12 replies to this topic

[AngrySpartan]

TL;DR
1. Current Hexagon tree has no distinct structure
2. Current tree has no flexibility for balancing different nodes to each other
3. Not user friendly, bulky and difficult to navigate UI.
4. It gives no freedom to customize your mech, instead forcing you to unlock "junk" nodes.
5. It's overcomplicated because it has too many nodes with low values (e.g. noone cares about 2% sensor range,whereas 15% will be different story)

Alternative - Witcher 3 style structure of a skill tree. Main Idea - to get to the most desirable nodes, you have to unlock more nodes in a respective tree.

To the topic (open images under Spoiler for better understanding):
Brief description and image for those who missed the Witcher (shame on you!):
1. 4 branches (different colors at the very top of the image = Trees in MWO terminology - firepower, protection, etc.)
2. Each branch has a line-up of 5 skills groups (vertical lines) - not relevant for MWO
3. Each branch has 4 "Tiers" of skills (horizontal lines), each consecutive "Tier" unlocks after you've spend enough points in the branch (e.g. 5 for Tier2, 10 for Tier3, etc.)
4. Each skill has several (1-5) levels - same as MWO's nodes (e.g. Velocity 1-10)
Image:

Spoiler

How to translate this to MWO:
Option 1. Direct translation:
1. Stack nodes the way module unlock worked before (e.g. Acceleration 1-10 is a single node with 10 lvls)
2. Arrange different nodes into Tiers (e.g. Arm speed, Acceleration and Decceleration is a T1, Turn speed, Torso twist speed is a T2, Speed tweak is a T3, etc.)
3. Set unlock thresholds into each tree (continuing with Mobility tree example: to unlock Torso Twist speed 5 nodes have to be unlocked, to unlock Speed tweak 30 nodes have to be unlocked). Non-linear increase in the thresholds is essential, cause it gives an incentive to specialize in a tree.
Something like this (Warning! Bad drawing):

Spoiler

Option 2. Indirect translation
1. Stack nodes into linear trees (e.g. Acceleration 1-10 arranged vertically into a column)
2. Arrange different nodes into Skill levels (e.g. Arm speed has 1-3 level, Acceleration and Decceleration has 1-10 lvl, Turn speed, Torso twist speed has 3-8 lvl, Speed tweak has level 8-10, etc.)
3. Set unlock thresholds into each tree for different skill levels (continuing with Mobility tree example: to unlock level 2 skill you need to unlock 2 nodes, to unlock level 3 skill - 4 nodes, level X skill - Y nodes, etc.). Non-linear increase in the thresholds is essential, cause it gives an incentive to specialize in a tree.
Something like this (Warning! Bad drawing):

Spoiler

How's it better than current hexagon abomination:

Spoiler

1. Both options have clear structure behind them.
2. Easier to balance! Every skill node (column of nodes for option 2) is independent from others, meaning it's easy to place them anywhere. Every change to individual node (column) does not affect other nodes, does not brake adjacent nodes.
3. Simple, compact and structured UI (less so for option 2).
4. Gives more customization freedom compared to Hexagon trees. At least you can pick your own poison when selecting "junk" nodes
5. Does not require obscure 1% "fill the gap" nodes to balance the trees. Balancing is easy to do by shifting nodes vertically between the Tiers, selecting proper thresholds, amount of levels for each skill and skill values.

How's it still imperfect:
1. Subjectiveness - there still could be opinions where to put each skill node. The same as current hexagon madness.
2. Trees have to be approximately the same vertical size to achieve better balance. Unless it's true, skill trees with fewer Tiers will be easier to unlock and (assuming Top tier bonuses are equally attractive) will be meta. Again, the same as current hexagon madness.

That's it. Hopefully, my English was fluent to explain the idea.Please leave a comment if you like the idea more than current PTS structure and even more important leave a comment why you don't like it.

Edited by AngrySpartan, 02 March 2017 - 07:57 AM.

[cazidin]

Honestly, why not make a skill tree like Mass Effect 3? Each class, or in this case, tree, would have large, specific bonuses and at certain points you'll get a choice. (I.e. 10% more Velocity, or 5% less laser duration)

That'd be simple, streamlined, and each skill upgrade would be meaningful rather than up to 15 levels of +1%.

[AngrySpartan]

cazidin, on 02 March 2017 - 08:15 AM, said:

Honestly, why not make a skill tree like Mass Effect 3? Each class, or in this case, tree, would have large, specific bonuses and at certain points you'll get a choice. (I.e. 10% more Velocity, or 5% less laser duration)

That'd be simple, streamlined, and each skill upgrade would be meaningful rather than up to 15 levels of +1%.

That's what Witcher 3 style skill tree will be able to achieve.

ME has waay fewer options that MWO and it's already class based game. Just look at each class as a separate tree with their own set of skills.

Edited by AngrySpartan, 02 March 2017 - 08:19 AM.

[cazidin]

AngrySpartan, on 02 March 2017 - 08:19 AM, said:

That's what Witcher 3 style skill tree will be able to achieve.

ME has waay fewer options that MWO and it's already class based game. Just look at each class as a separate tree with their own set of skills.

I played Witcher 1. That was an incredibly dull and uninteresting game with poor voice acting. I'm not playing 2 or 3. I don't care how many people think it's improved. The house may be built of solid concrete, shatter resistant windows and a wood burning pizza oven, but If the foundation is on a mud-filled faultline? I'm not living in that house, and I also don't care if that analogy is poor or confusing.

Back to your suggestion. Yeah, sure. It seems we agree!

[AngrySpartan]

cazidin, on 02 March 2017 - 08:21 AM, said:

Back to your suggestion. Yeah, sure. It seems we agree!

Well, thanks for feedback!

Offtop:
W1 was made on a knee with a budget you could spend on decent car on Aurora engine (just google Neverwinter Nights 1) and it's amazing what developers did with those resources. I sincerely wish you to try W2 and W3, cause they are amazing (and with proper AAA budget compared to W1).

[Nikkkname]

Hey, that's actually well thought! That's way better than current tree and It worked great in Witcher3

[AngrySpartan]

Nikkkname, on 02 March 2017 - 11:16 PM, said:

Hey, that's actually well thought! That's way better than current tree and It worked great in Witcher3

Thanks for feedback!

I really wonder why PGI have to invent the wheel all the time they want to change game? I'm curious if they ever looked at how other games manage to do that kind of skill systems.

[AngrySpartan]

As it seems, I am not the only one with similar thoughts on skill tree's structure, I'll leave a link to Pyed's topic here:
https://mwomercs.com...kill-tree-ftfy/

Same structure, with different branches, skill and Tier placement.

Edited by AngrySpartan, 04 March 2017 - 06:34 AM.

[Pyed]

AngrySpartan, on 04 March 2017 - 05:39 AM, said:

As it seems, I am not the only one with similar thoughts on skill tree's structure, I'll leave a link to Pyed's topic here:
https://mwomercs.com...kill-tree-ftfy/

Same structure, with different branches, skill and Tier placement.

I came up with basically the same idea because neither of us are the first to think of it. This has been done before, and there are dozens, probably hundreds of AAA games that have used the same kind of tiered skill structure.

There's a reason you never see a skill tree like this web concept (please let me know of another game that uses it if you do--bonus points if that game is actually successful). I'm sure PGI isn't the first to come up with it, and I'm sure it has always gotten passed up for something better.

[AngrySpartan]

Pyed, on 04 March 2017 - 04:02 PM, said:

There's a reason you never see a skill tree like this web concept (please let me know of another game that uses it if you do--bonus points if that game is actually successful). I'm sure PGI isn't the first to come up with it, and I'm sure it has always gotten passed up for something better.

That's a single game that comes to my mind: Path of Exile (Diablo clone, free-to-play as far as I recall). Never tried it myself, but as far as I am aware skill tree is the only thing worth attention in it. It takes spider web skill tree to the totally different level of insanity, and to be honest my eyes start to bleed every time I see this:

Spoiler

On the other hand: Witcher 3, Mass effect, every Ubisoft game, literally every modern Blizzard game (and they know how to balance things). All those use simple and structured skill systems. It takes 30 second to minute to understand how it works and what might be important. Even AAA RPG games, where levelling character is waaay more important than in the arena shooter (like MWO) use simple rule: 'less is more'.

What bugs me even more is that by using spider-web structure PGI traps themselves. It's so much easier to balance something like Witcher 3's system than what we have right now.

[Alistair Winter]

I don't think hexagon structure is innately flawed, but PGI is limiting themselves by making it such an intricate web, where practically every hex is connected to 4-6 other hexes, on top of the fact that they have randomly placed associated skills all over the branches, to the point where their actual location isn't really that important, because you're forced to unlock the whole branch anyway.

I like the Witcher Structure, and I like the more traditional linear skill tree. The hexagon structure could easily be used to represent either the Witcher structure or the tradtional linear skill tree, or both.

UI design has never been PGI's greatest strength and the number 1 critique of UI 2.0 for MWO is that it's style over substance. It looks pretty at a glance, but it's been (especially in the early days) rather unintuitive and required a lot of unnecessary clicking. So it comes as no surprise that the new skill tree looks cool at a glance, but fails to do its job properly.

[AngrySpartan]

Alistair Winter, on 04 March 2017 - 04:35 PM, said:

UI design has never been PGI's greatest strength and the number 1 critique of UI 2.0 for MWO is that it's style over substance. It looks pretty at a glance, but it's been (especially in the early days) rather unintuitive and required a lot of unnecessary clicking. So it comes as no surprise that the new skill tree looks cool at a glance, but fails to do its job properly.

Yep, that's definetely Style over substance right now too. Linear tree with 2-3 choices with fancy hexagons definetely looks good. However it loses it's style as the complexity grows and than the UI stands in the way of balance. And that's how we came to the current PTS skill tree...

Alistair Winter, on 04 March 2017 - 04:35 PM, said:

I don't think hexagon structure is innately flawed, but PGI is limiting themselves by making it such an intricate web, where practically every hex is connected to 4-6 other hexes, on top of the fact that they have randomly placed associated skills all over the branches, to the point where their actual location isn't really that important, because you're forced to unlock the whole branch anyway.

True, just wish PGI would acknowledge that too. Pull a node and the whole web collapses. At that point it makes no sense to have connections between nodes at all, hence Witcher 3 structure works so well. It also address granularity - no one cares about 1% bonuces spread all over the place.

Edited by AngrySpartan, 04 March 2017 - 04:52 PM.

[Pyed]

Alistair Winter, on 04 March 2017 - 04:35 PM, said:

I don't think hexagon structure is innately flawed, but PGI is limiting themselves by making it such an intricate web, where practically every hex is connected to 4-6 other hexes

The hexagonal shape isn't the problem. It's just the web part that's the problem.