoints. This means you can get 13 of these, which is roughly equal to how you can spread your points currently. The firepower tree is a little difficult to condense in this system, but here's a first draft anyway.

Mobility:
T1: Torso Yaw+Pitch +10% vs Accel+Decel +20%
T2: Torso Speed +20% vs Anchor Turn +20%
T3: Speed Tweak +7.5% vs Torso Yaw+Pitch +10% & Accel+Decel +20%

Structure*:
T1: Armor +5% vs Structure +7.5%
T2: Critchance -10% vs Overheating damage -25%
T3: Armor +10% vs Structure +15%
*Armor&Structure dependent on mech weight, AMS moved to other tree because it's got nothing to do with integrity. Falling damage reduction is niche/situational, uninteresting as module.

Jumpjets:
T1: Vectoring Thrust 100% vs Thrust Heat -20%
T2: Thrust Duration 20% vs Initial Thrust 20%

Operations**:
T1: Heat Containment 5% vs TAG decay +2 vs NARC Duration vs MASC Duration
T2: Coolrun 5% vs AMS Overload vs ECM +20%
T3: Heat Containment +10% vs Coolrun +10%
**Needs a complete redesign of contents, the filler skills are not good enough.

Sensors:
T1: Advanced Zoom vs Target Info+Retention vs Target Decay - Least powerful modules/skills
T2: Seismic 200 vs Radar Derp 50%. - More powerful 'modules'

Comsumables (2 slots is standard):
T1: UAV Range+Duration vs Coolshot Effect vs Strike Accuracy+Range
T2: UAV slot vs Coolshot Slot vs Strike slot -(maybe freely equipped?)

Firepower (difficult, maybe keep some kind of tree-like single point investments between larger multi-point nodes):
T1: Velocity +10% vs Heatgen -5% vs Cooldown 5% vs Gauss Charge
T2: Range +5% vs LBX&Missile Spread vs PPC ECM scramble +2.5 sec
T3: Range +10% vs Heatgen -10% vs Cooldown 10% vs Ammo reserves 10%
T4: Laser duration -10% vs UAC Jam Chance vs High Explosive

Constructive criticism & own opinions on how to change the skill tree (NOT THE ECONOMY) are welcome responses.

Edited by Excalibaard, 29 May 2017 - 06:17 AM.

[Leopardo]

I can only agree .

[Gasboy]

What I would like is for the trees to not move. Half the frustration is clicking, holding down the mouse button a fraction too long, and moving the tree instead of a node getting clicked.

Make it so that you need to hold a button down + move the mouse, so that you can click the dang nodes easier.

As for your skill tree suggestions, Excalibaard, I agree in full.

[50 50]

Don't think there is a difference with the XP based on what you initially unlocked the node with.
Once it's undone, it's just the 400xp to reallocate but shouldn't matter if you chose to pick mech XP or general XP.

[Excalibaard]

50 50, on 22 May 2017 - 07:14 PM, said:

Don't think there is a difference with the XP based on what you initially unlocked the node with.
Once it's undone, it's just the 400xp to reallocate but shouldn't matter if you chose to pick mech XP or general XP.

If I selected a build path beforehand, using some respecs that would cost 400XP/GXP, and then buy the necessary skill points for the rest of the build, It's happened multiple times that after buying skill points (sometimes with the same type of XP I used to respec the nodes, I didn't have enough XP left to re-spec those nodes. That's >very< annoying.

[Leopardo]

I can only hope that PGI will read this and take a tone .... god plz.

[Remover of Obstacles]

Thank you for taking the time to put together a well thought out response.

[Trev Firestorm]

Mostly agree but I think if you buff ECM again do so in small steps, like nodes at 15-18% each rather than just jumping straight to 10% each.

Also Hill Climb is way way way more useful than speed retention, it was a terrible choice when it was a module, but that was only because it cost an extremely limited slot, now it costs very little. Speed retention on the other hand, once you're legged you're basically done anyway in most mechs, and legged assaults (the only ones likely to survive legging long enough to benefit) won't see any meaningful speed boost anyway.

Improved Gyros, yep the explosion sprite on the cockpit is the real problem, worse with this patch(or had it been doing this already and I didn't notice before?) hits to /any/ part of your mech (even your rear) will also block your vision.

Radar Dep, I'd rather see it a flat value directly one to one counter to target decay nodes for optimal balance but having a sub-100% max would also work.

Completely agree with all the rest OP.

My own additional suggestion, UAC Jam Chance nodes... It's been a bit unclear on how effective they have been, I want to see them converted instead into UAC Jam Time Reduction nodes. This makes them very apparent exactly how effective they are and makes them more consistant, rather than modifying RNG they work to mitigate the effects of that RNG.

[Sarsaparilla Kid]

Lots of good points, and one I'd like to throw in, because I haven't seen it discussed elsewhere...what happens when they add the new tech and adjust the weapons tree or some other tree? Are they going to reset everyone's trees, or are we going to have to set aside more SP or XP & C-bills to respec every time major changes occur? They did say these trees are not set in stone, but they had better be equitable about the process of constant changes, and hopefully without having to respec the full tree again. Every other game I've played, once you've earned enough XP to open up your skills, you can switch back and forth without cost, even after changes have been made to the various nodes or skills. With that uncertainty, I don't really want to commit to setting up the skill trees for all my mechs, especially my favorites, because adjusting the trees for lots of mechs over and over could get very expensive. If anything, I'll just experiment with underperformers until I know for sure...

Per the Skill Tree Q&A: The Skill tree will be monitored heavily after release. We will have new nodes introduced with the new tech in July, and we will continue to refine the overall Skill Tree balance as we move towards introducing the nodes for that new tech.

[Brom96]

That is actually scary, you know, the prospect that we could end up with the need to respect the trees after every major update.

About ergonomics, I actually had to ask people in chat how many trees there are. No matter how I set the resolution of the screen I could never see all 7, and in usual resolution I could see only 3.

[Sarsaparilla Kid]

I set up my mouse wheel to scroll my Zoom in/out while in combat, so when it comes to looking at the Skill Tree, I can't use the mouse wheel to even zoom it in/out...I get to see maybe a half of a tree at a time. Would be nice if they had picked a different way to Zoom the Skill Tree instead of the mouse wheel, as I am sure others may have the same set up.

Edited by Sarsaparilla Kid, 23 May 2017 - 03:49 PM.

[ALEXANDER CARLYLE]

They tweaked this over many months. I don't think they want to re-tweak it now after only one week. Give em a month to fix bugs and not reinvent the wheel.

[Sarsaparilla Kid]

ALEXANDER CARLYLE, on 23 May 2017 - 04:35 PM, said:

They tweaked this over many months. I don't think they want to re-tweak it now after only one week. Give em a month to fix bugs and not reinvent the wheel.

They won't be reinventing the wheel...they'll just be adding a lot more spokes...

[Brom96]

During this year or so, I honestly must say, I saw some really nasty mess, the most memorable being that mini map fiasco during the last summer. Ergonomics of that one reminds me of the functionality of this tree thingie. Now, could they just make the ribbon with the tree names to be scalable and have, I don't know, navigation keys?

[Excalibaard]

I've been away for a few days, thanks for all the responses, I'll try to get back to them as soon as possible.

[Insanity09]

Good bit of work there, OP. By and large I agree with what you stated.

Another improvement I might suggest, to bring roles into it a bit more, would be to apply some sort of classes/roles to the tree. You decide what role you would like to be and that modifies the skill tree by either making certain nodes unavailable, or by altering the numbers for certain trees.
For example, if you choose to make a mech a tank, you would have access to the entire survival tree, but the mobility nodes you could choose from would be reduced. Or, for the number mod method, for a tank the mobility nodes would be maybe 3/4 value, while survival would be 5/4.
Those are just spitball ideas, the actual numbers could be played with.

Edit: Ah, one other thing, most games with skill trees have the idea of a capstone, or significant skill at the top of any given tree or spec line, which you can get access to by spending a sufficient number of points in a particular area, it would be nice to see something similar here, especially with most of the node values feeling very lackluster.

Edited by Insanity09, 27 May 2017 - 12:35 AM.

[Excalibaard]

Trev Firestorm, on 23 May 2017 - 10:37 AM, said:

Mostly agree but I think if you buff ECM again do so in small steps, like nodes at 15-18% each rather than just jumping straight to 10% each.

Also Hill Climb is way way way more useful than speed retention, it was a terrible choice when it was a module, but that was only because it cost an extremely limited slot, now it costs very little. Speed retention on the other hand, once you're legged you're basically done anyway in most mechs, and legged assaults (the only ones likely to survive legging long enough to benefit) won't see any meaningful speed boost anyway.

Improved Gyros, yep the explosion sprite on the cockpit is the real problem, worse with this patch(or had it been doing this already and I didn't notice before?) hits to /any/ part of your mech (even your rear) will also block your vision.

Radar Dep, I'd rather see it a flat value directly one to one counter to target decay nodes for optimal balance but having a sub-100% max would also work.

Completely agree with all the rest OP.

My own additional suggestion, UAC Jam Chance nodes... It's been a bit unclear on how effective they have been, I want to see them converted instead into UAC Jam Time Reduction nodes. This makes them very apparent exactly how effective they are and makes them more consistant, rather than modifying RNG they work to mitigate the effects of that RNG.

ECM: yes, that's all subject to balance and further internal testing. I was making this write down with 50% + 10% additively per node. minimum 50% range reduction, maximum 70% (instead of 30% to 75%).

Hill Climb: I haven't done it justice putting it in the same line as the proportionally less useful speed retention, indeed. I'll change the writing there.

Derp: I specifically didn't choose linear values, because Decay does that already. If Decay and Derp (which aren't worded as opposites so may be confusing to directly oppose) both were linear to balance each other out, the situations in which only derp is present will be unbalanced --> no decay specced and someone gets -2sec Derp on 2 sec, is the same as 100% Derp. By using percentages I think it's easier to distinguish and to balance out the most extreme cases especially approaching a 0 as value.

I second your opinion on the UAC nodes, that is a great alteration and I'll put it in the main thread when I clean it up a little.

Edited by Excalibaard, 27 May 2017 - 02:28 PM.

[Excalibaard]

Insanity09, on 27 May 2017 - 12:24 AM, said:

Good bit of work there, OP. By and large I agree with what you stated.

Another improvement I might suggest, to bring roles into it a bit more, would be to apply some sort of classes/roles to the tree. You decide what role you would like to be and that modifies the skill tree by either making certain nodes unavailable, or by altering the numbers for certain trees.
For example, if you choose to make a mech a tank, you would have access to the entire survival tree, but the mobility nodes you could choose from would be reduced. Or, for the number mod method, for a tank the mobility nodes would be maybe 3/4 value, while survival would be 5/4.
Those are just spitball ideas, the actual numbers could be played with.

Edit: Ah, one other thing, most games with skill trees have the idea of a capstone, or significant skill at the top of any given tree or spec line, which you can get access to by spending a sufficient number of points in a particular area, it would be nice to see something similar here, especially with most of the node values feeling very lackluster.

The idea of classes is good indeed. Personally, I think the way you suggest number mods makes it a bit confusing, but you could motivate people into certain trees by unlocking some nodes for free, or give an increased skill point cost on trees outside the intended role of the mech (f.e. Dragon gains Structure and Mobility as its assigned trees, investing in other trees costs double skill points (cost balance ignored for now), but this could also vary between variants of a mech. That gives it roles and could be a way to balance mechs against each other without taking away the opportunity to spec into something else.

Nevertheless, I've added the general idea of classes to the main post with your name in the credits.

Edited by Excalibaard, 29 May 2017 - 06:09 AM.

[tokumboh]

I think the problem of making progress will always be an issue as is how you spend your SP if you are picking them up 1 at a time what is seen as important and the fact that often the best utilisation is when you all 91 at disposal

[Inappropriate1]

And as predicted.. the marketplace has spoken. There should have been a huge uptick in use after the kill tree but das not happening. http://steamcharts.com/app/342200

Now Battletech approaches. They will struggle even more. REPENT PGI.