More weapon balance stats

[george moromisato]

I've created a spreadsheet (with TransData) which attempts to quantify every weapon in vanilla+CC:

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

For each weapon I've defined a "balance" value. If balance is positive, the weapon is overpowered (for its level). If negative, the weapon is underpowered. The magnitude is represented in percentage points over the standard damage amount. For example, if a weapon has a balance = 10, it means it is overpowered by the equivalent of 10% damage hp over standard (thus decreasing damage by 10% would bring it into balance).

To calculate balance I translate various weapon properties in terms of balance percentage. For example, omni-capability counts as 50% extra damage (or 50 balance points). The spreadsheet breaks down each component (comments at the top of each column describe it).

A few comments on the results:

1. The changes in Ministry (https://ministry.kronosaur.com/record.hexm?id=30518) have already been applied to get these stats. If you still think a weapon is over/under powered based on this spreadsheet, please comment.

2. In general if a weapon is over/under powered by 50% or less, then we can compensate by price or other soft metrics. For greater deviations, we probably need more radical changes.

3. The computation is very subjective. For example, is omni really worth 50% damage? Would you trade 50% damage to get omni? Would you give up omni for 50% more damage? You can look at each column to see how much each feature is valued. If you disagree with a valuation, please mention it in the comments and we can discuss and adjust as necessary.

4. Fragmentation and shockwave weapons are most difficult to balance because a great deal depends on what percent of its entire damage is expected to actually hit the target. Right now, it's set to 12.5%. If we were to pick a lower number, then we'd have to increase damage to be balanced. For higher values, we would have to decrease damage. Again, this is subjective, so I'm interested in opinions (particularly those based on game play).

5. In general, this is meant to be one more data point--not a definitive guide to weapon balance. If your intuition disagrees with the numbers, please share in the discussion so we can adjust.

Thanks for all your comments!

[Xephyr]

I see CC was included in this - are you considering creating a table for EP as well?

[george moromisato]

I see CC was included in this - are you considering creating a table for EP as well?

Yeah, I will once we work out the bugs in the algorithm. [p.s.: Once I release 1.7 Alpha 1 you'll be able to generate these yourself with TransData.]

[Song]

Definitely a few weapons that stand out as having things that can't be accounted for very well. Here's ones I think can be discounted from consideration:

-NAMI Mine Launcher. It's very, very hard to use and also quite hard to find ammo for it. If the CRM500 becomes more common it might be overpowered.

-Dual Mark III: Crafted weapon requiring two at-level weapons *and* an enhancement device....thus a very high cost (and no resale value). Dual slot weapon. Tricky to use effectively because of high recoil and low ROF. Also suffers from blast damage becoming obsolete very rapidly after the point in the game where you're most likely to make it. Personally I consider it the best-designed crafted weapon in SOTP.

-Blinder cannon: Is literally useless for any practical purpose.

-Iocrym Fracture Cannon: Very short range (in the vicinity of 10ls for full damage, IIRC).

-DM600: Very limited, non-renewable ammunition. Disposable launchers are terrible in general but this one is especially not-useful.

-Advanced Urak Mass Driver: Inaccuracy of projectile means it's not a very reliable weapon for doing damage. I don't think this can be modelled accurately though (as reduced effective damage is dependent on range)

-TK10 Trident: Also a very short effective range. It is, however, probably one of the best autosentinel guns in the game.

-Mining laser: Needs the extra damage to get past damage thresholds. Short range.

-Ion disruptor: Tends to knock out shields and reactors, which no other gun can achieve.

-XM300 Reaper: Radiation effect is a one-hit-kill on AI right now. If that gets changed, this will need a buff.

-Thera Cannon: May actually be underpowered, but personally I consider this thing to be extremely dangerous.

-Ares Warhammer & Micronuke: See above. Their shockwaves are exceptional and the Ares launcher is a very popular launcher for players because it performs quite well.

-------------------

All in all there's probably some variable tweaks that could be useful, but this is definitely going to be a useful tool. Might be worth making a template (when all this is done) so that modders can use it for balancing their own gear against vanilla.

[NMS]

I'd say you're dramatically undervaluing WMD, based on how it currently works. Also, the penalty for weapons with particularly low shot speeds (Moskva-11, MAGs, Iocrym Fracture cannon) is not large enough.

And a few weapons' damage stats are still funny:
- The Ares lightning turret uses its sustained fire rate (2) rather than its burst rate (15), but the advanced tritium cannon uses its burst rate (10), rather than sustained (2.5).
- The TM7 assumes most of the fragments miss, which is not a good assumption for tracking fragments.

-Thera Cannon: May actually be underpowered, but personally I consider this thing to be extremely dangerous.

The spreadsheet doesn't say it's underpowered. It says it's powerful for its level, but not very good for the player because of how expensive it and its ammo are. (Of course, this ignores the fact that you can't purchase them.)

[PM]

- The TM7 assumes most of the fragments miss, which is not a good assumption for tracking fragments.

The tracking for TM7 is poor. What makes it deadly is within a certain range, all fragments auto-aim at targets. For example, one target within multitarget range, and all three fragments get the same velocity and seem to overlap and effectively merge into one fat missile with the power of three. When player can maintain optimal range, he bombards a target with all three fragments.

Same thing happens with S3 Medusa for NAMI heavy launcher, though it has better tracking.

[PM]

Does "balance" mean many of the seemingly baseline weapons, such as dual turbolaser, X-ray laser, lancer, and various howitzers are highly overpowered, and need their DPS cut?

Do the damage numbers account for damage type and expected armor resistances?

Does the algorithm use only the stats in the given item type? (For example, mod weapons that use original stats mostly for display purposes, then use onFireWeapon event to fire other missiles for heavily custom effects that bear no resemblance to the attacks the game expects.)

[JohnBWatson]

A question: How is balammo determined? It fluctuates quite a lot.

---

3. The computation is very subjective. For example, is omni really worth 50% damage? Would you trade 50% damage to get omni? Would you give up omni for 50% more damage? You can look at each column to see how much each feature is valued. If you disagree with a valuation, please mention it in the comments and we can discuss and adjust as necessary.

I think omni is something of a special case, as not only can you score many more hits in conventional combat, but also fight in ways that allow for much greater advantages, like dodging while attacking. A range decrease, for some omni weapons, might be a better balancing factor than a damage decrease, and would also provide a clearer niche for "middle of the road" weapons that are neither long ranged bombardment guns or omni turrets. It would also make using an omni weapon on a hardened target more exciting, as at present the strategy is to wear the enemy down over a long duration of time, in which there isn't much threat from the target(creating a de facto high defense low offense set of enemies, which you rightly try to avoid).

The Moskva turret, for all its criticisms in previous versions, definitely felt like a balanced gun when I used it recently. I could hit an enemy from any angle, and could dodge while attacking, but I had to get a lot closer to do so.

---

Iocrym Fracture Cannon: Very short range (in the vicinity of 10ls for full damage, IIRC).

I don't know about that. While this makes it weak in the hands of a (slow, unmaneuverable) AI, it's much more powerful than weapons of similar difficulty to acquire in the hands of the player.

Very limited, non-renewable ammunition. Disposable launchers are terrible in general but this one is especially not-useful.

Agree. I will once again suggest making them slotless.

Advanced Urak Mass Driver: Inaccuracy of projectile means it's not a very reliable weapon for doing damage. I don't think this can be modelled accurately though (as reduced effective damage is dependent on range)

Definitely. Right now it's a point defense weapon for the Urak Destroyer, and even there it serves a similar function to the Iocrym repeller. I'd support a complete overhaul.

May actually be underpowered, but personally I consider this thing to be extremely dangerous.

This is a common situation for high damagetype weapons. They hit the resistance reliant player very hard, but against AI enemies which rely more on numbers and compartments, they are less powerful. The rebalance of WMD should fix a lot of this, but not all.

---

Another note: Weapons which have high DPS but low firerate are best against heavier targets. If they lack WMD, they are often left without a niche. This appears to be the situation for the heavy ion blaster.

[NMS]

- The TM7 assumes most of the fragments miss, which is not a good assumption for tracking fragments.

The tracking for TM7 is poor. What makes it deadly is within a certain range, all fragments auto-aim at targets. For example, one target within multitarget range, and all three fragments get the same velocity and seem to overlap and effectively merge into one fat missile with the power of three. When player can maintain optimal range, he bombards a target with all three fragments.

Same thing happens with S3 Medusa for NAMI heavy launcher, though it has better tracking.

Even beyond optimal range, I find that generally at least 2 of the 3 TM7 fragments will hit, out to a reasonable range (assuming you have a target and fire roughly towards it). The third may start out in the wrong direction and be unable to turn around in time. Even if you assume 2 hits per shot, that's 5 1/3 times as good as the damage calculation algorithm is assuming.

[PM]

Re: Iocrym Fracture Cannon
It was overpowered pre-1.1 when Iocrym weapons could be powered-up to +150% by multiple longzhu spheres, and other plasma weapons were weaker (e.g., EI plasma cannon only had 156 DPS, instead of 217 DPS). Without the boost, it is powerful, but not always enough to offset the pain of using it, due to slow speed and short range. Today, fracture cannon remains unchanged, but other plasma weapons crept up in power, and longzhu spheres no longer work on Iocrym plasma weapons. Fracture Cannon only has 25% more DPS than current EI plasma cannon. (Before, it was 80%, or more if enhanced.) Today, I would use Fracture Cannon only if it is the only endgame weapon I have, but even then I can go to a Teraton fabricator and transmute it into something better, or at least easier to use.

[JohnBWatson]

Does "balance" mean many of the seemingly baseline weapons, such as dual turbolaser, X-ray laser, lancer, and various howitzers are highly overpowered, and need their DPS cut?

Dual turbolaser shows up a while after its level would suggest, so I'd say it's reasonable based on that. It also uses quite a bit of power compared to the alternatives.

X-ray is incredibly rare, and quite expensive. I'd say that justifies its power enough.

The others, I'd definitely call overpowered.

---

It was overpowered pre-1.1 when Iocrym weapons could be powered-up to +150% by multiple longzhu spheres, and other plasma weapons were weaker (e.g., EI plasma cannon only had 156 DPS, instead of 217 DPS). Without the boost, it is powerful, but not always enough to offset the pain of using it, due to slow speed and short range. Today, fracture cannon remains unchanged, but other plasma weapons crept up in power, and longzhu spheres no longer work on Iocrym plasma weapons. Fracture Cannon only has 25% more DPS than current EI plasma cannon. (Before, it was 80%, or more if enhanced.) Today, I would use Fracture Cannon only if it is the only endgame weapon I have, but even then I can go to a Teraton fabricator and transmute it into something better, or at least easier to use.

Fair enough, I'm probably influenced by the roundoff for compartments benefitting it greatly and the previously available ability to enhance it with an optimizer.

On the subject of the plasma cannon, its DPS is surprisingly decent. If its shots weren't so easy to shoot down it would likely be a lot more viable.

---

With regard to burst weapons, I believe we should take role into account. The Ares Lightning Turret, designed to kill smaller targets, rarely needs more than one burst to kill its target. Thus, the cooldown isn't as much of a penalty to it. The Advanced Tritium Cannon, on the other hand, is used against heavier targets, and the burst cooldown is much closer in terms of balance to a decrease in DPS(of course, this is not exact, as cooldown time can be spent dodging rather than lining up a shot.)

[george moromisato]

Does "balance" mean many of the seemingly baseline weapons, such as dual turbolaser, X-ray laser, lancer, and various howitzers are highly overpowered, and need their DPS cut?

I don't know. For X-ray laser in particular, maybe the algorithm is not discounting laser damage enough (at level 5, when armor has laser resistance). The balDamageType column shows -65, which means we expect DPS to be 65% higher to compensate for laser resistance. But at level 5, armor has 25% damage adjustment, which means DPS needs to be 400% higher to compensate.

I'm hoping people's game-play intuition will help refine the algorithm.

Do the damage numbers account for damage type and expected armor resistances?

Yes, to damage type. The current discounts are based on the difference between the weapon level and when the damage type first appears. For example:

particle/blast first appears at level 4. Therefore, the balance adjustments are as follows:

particle/blast @ level 1 (-3 levels): +15% (i.e., the weapon is considered 15% overpowered, all other things being equal)
particle/blast @ level 2 (-2 levels): +10%
particle/blast @ level 3 (-1 level): +5%
particle/blast @ level 4: +0
particle/blast @ level 5: (+1 level): -10% (i.e., the weapon should have +10% DPS to compensate).
particle/blast @ level 6: (+2 levels): -25%
particle/blast @ level 7: (+3 levels): -40%
particle/blast @ level 8: (+4 levels): -65%

It's the same for other damage types, but at a different level. E.g., laser/kinetic is level 1, ion/thermo is level 7, etc.

Does the algorithm use only the stats in the given item type? (For example, mod weapons that use original stats mostly for display purposes, then use onFireWeapon event to fire other missiles for heavily custom effects that bear no resemblance to the attacks the game expects.)

Yes, unfortunately, we only rely on stats.

[pixelfck]

@George, May I ask why we 'need' the inherent damage type resistance curves on armor? If you ask me, the level dependent armour resistance makes the whole weapon balance challenge needlessly complicated.

If the armour would, with increasing level, 'only' increase in expected hp and not have an automatic* damge type resistance increase, the whole weapon balance issue would get a lot less complicated, while the end result would stay more or less the same: you will still need higher level weapons to successfully beat higher level armor.

*: I'm not arguing against per-armor type damge type resistance.

Sure, high level armor would not be immune to laser hits. But it would take a very long time to munch through a level 8 armor plate.

(If you really wanted some sort of immunity for high level armor against low level weapons, you could probably come up with other mechanics, such as level based minimal damage threshold).

~Pixelfck

[Song]

@George, May I ask why we 'need' the inherent damage type resistance curves on armor? If you ask me, the level dependent armour resistance makes the whole weapon balance challenge needlessly complicated.

If the armour would, with increasing level, 'only' increase in expected hp and not have an automatic* damge type resistance increase, the whole weapon balance issue would get a lot less complicated, while the end result would stay more or less the same: you will still need higher level weapons to successfully beat higher level armor.

*: I'm not arguing against per-armor type damge type resistance.

Sure, high level armor would not be immune to laser hits. But it would take a very long time to munch through a level 8 armor plate.

(If you really wanted some sort of immunity for high level armor against low level weapons, you could probably come up with other mechanics, such as level based minimal damage threshold).

~Pixelfck

While I'm on record many times as really not liking the damage resistance curve (because I still think it hampers design rather than enhancing it, and makes things stupidly complex for people learning to play who don't know about it), I think that moving away from it would be a huge change, and take plenty of releases to balance by itself. It'd basically require a revamp of every armor in the game, and possibly several enemies as well.....the upside would be things like bringing back the old vulnerability to laser/kinetic damage that Hexphase used to have (which was awesome and made sense for an optimised armor), and having a far more intuitive system for people to learn....but at the same time, it would also cause serious problems if handled badly and would be a very massive shift for the game.

(It would also still require a virtual balance curve to be calculated to see how Gun X fared against the average resistances for that level)

.....I think that Armor balance in general also needs a serious look at some point (in the same way that weapons have been a big focus for 1.7), but one thing at a time for now. Accounting for the existing curve is pretty much essential for setting up the guns in the game.

[gunship256]

3. The computation is very subjective. For example, is omni really worth 50% damage? Would you trade 50% damage to get omni? Would you give up omni for 50% more damage? You can look at each column to see how much each feature is valued. If you disagree with a valuation, please mention it in the comments and we can discuss and adjust as necessary.

4. Fragmentation and shockwave weapons are most difficult to balance because a great deal depends on what percent of its entire damage is expected to actually hit the target. Right now, it's set to 12.5%. If we were to pick a lower number, then we'd have to increase damage to be balanced. For higher values, we would have to decrease damage. Again, this is subjective, so I'm interested in opinions (particularly those based on game play).

At least for me, the value of going omni depends a lot on characteristics that are hard to quantify. A highly accurate single-barreled weapon like the light ion blaster is worth at least 2.5 times as much for me if it becomes omni, since it's hard to hit small ships with it, and its lack of WMD makes it nearly useless against larger targets.

Weapons with a rapid fire rate and some shot inaccuracy gain less from being omni due to their improved ability to hit small gunships, block missile fire, and damage compartments. This is particularly true of wide-angle cloud weapons like the PK Morning Star and the actinide waste cannon, although the latter is nearly worthless because stations won't install it until midgame. The level 3 Urak mass driver and dual Moskva are also pretty good as dual weapons with a relatively high inaccuracy. Ditto for the Moskva repeater. The advanced Urak mass driver isn't as good because better weapons are available by the time I can actually loot one.

Fragmenting and radius weapons gain the least from being omni, particularly when they're used against swarms. A fast ship can run away, turn around, and blast all the gunships that are chasing it at the same time. In that scenario, more than 1/8 of the fragments are going to hit.

My subjective feel is that I normally get about 1/5 of the fragments to hit, on average, when I fire a fragmenting weapon in various scenarios. That includes the fragments that block missiles, and it includes the times when I fire Starburst missiles. (EDIT: That doesn't mean that I think fragmenting weapons should have 5 times their displayed DPS. I don't get 100% of my shots to hit when I use dual or spread weapons, either.)

My operating assumption is that omni weapons are 2.5 times as powerful as equivalent fixed-angle weapons, but a couple of people have suggested to me that omni is worth more:

https://forums.kronosaur.com/viewtopic.php?f=25&t=7518

The difference may be that I prefer cloud screens and fragmenting weapons, which benefit less from swivel; players that use lasers with high accuracy and long-ranged howitzers probably benefit more.

EDIT:
Howitzers show up in the spreadsheet as overpowered, but it's possible that this is actually helpful because of the way the late game is balanced. Early on, I'm fighting a pirate frigate with level 4 armor and weapons, and I have level 4 armor and weapons, if not better. In the midgame, Earth Slavers have level 6 armor and weapons, and I'm fighting back with level 6 or 7 armor and weapons.

In the late game, though, Ares ships and turrets have level 8-10 equipment, and I'm usually stuck with level 8 armor and weapons at around Point Juno. If I get too close to Ares bases, I have to face a large number of enemies with advanced equipment. It's hard to take out the bases with missles unless I'm far enough away that the missiles won't get shot down. But if I'm that far away, why not just use a howitzer?

It'd be helpful to have greater availability and variety of high-level weapons going into Point Juno so I don't have to switch to a howitzer-dominated play style just to make it through the final leg of the game. In particular, CSC's could stock a greater variety of missiles, and it might be helpful to have level 9 and 10 weapons with shot inaccuracy, radius, and fragmentation effects.