So, this whole armor thing

[Eternaly_Lost]

I am going to suggest something a bit weird here. But why not use both armor-as-percentage and Armor as a flat vaule on a few special units.

Most ships would use armor-as-percentage, and that would be it. But a few special units (Those very heavy armored ones that should shut down the lightest of weapons) get a flat value on top of their armor-as-percentage.

[Tridus]

I am a bit puzzled about the armor-as-percentage idea --- am I right to think it would be the same as extra HP except for some bonus ships with armor effects?

You too? I'm not seeing much of a difference between a ship with 50% armor and X HP, and a ship with 0% armor and 2X HP, outside of the case of stuff with 90% armor vs stuff with 100% armor piercing.

[Draco18s]

Ain't going to be in the same sector as this one.

The current system is bad, the idea at the top of this thread is also bad, so are most other options, all for a variety of reasons.

[Vitka]

Why not remove armor altogether and bake its effects into hull type/damage bonus?

[Ranakastrasz]

The thing is, even if it averages out on huge quantities of ships... what about people who use starships? What about the early game? What about a streak of bad luck where you have this decisive engagement where all of your bad rolls got stacked all at once and you miss every single shot? Luck can completely ruin you, even if it averages out quickly. I've seen it happen countless times in other games. Notably, board or pen and paper games. I'd burn all my great Pathfinder rolls on a shooting gallery, and miss every time in actual combat while the DM would score regular confirmed criticals. I've seen opponents with great figures in Heroscape, and yet they just lose because every time they attack they miss completely.

Sounds like someone needs a less random random number. Real random is a horrid Idea when it comes to video games, you need to use a custom random, which is less random than real random, and looks at previous rolls for new rolls.

~Shuffle Bag. Have a bunch of booleans in a bag, hit or miss. (20 of them?) And pull one out each attack. And when you use all 20 (5% threhholds) reset the bag and start over. Have one for each set of chances from 5% to 100%, and for each player, and whichever one applies, use that one. Will garentee that one of every 20 shots hits for a player. (Still not ideal, but you would have to attach data to EACH SHIP or at least ship Type, for the alternative)

Also, starships tend to have a lot of attacks.
~Bomber starships, and bombers, would have shield bypass ability. On account of them intended to counter hardened targets, and short range.

Hm, that's a new idea. Bring back the old shield mechanic, but interpret the resulting percentage as a multiplier to damage instead of a probability, putting determinism back into the system.

However, it still suffers from the other problem the old shield system had; hard to estimate at a glance what kinds of expected orders of magnitude of damage reduction would be getting in ordinary matchups, aka, not intuitive. (Remember, the tool-tip only helps so much, sometimes there are no enemy ships around that you can hover over, and the tool-tip won't help you with estimating how much damage you will be taking from the enemy ship, though the new reference tab in the stats screen helps somewhat).


Still, this seems like a viable new option to add to the growing list of proposals.

Yep, Hence why I had that secondary suggestion. If you cannot accept, even a less random random number, this ought still work. And the Random Factor was totally Unnessesary, since having a simple damage scale-down instead of chance to hit works just as well.

Also, While I didn't figure the shield scale on my own, it was on the wiki. However, I know it was simple to figure out closing the distance increased your chance to hit. Which made intuitive sense, as well.

The BIG question, however, is why hasn't anyone realized yet, that Radar dampening was put in to fill the void added when shielding was removed?

My Current Theory as to what the Equation used to be, is '% = (Range - Distance)/Shield' bound between 5% and 100%. At 'Range - Shield' distance, the % is 100% exactly, or at least it should be. Lightly shielded/armored craft, which Fighters should counter, have little to no shielding. Heavily shielded/armored craft have Plenty, and require specilists or lots of Dakka to take down. Bombers and similar anti-heavy should get shield bypass, and generally will be more effective on heavily armored/shielded targets. And, as the fun bonus part, Guard posts, any anything else with radar dampening become hardened shielded targets, with 20k+ Shielding, requiring either significant damage penelties, Shield bypass (Sniper) , or else getting dangerously close.

[LaughingThesaurus]

Armor as a percentage would basically be that HP multiplier that only armor piercing ships can really get around. The fact that you can get around the % damage reduction means it isn't 'just a health multiplier'. In a way it ends up being similar I guess to hull type damage multipliers, but it's not at all connected to the hull type.  I'm kind of into maybe adding flat damage reduction on top of that as something that can't be pierced at all. So, what you get is the occasional heavily armored ship that will just flat-out take less damage even if armor pierced or armor rotted. The thing is, the idea is super simple and I don't think the game remotely needs 30 complex formulas to go into if you want to actually predict the results of combat. If somebody has a very simple idea of what to do that you can analyze in-game at a glance, that doesn't involve hit chance, then that's going to be what I go with.

The thing is, I'm also just in favor of simple ideas, hit chance or no. If you have this complex series of formulas and 6 different steps that go into calculating damage, why even bother telling the player the armor value? They aren't going to be able to interpret it in any meaningful way because interpreting that data is just going to be too complicated to do. This is a game where it already takes long enough to analyze all the differences between each ship type. I'd rather not throw formulas into the mix. Even if it's a simple case of "higher number is better", that isn't really good enough. I'd rather remove armor entirely than have a new armor system that people can't actually predict without busting out the calculator for some hardcore analysis. A percentage damage reduction based on deflection/shields is okay, because then instead of a 20% chance to miss entirely, 20% of the damage is taken off while you always hit. There's not really a difference in the long run, the same amount of damage is lost, but you can't have a big cluster of complete misses anymore. That being said, I still think we need something simple that you can interpret at a glance, or flat-out no armor system at all.

[Ranakastrasz]

Armor as a percentage would basically be that HP multiplier that only armor piercing ships can really get around. The fact that you can get around the % damage reduction means it isn't 'just a health multiplier'. In a way it ends up being similar I guess to hull type damage multipliers, but it's not at all connected to the hull type.  I'm kind of into maybe adding flat damage reduction on top of that as something that can't be pierced at all. So, what you get is the occasional heavily armored ship that will just flat-out take less damage even if armor pierced or armor rotted. The thing is, the idea is super simple and I don't think the game remotely needs 30 complex formulas to go into if you want to actually predict the results of combat. If somebody has a very simple idea of what to do that you can analyze in-game at a glance, that doesn't involve hit chance, then that's going to be what I go with.

The thing is, I'm also just in favor of simple ideas, hit chance or no. If you have this complex series of formulas and 6 different steps that go into calculating damage, why even bother telling the player the armor value? They aren't going to be able to interpret it in any meaningful way because interpreting that data is just going to be too complicated to do. This is a game where it already takes long enough to analyze all the differences between each ship type. I'd rather not throw formulas into the mix. Even if it's a simple case of "higher number is better", that isn't really good enough. I'd rather remove armor entirely than have a new armor system that people can't actually predict without busting out the calculator for some hardcore analysis. A percentage damage reduction based on deflection/shields is okay, because then instead of a 20% chance to miss entirely, 20% of the damage is taken off while you always hit. There's not really a difference in the long run, the same amount of damage is lost, but you can't have a big cluster of complete misses anymore. That being said, I still think we need something simple that you can interpret at a glance, or flat-out no armor system at all.

I understand that as well. I think that removing the Armor system totally may then be the best option.
Its a single formula, either way. One Ensureing Armor doesn't become irrelivent with tier differences, the other one with a Min/max to keep the entire thing VERY simple. And the reduction effect is only in a set threshhold, where it is entirely linear. It is seems like an exaduration. I don't see 28 other equations.
Regardless, Dropping the Armor system would work, but it would likely be nearly impossible to then justify a ton of units that currently exist. I have no idea how space tanks will have a niche, without something to make them really durable in all but a handful of cases. (Give them a unique Armor type? That makes it MORE complex...)

Still. Radar Dampening Vs Old shields. Has anyone noticed? I haven't seen anyone post about it before...

[LaughingThesaurus]

For space tanks, I'd just give them 95% armor under my very simple idea, so that armor piercing ships will have a field day with them but they're super durable against anything else that only has very light or no armor piercing.
But yeah, I was exaggerating. Your idea isn't that bad, I just would rather keep formulas out of the data the player actually has to interpret to make decisions. That way, any player can at any time interpret the information on the screen very easily. This is something that grew to be my own philosophy on game design. You keep elements the player needs to understand simple, so that the player can actually make informed decisions relatively easily. Any information that is out of the player's hands can be as complex as you want it to be because the computer is the only thing that's looking at it and the player's probably not making decisions based on it. Heck, even accuracy ratings for a turn based game (if you went that way) could be calculated via very complex formulae, as long as you can just select to attack without confirming it and get the result of the equation instantaneously. In this case, I don't really see how you could without actually calculating it manually. Even if you did, it does kind of flood AI War with a ton more data, which is probably the last thing the game needs.
I'll admit though, I did flip out at your idea initially because of the idea of a hit chance, so I didn't respect it as much as I probably should have. It just struck me as too complex.

[TechSY730]

I'm not a fan of totally removing the armor system. Some way to reward high damage shots vs high rof, low damage shots is nifty (though the game is laking something that directly go the other way).

However, one possible solution is to make armor (either as it is now or in some revised form) be the exception rather than the rule; only some ships would have armor. If this is done, the few units that would get armor, and the units that counter it directly, could be made more "extreme" in their gimmick (and in the case for those that counter it, buff to give more general purpose utility as well). (Also, I'd like to see the "reverse armor" done as well, something that stops more damage from low rof, high damage than high rof, low damage, and give that to a few units, like the space tank or something)

I would personally like the system to be deterministic (complicated maybe, but deterministic).

Also, this game has "prided" itself in embrasing true randomness, instead of some "repeat pattern avoidance" semi-random generator. I really like it, because of its conceptual purity, and it keeps event probability independent of previous results.
This has had implication about what things can be randomized from a balance perspective, as that somewhat rare but still will be seen sometimes "lots of unlucky rolls back to back" can indeed break balance of certain types of situations. This was one of the reasons why randomness was brought out of any damage formula; it is too central to trust that the RNG will deliver "average-ish" results when it matters. In some cases (like map seeding), checks had to be out into place to force some amount of "rerolling" if a worse case roll was detected (though thankfully, this is rare. Usually this is done by having a fixed pool of stuff they can "draw from" when seeding, and using that RNG to pull from that pool, sometimes removing the chosen item and/or points, with carefully constructed pools such that the"worst theoretical" cases cannot happen)

[LaughingThesaurus]

The thing is, the randomness the game does use is actually randomness that you can react to. If you get a massive wave of only bombers barreling down your homeworld, well, you'd better get your things that are good against bombers ready and fight them. If you get a planet that's a massive chokepoint that you can't get by, there's the new hacking method to get around it. Even without hacking, you can probably get by it just by playing well enough and thinking of the best strategy. You can deal with that stuff. Now, what happens in those do-or-die situations if you get super unlucky with hit chances?

[Ranakastrasz]

Exactly why I suggest less random random numbers. Saying we have always used true random, hence we should always use true random, is circular reasoning, and not at all a valid reason to reject it. Unless you support and want streaks of good or bad rolls, and don't care about / enjoy them punishing players, or the effect is, as said, noted in advance, you should NOT be using True random. (or best true random you can generate)

[Tridus]

I'm not a fan of totally removing the armor system. Some way to reward high damage shots vs high rof, low damage shots is nifty (though the game is laking something that directly go the other way).

The armor system as proposed at the start of the thread doesn't reward high damage vs high rof, 25% reduction is 25% reduction no matter how many shots it comes from.

As for something that goes the other way... Protector Starships do, but the AI doesn't have that mechanic. Laser Gatlings would basically be immune to Protector Starships because they fire so many shots that most would get through (and the blocked ones would be blocking trivial DPS vs blocking a fortress attack).

[LaughingThesaurus]

I still don't see how it's necessary for one in every 20 shots to completely miss when instead you could make it so every shot does 1/20th the damage. I mean, if it were something like a meter you could see fill up that shows you when you'd miss your next shot, that's still just overcomplicating the combat.

[TechSY730]

Exactly why I suggest less random random numbers. Saying we have always used true random, hence we should always use true random, is circular reasoning, and not at all a valid reason to reject it. Unless you support and want streaks of good or bad rolls, and don't care about / enjoy them punishing players, or the effect is, as said, noted in advance, you should NOT be using True random. (or best true random you can generate)

One reason I like "true random" (or rather, independent "rolls" randomness) for when probability is listed is because if it says 20%, it means 20% chance for, in this case, each shot, regardless of what the recent shots were like. If you make it "less random" by trying to avoid "streaks", then sometimes that listed probability will be flat out wrong (it would have a much greater or much less than 20% chance if many of the last N shots were a streak of missing or hitting, respectively).

Now one way this can be avoided is to not list probability exactly. Maybe like "about 20%" or "~20%", or "not very likely", "somewhat likely", "very likely", or something like this. When you don't list (or claim to list) an exact probability for a shot to hit, you don't have to worry about any potential difference in per shot chances vs. long term expected hit rate.
Another way would be to show the "long term" probability (based on the formula, which would remain constant for a set of stats for the attacker and defender) AND the probability of the next shot (which would vary each shot based on recent history by whatever algorithm the "de-streakifier" uses)

Note, "true random" does not have to mean uniform distribution. There are other distributions out there that have independent events but do have a higher chance of returning something "closer" to the average, like Gaussian.

[LaughingThesaurus]

The problem with true random is just stuff I've mentioned though. You can have a long streak of misses at a critical moment and lose the game as a direct result of that.