Starting ship balance

Post ideas & suggestions you have pertaining to the game here.
Post Reply
george moromisato
Developer
Developer
Posts: 2957
Joined: Thu Jul 24, 2003 9:53 pm
Contact:

Tue Nov 13, 2018 7:39 pm

I've added some balance equations to TransData to calculate the value/balance of starting ships. The results are here:

https://docs.google.com/spreadsheets/d/ ... sp=sharing

NOTE:

1. This excludes starting devices/armor. I.e., we just care about the value of the hull properties (max devices, etc.).
2. hullPoints is the total value of the hull, higher is better. This is calibrated so 10 points = 1 unrestricted device slot.
3. hullPoints is the sum of all the other columns.
4. The Eternity Port stats are based on changes in 1.8 Beta 4. See: https://ministry.kronosaur.com/record.hexm?id=83042

I'd love your thoughts: is the algorithm over/under-valuing some properties? How does this correspond to your subjective feelings about each class?

User avatar
Song
Fleet Admiral
Fleet Admiral
Posts: 2733
Joined: Mon Aug 17, 2009 4:27 am

Tue Nov 13, 2018 8:44 pm

While I do like having algorithmic support for balancing ships, I think we need to resolve the disconnect between the two measures used for the mass of armor. It's two sets of very reasonable scales that don't actually match each other, and that causes some significant issues. I've run into this a lot on Elemental Shift, since I'm trying to do as much 'vanilla-like' content as possible, including a large range of playerships....and armor mass is a critical balancing property for playerships. The algorithm I assume is using the standard for playerships, but that doesn't entirely work when the ingame armor masses are reduced:

Standard armor mass for the armor items is usually about 3 tonnes, sometimes 3.5. "Heavy" armor is only about 4 tonnes, maybe 5. Super-heavy is usually 8 or 9 tonnes (if you use the "take the heavy armor and double it" standard). "Massive" armor is generally more than that, up to the 40+ tonne armor used by CSCs. There are very few armors above 10 tonnes (even modding in extra super-heavy armors, as I have...10t is about as high as you ever get). But we have light gunships with armor limits at 12 tonnes, and freighters with limits at 20 tonnes. It's arguable that setting maximum armor that high helps control the speed alterations when using the actual heavy armors....but that can be done just as effectively with lower values for both armor limits and speed reduction. Or more granular speed reduction settings, but that's the complicated solution.

I think that there's a need to alter one or both scales (either drop the max armor values on ships across the board, or increase the range of mass of armor segments, particularly at the moderate to high end). We need to get them matching, or the high armor limits don't really have an impact except to make heavy armors not give speed penalties.

(Edit: I wrote this when I'd just woken up, so it's probably not as clear as I'd like it to be, even after a lot of edits. I'll check in when I've woken up properly and edit again if necessary)
Mischievous local moderator. She/Her pronouns.

NMS
Militia Commander
Militia Commander
Posts: 475
Joined: Tue Mar 05, 2013 8:26 am

Tue Nov 13, 2018 9:59 pm

In general, I'm not sure you're taking into account that most stats have decreasing marginal utility. For instance, the 3rd and 4th unrestricted slots on a gunship are much more valuable than the 6th and 7th non-weapon slots on the EI500, though the marginal benefit of weapon slots probably decreases faster than non-weapon slots.

For these specific ships and the existing universe, I'd say you're significantly overvaluing max armor mass and undervaluing max speed. If armor hp were directly proportional to mass and enemies had a wide range of speeds, that might not be the case. But a huge fraction of enemies have 20 speed, so going from 19 to 20 or 20 to 21 is a tremendous benefit. And most armors can be used by any ship and don't confer a speed penalty or bonus to most ships, and the ones that do are not that much better or worse than the others.

george moromisato
Developer
Developer
Posts: 2957
Joined: Thu Jul 24, 2003 9:53 pm
Contact:

Tue Nov 13, 2018 10:28 pm

Song wrote:
Tue Nov 13, 2018 8:44 pm
While I do like having algorithmic support for balancing ships, I think we need to resolve the disconnect between the two measures used for the mass of armor. It's two sets of very reasonable scales that don't actually match each other, and that causes some significant issues...
This is a good point. Maybe what we need to do is keep the armor item scale and have ship classes specify the bonus/penalty at various scale points.

E.g., we would define a Sapphire as:

ultra-light: +2 speed
light: +1 speed
medium: +0 speed
heavy: -2 speed
super-heavy: N/A

This makes things a little more transparent. [Of course, we may need to adjust the armor scale to be a little more granular.]

george moromisato
Developer
Developer
Posts: 2957
Joined: Thu Jul 24, 2003 9:53 pm
Contact:

Tue Nov 13, 2018 10:33 pm

NMS wrote:
Tue Nov 13, 2018 9:59 pm
In general, I'm not sure you're taking into account that most stats have decreasing marginal utility. For instance, the 3rd and 4th unrestricted slots on a gunship are much more valuable than the 6th and 7th non-weapon slots on the EI500, though the marginal benefit of weapon slots probably decreases faster than non-weapon slots.

For these specific ships and the existing universe, I'd say you're significantly overvaluing max armor mass and undervaluing max speed. If armor hp were directly proportional to mass and enemies had a wide range of speeds, that might not be the case. But a huge fraction of enemies have 20 speed, so going from 19 to 20 or 20 to 21 is a tremendous benefit. And most armors can be used by any ship and don't confer a speed penalty or bonus to most ships, and the ones that do are not that much better or worse than the others.
Good point. I hesitate to design the balance algorithm based on the current distribution of items/ships, but the concept of decreasing marginal utility makes sense.

p.s.: You're not wrong about the speed distribution. I used TransData to generate a distribution of ship speeds (based on the number of ships actually encountered, not just ship classes). Ships with 22 speed are actually very common (at least 15 classes). We could flatten this distribution a bit, but that's a different topic.
Ship Speed Distribution.jpg
Ship Speed Distribution.jpg (52.79 KiB) Viewed 857 times

User avatar
Song
Fleet Admiral
Fleet Admiral
Posts: 2733
Joined: Mon Aug 17, 2009 4:27 am

Wed Nov 14, 2018 1:52 am

I think more speeds and turn rates is a good thing to do. I'll....admit that I don't really shift from those much (though I have changed a few values, notably for the Scorpions). Ship balance will always have some qualitative elements of course, as well as technically quantitative measures that aren't really possible to measure. Some vessels have unique gimmicks and penalties that may not be possible to track in an algorithm (though they could be documented in a set of guidelines. For example: not being able to use shields is common enough that it could have a manual modifier stated).

For example: I decided to add two ship into the game as semi-jokes. One has stayed mostly a joke, the other ended up as arguably the most powerful vessel in Elemental Shift....but I still wouldn't recommend playing it if you weren't after a challenge:

The first is the Mouse-class shuttle. It's tiny, it's terrible, and it's hilarious. While it has terrible armor (it starts with XMH of all things), the combination of a very, very small ship with reasonable speed and mediocre but adequate agility is a challenging but interesting experience. Until you find an omni weapon, at which point it becomes remarkably strong (because you're small enough to dodge and can shoot back while doing so). In this case it's quantitatively bad in most respects apart from size and possibly speed...but qualitatively, it's quite fun in the hands of a skilled player. I wouldn't even know how to start measuring that. That's fine though: the whole point of the Mouse is to have an interesting early-game experience, then ditch it at a shipyard (once you fix the "you can't buy a vanilla ship if you're flying a modded ship" bug, that is) and buy something practical.

The second is the EI100/M, which is an extremely blatant gimmick ship. It shares most of the attributes and elements of the stock EI100. That is to say, the hull is complete garbage and it's so slow I keep trying to turn on autopilot only to notice that I was already using it and just hadn't noticed. It is a truly terrible ship, starting out. But the inclusion of a powerful gimmick (the EI100/M gains an extra weapon slot and +Omni to missile lunchers) means that, with decent RNG, it becomes an unstoppable killing machine in certain circumstances. The gimmick is in my view severely overpowered in the early game and only really becomes more of a problem in the endgame when missile prices skyrocket......but the slow speed counters the early-game imbalance (possibly excessively so) because the access to decent drives is usually impossible before St. Katherine's Star (where the most powerful early-game launcher, the Burak, starts to be less effective). Thus I have two extremely unbalanced things that in some ways swap in terms of which thing is a bonus and which is a problem. In the end I removed variant-specific buffs to the 100/M, leaving the hull almost exactly like the base EI100 because the massive advantages and disadvantages conflicted in an interesting way to make a challenging but interesting vessel.


I think ultimately though, these don't prevent algorithmic balancing, but sit on top of it: get a balanced hull (through whatever mechanism ultimately ends up working), then apply gimmicks and qualitatives on top of that.
Mischievous local moderator. She/Her pronouns.

NMS
Militia Commander
Militia Commander
Posts: 475
Joined: Tue Mar 05, 2013 8:26 am

Wed Nov 14, 2018 8:12 am

Well, there are fewer 20s and more 19s, 21s, and 22s than I expected. But there's still a big cluster in the 18-22 range which contains most of the enemies that are both tough and potentially able to chase the player down. Most of the 25s, 30s, and 35s are Corsairs, Wind Slavers, and Hornets, respectively. But some capital ships, including the Phobos, are at 16, and armor may adjust your speed up or down a point or two, so I guess there isn't really a single key speed you need. But it does look like there would be decreasing marginal utility above 23 or so, in terms of chasing or escaping from enemies. That's on top of travel time being proportional to the reciprocal of speed.

Post Reply