Fractal Softworks Forum

Please login or register.

Login with username, password and session length
Advanced search  

News:

Starsector 0.98a is out! (03/27/25)

Poll

What do you think the tree needs the most? (you can vote on as many as you want simultaneously)

More and clearer explanations of each skill (in-depth explanation of what the thing really do)
- 21 (8.3%)
Less multiple bonus skill nodes and a greater number of them ("Let's do this bit by bit" said Jack the Ripper)
- 18 (7.1%)
A clearer ascending progression of skill power (from poopy to spoopy) (sorry)
- 13 (5.2%)
More synergy within the same branches (going all in for a tree to give better results than now)
- 32 (12.7%)
Less linearity (Modifying the tree to a more branching shape)
- 42 (16.7%)
More emphasis on top tier skills (having them consolidate a build more clearly)
- 22 (8.7%)
Simply more skill points (everything the same I just wanna get all of them stuffs)
- 33 (13.1%)
Skills locked behind special missions (could be high tier ones that feel cheap or not immersive to learn without doing something first)
- 17 (6.7%)
Less DP limits in skills (diminishing returns make me too sad)
- 46 (18.3%)
Everything is perfect as it is, and if you criticize my skill tree waifu any longer I'll stab you
- 8 (3.2%)

Total Members Voted: 84


Pages: [1] 2 3 ... 6

Author Topic: About skill tree design  (Read 9162 times)

Ad Astra

  • Captain
  • ****
  • Posts: 263
  • Are Reapers strawberry flavored?
    • View Profile
About skill tree design
« on: April 24, 2021, 09:47:14 AM »

Hello everyone, I hope you are all doing well!

First of all I'd like to clarify that this isn't a take on skills themselves, which has already been discussed extensively and I don't have any insight that hasn't been shared, but an attempt at discussing the structure of the skill tree itself.
What this refers to, is the way these nodes are arranged, how they interact with one another, how the progress through them functions, and what the player can expect from skills when they start a new game.

I'll be talking about the subjective and emotional reactions of the player towards the design, so naturally everything here will be merely opinions, and hopefully useful ones at that.

Let's start this by talking about classic skill tree design, and that is linked to the very name we give to these things.
A tree possesses roots, a trunk and branches. A tree has been used to depict both heritage and hierarchy through the ages, a tree grows organically and intertwines its segments often. All of these are reasons why the progression of skills throughout a character's (and the player) journey, ingrained iconography, "sense" of progression and notion of growth (in power mostly), and lastly a very important matter for the designer, the freedom that an asymmetric structure brings for modification and development.

If we contrast this to the current  (and previous) skill tree designs we see that several of these advantages are not present, being the most subconscious reason why every implementation of the skill system faces so much discord.
We don't "feel" the power on many of the higher tier skills clearly, meaning that their position, their hierarchy comes off as meaningless, we don't "feel" freedom in the path that is taken to the upper branches because there are too little points and as such, little choices to be made, and skills that are desirable within a same playstyle often exclude each other.
The constraining and unnatural feeling that the skill tree causes in the player, is a consequence of the constraint a symmetrical structure causes on the one who creates it, the self applied demand to make every tree "equal" is very limiting on design, and those limitations are then felt by whoever interacts with such a system. This "unsettling" feeling only becomes greater when immersion within the world is significant, causing yet a further obstacle to the flow of the experience.

The same resources at hand, could be used to produce far greater results should they be arranged differently. Each point right now involves several stacked bonuses, while this is probably intended to make each choice more significant, the "why" each of those bonuses is stacked to each other, and then contrasted to other such node is never clear, never truly reasonable within the world, always making the fact that they are design choices be felt too strongly, something that within a classical tree would be mellowed by separating branches thematically, and allowing every early node to be chosen, knowing full well that the latter, "more powerful" or interesting choices are being sacrificed every time a small convenient node is chosen. Specialist vs Generalist, in every ramification of the tree the choice becomes clearer, heavier. As long as the player is drawn to and enjoys what each choice provides, then no wrong choice is ever made, sub-optimal only hurts when it equals to unviable.

If branches are separated on each of the 4 main ones we have, the choice of what affects player ships, what affects fleets in combat, and what affects the campaign level would be so much clearer for new players, if nodes are separated in a greater number of less significant choices, further explanation can be applied to each and every one of them, helping a newbie understand without having to wiki search what every one of those small stacking bonuses mean on every node, just hover a mouse over them and there would be the explanation with examples and all.
Asymmetrical trees allow for padding to be removed, we don't have an initial skill point investment to even access each section anymore, having more or less nodes carries no harm. Being able to add and remove nodes without a hitch allows for modifications to be easier, freeing creativity within development. Nodes can intertwine at chosen spots to allow different paths to reach the same destinations, versatility and variety within a limited system make for great fun experimenting.

Driving ourselves into a corner is what our mind is best at, seek success in freedom, freedom from our own self placed constraints.

Now to the most important aspect of all. You might be thinking "there are no trees in space dumbass", that's accurate. So while an organic looking design might be out of place, an intertwined mess of cables and munition belts, strategic points in a war table, shapes in an electric board, or trade routs in a merchant's map would all conform to these abstract shapes that are so precious for facilitating design, while fitting the themes of the skill tree perfectly (It could be anything really, I'm just throwing stuff on the table to make it easier to imagine).

To end I'd like to say that by no means is this an attempt of telling how things should be done, or a disregard of the systems we've seen so far (with their own special charm), but rather an attempt at creating discussion and the following inflow of ideas and clarity it causes.
If after reading this, sorting out how to solve the rejection towards a skill system when its established feels even a tiny bit easier, I'd be truly glad and consider this attempt a success.

Thank you for reading this wall of text and please share any opinions you have!

tl;dr: rows=pretty and original but tree=easier. ???
Logged
You can park your spaceship anywhere you want if you get along with pirates

Mordodrukow

  • Captain
  • ****
  • Posts: 275
    • View Profile
Re: About skill tree design
« Reply #1 on: April 24, 2021, 10:16:35 AM »

Quote
What do you think the tree needs the most?
Different players  ;D
Logged
Spoiler
[close]

Ad Astra

  • Captain
  • ****
  • Posts: 263
  • Are Reapers strawberry flavored?
    • View Profile
Re: About skill tree design
« Reply #2 on: April 24, 2021, 10:20:57 AM »

Different players  ;D

Going for the quick fix I see, always worth considering I guess  ;D
Logged
You can park your spaceship anywhere you want if you get along with pirates

Mordodrukow

  • Captain
  • ****
  • Posts: 275
    • View Profile
Re: About skill tree design
« Reply #3 on: April 24, 2021, 10:29:49 AM »

You see, before deciding what the skill tree needs we need to agree what goals we are pursuing. And we cant do even that, because some people, for example, like limitless skill points, while the other want a cap. And this is only one example, there is more. And we cant make happy everyone simultaneously.

Current skill tree pursued specific goals like "make it more newbie-friendly". I guess, it needs few tweaks, but it is by any means not a bad skill tree. Old one was not bad either.
Logged
Spoiler
[close]

Ad Astra

  • Captain
  • ****
  • Posts: 263
  • Are Reapers strawberry flavored?
    • View Profile
Re: About skill tree design
« Reply #4 on: April 24, 2021, 10:38:23 AM »

Yeah, I agree with that. I started this thread as an attempt to see what most people expect from the tree in general, rather than to try and add more mechanics or fine tune the ones present. Many times attempting to balance skills comes from the frustration of not feeling comfortable with the way the tree is designed, rather than the skills themselves. Trying to get the general feel of a desirable system might help the developers with satisfying those who feel bad with the current skill tree, by modifying it towards a more "globally agreeable" state. Of course some fringe opinions will always go unsatisfied, but that's an inevitable part of not being willing to negotiate a bit I guess.
Logged
You can park your spaceship anywhere you want if you get along with pirates

Mordodrukow

  • Captain
  • ****
  • Posts: 275
    • View Profile
Re: About skill tree design
« Reply #5 on: April 24, 2021, 10:48:15 AM »

In fact, the example of great skill tree for me is one from Path of Exile. It is complicated, so, you dont see obvious best choices with your naked eye. Its bonuses divided by many skill nodes (you need to pick "useless" +10 dex to get to essential nodes etc). It offers multiple routes and allow to swap em. But it also has flaws. First of all: it is complicated  :D. I can deal with that. Many people will not want to.

Another good example: D&D 5 edition and D&D 3.5 edition (or Pathfinder 1st edition). 5E is simple, really best for new player. It also fixed some issues we had in previous editions. But i ll choose 3.5 one, because 5E is too primitive for me.

Will PoE skill tree or D&D 3.5 system good for Starsector? Well, may be, but i m not quite sure.
Logged
Spoiler
[close]

SCC

  • Admiral
  • *****
  • Posts: 4573
    • View Profile
Re: About skill tree design
« Reply #6 on: April 24, 2021, 10:56:45 AM »

Better tiers 4 and 5. I even made a thread about this.

For Combat, Phase Mastery should definitely be tier 5, because it really is strong. Shield skill is okay, but it could fit at any tier in the tree. My feeling is similar about Systems Expertise, but I suspect that a certain phase cruiser made Alex put it up there. Missile Specialisation is alright, but I could see it at any tier above 2. It's not something that really only fits at the top tier. Of these skills, however, the only one that I would call nearly universal is the shields one and maybe missile spec. Phase skill is obviously specialised, while Systems Expertise heavily depends on base ship system being desirable in the first place (or, in the case of Fortress Shield, having anything you can buff with it).

In Leadership, there are two kinda fine skills at tier 4 (more officers skill needs help), but tier 5s are jokes. They easily could just as well be tier 1s.

Technology is mostly fine. Flux regulation is a universal buff. Phase corps is neat for phase ships, but phase ships are also kinda busted at the moment. Spec mods gives you more variety in loadouts. Automated Ships gives you a pet death god, though I wish options other than Radiant were competitive.

Industry has one okay tier 4, one terribly busted tier 4 and two either desired or useless tier 5s.

Ad Astra

  • Captain
  • ****
  • Posts: 263
  • Are Reapers strawberry flavored?
    • View Profile
Re: About skill tree design
« Reply #7 on: April 24, 2021, 11:08:43 AM »

Yeah, when talking about branches I was thinking about trees like POE, of course such level of complexity is unneeded here, but the branching paths mechanism can be imitated with just 3 or 4 rows in each section, ideal tier 5 skills to my opinion should work like Outward's "breakthrough skills" which are limited in number for any playthrough, essentially locking you into a build for the remainder of the game (which is fitting for a tier 5 skill).

Another example of great build delimitation is Dragon Age: Origins, your build could choose 2 specializations, most of which needed to be unlocked through story missions, they added great flavor to for example becoming a blood mage (Blood mage's in Starsector would be automated ships owners, both carry great danger and social rejection in their respective worlds).
Logged
You can park your spaceship anywhere you want if you get along with pirates

Serenitis

  • Admiral
  • *****
  • Posts: 1541
    • View Profile
Re: About skill tree design
« Reply #8 on: April 24, 2021, 12:02:37 PM »

What I want out of the 'tree' is to be able to pick all the QoL skills I want, and then fill in the leftover points with whatever skill seems useful at the time.

The previous skills setup in 091 wasn't perfect by any stretch. The 'dead' aptitudes in particular were not great.
But it did have the flexibility to accomodate every kind of player by virtue of not placing any kind of restriction (beyond the level cap) on them.

The major failing (imo) of the new system is that many of the choices it gives you are not choices at all, some of the pairs have a definite "I will always pick this".
And for others I want both enough that I have to spend multiple points in order to get 1 skill.
There are also some pairs where I don't really care about either choice, which isn't a great incentive for getting past them.

One possible way out of this is to ditch the pairing thing entirely, and arrange each line of skills in order of 'power'. However you define that.
As you spend points on skills at the lower end, you gradually unlock skills higher up in all lines simultaneously.
The total number of points you spend determines the skills you have access to.
Spoiler
[close]
This would allow the player some amount of freedom to pick whatever they want while still 'gating' the more powerful abilities.
Logged

IonDragonX

  • Admiral
  • *****
  • Posts: 816
    • View Profile
Re: About skill tree design
« Reply #9 on: April 24, 2021, 12:07:54 PM »

Oopz. Nvmd
« Last Edit: April 24, 2021, 12:19:02 PM by IonDragonX »
Logged

Serenitis

  • Admiral
  • *****
  • Posts: 1541
    • View Profile
Re: About skill tree design
« Reply #10 on: April 24, 2021, 12:12:33 PM »

Seems similar to the 3 aptitude points that you'd spend in 0.91 but using story points instead of skill points.
The intent was spending skill points unlocks more places to put future skill points.
No story points required. (Although that might be a potentially interesting use...)
Logged

TaroEld

  • Ensign
  • *
  • Posts: 3
    • View Profile
Re: About skill tree design
« Reply #11 on: April 24, 2021, 12:24:45 PM »

This would allow the player some amount of freedom to pick whatever they want while still 'gating' the more powerful abilities.
Sounds like the system Battle Brothers has. You unlock the next tier by spending a skillpoint, which you can also spend in the previous tiers. So to unlock tier 3, you could spend one point in tier 1 and one in tier 2, or two in tier 1. Higher tier perks are generally stronger than lower tier perks, though with some exceptions. Of course, that system works on a unit-by-unit basis, so you'll pick different things depending on how you might want to style that individual unit with its pros and cons in terms of stats.
Logged

V

  • Ensign
  • *
  • Posts: 4
    • View Profile
Re: About skill tree design
« Reply #12 on: April 24, 2021, 12:44:07 PM »

Quote
What do you think the tree needs the most?

Rewrite or something. There just no point in current trees, strange logic of skills like they randomly mixed between current 4 branches.

Different short trees like
Code
1 -> 2 -> 3 -> 4
or
Code
           3a 
1 -> 2 ->  or  -> 4
           3b

and more thematic groups like "officers", "colony", "lasers", "shields", "armor", "phase", "piloted ship" and etc. And it will be easier for balance and changes.
Logged

Rudette

  • Ensign
  • *
  • Posts: 10
    • View Profile
Re: About skill tree design
« Reply #13 on: April 24, 2021, 01:46:04 PM »

Over all, I think the current system could work it just needs adjusting. The placement of the skills, some skill balance, and DP caps make a lot of skills feel bad. I do think more synergy between skills would be fun too. Inter and cross tree synergy is what makes trying out different character builds fun, adds to replay value. I like a sense of progression more than I like free form system, but this doesn't quite nail that for me.

I think something that would also go a long way is if the the first 2 to 3 tiers in a tree presented the choice of Generalist vs. Specialist or Generalist vs. Generalist, and the later skills (3, 4, 5) were Specialist vs Specialist to give you a sense of solidifying a playstyle. It would also frontload more stuff that would be generally good for every fleet, making it both less punishing if you don't spend your scant few points specializing and more interesting when you do.

Some examples of good and bad in the tree:

T1: Navigation vs. Sensors feels like a proper choice between general and specialist- general burn speed is great for everyone, being sneaky helps smugglers and the like. This choice is solid, feels good to make. Damage Control vs. Reliability Engineering is another good one. Gunnery vs. Energy, etc.

Leadership feels borked because the tree is, frankly, all over the place. DP caps make Weapon Drills feel weak (and 5-10% damage is rather boring as far as bonusses go) Auxiliary Support is great, but you don't always want it (And, thematically, arguably an industry skill) Then, L2 is a choice between two frigate specialties. A choice between two specialist skills, especially in the same field, should not be present so early in a tree.

Leadership does do one thing well, and that's build into a frigate playstyle. Auxiliary support to militarize some early game kites and then later a Venture or Colossus MKIII, Wolfpack and Crew Training for more bite and CR. But, due to Weapon Drills feeling weak and boring and L2 being all about frigates? Everyone else is less enthusiastic to climb the tree for the officer skills. Fixing Weapon Drills and Moving Wolfpack tactics or coordinated maneuvers to provide another choice would fix the tree.

Industry feels awkward and has a lot of problems because the skill placement a lot of the time is logistics skill vs logistics skill. But, something it does do right that feels good, is synergy between it's skills. Salvaging > Damage Control (on officers too) > Containment Procedures > Field Repairs. You lose less crew, you use less supplies when you fight, Salvaging and Containment Procedures synergize to top off your fuel and you can burn that fuel in excess to chain down fleets. Your fleet can sustain itself for a pretty long time this way.

On the other hand, Industry feels awkward because a lot of the time you're choosing logistics vs. logistics instead of say being able to specialize without wrapping the tree. And that's due to skill placement. The first choice in the tree is kind of sour because it's like asking you "Hey, do you want more loot or do you want to actually be able to carry that loot?" and the +1 burn you get back with militarized subsystems and expanded cargo/auxilarytanks/berthing combo. (I think the +1 burn would feel better rolled into Auxiliary Support to round out that playstyle). The DP caps hamstring the skills longterm value, relegating them to "do you want more salvage gantries, or do you want more logistics ships in your roster?" Which I suppose is a certain value.

Though, I think Auxilary Support (making junk civilian ships work in combat) has more synergy with the industry tree, Bulk Transport and Derelict Contingent in particular, than it does leadership. Both mechanically and thematically.

The colony skills do make Leadership and Industry feel more chaotic. Moving their big gameplay changing capstone abilities down to t4 instead of t5 like everyone else.
Logged

sector_terror

  • Commander
  • ***
  • Posts: 201
    • View Profile
Re: About skill tree design
« Reply #14 on: April 24, 2021, 06:15:09 PM »

I'm not susprised that the DP limits got the highest votes right now. I mean, I would love the fight bay cap to run to around 30 or so myself(5 astrals) but the desire to non-liniarity surprised me.
Logged
Pages: [1] 2 3 ... 6