[ROTP] What are your ship design tactics?

[RefSteel]

Updated the Gunner algorithm above with hopefully-improved defensive numbers as well.

[Ray F]

Sweet, thank you! Once I am finished with the diplomacy model, I am going to work on the AI. At that point, I'll start integrating your design algorithms into the game.

Below is the latest project timeline. Please note that most of the development tasks after the Alpha are really stretched out since the long pole in the tent is all of the artwork remaining.

http://www.java-moo.com/wp-content/uploa...meline.jpg

[RefSteel]

Cool - the timetable looks good (and realistic based on your progress so far I think, so that's especially great).

I 'll probably write up a bomber next, since in theory it should be relatively simple, but some complexity is introduced when attacking a competent general + tactician (whether player or AI) that I'll want to think about some more. After that, I'll hopefully take a stab at support ships, which might force me to finally think through some of the complex interactions of special systems, as alluded to under anchors. As it is, there are probably situations where anchors should mount stream projectors or pulsars in lieu of some or all of their weapons, especially when they would otherwise have a Special slot open.

[Ray F]

Just wanted to add that the ROTP "gunner" ship design algorithm has now been extensively modified to incorporate the same elements that RefSteel's algorithm uses. For programming reasons it's not the same algorithm, but it's very close and takes all of the same factors into consideration.

In summary:
-- design size is based upon the mininum of racial preference, and largest constructable by most developed colony (up to Large). I chose this approach because I wanted there to be a more pronounced racial preference for ship size
-- obviously, fastest engines and manv=2 where possible. large designs get best computers & scanners
-- an enemy target template is created for each hostile race... if no hostiles, then each contacted race... if no contact, then player is used
-- enemy targets currently track shield level, have repulsor? and have interdictor?
-- for each target, the max possible damage is calculated for each available weapon. streaming, shield reducers, missiles & torps are all adjusted for. Any weaponized design that cannot damage any of the targets is discarded.
-- heavy weapons, beam focus, teleporters, cloaks, missiles are considered against repulsor designs
-- if no design can damage all of the enemy targets, retry at a larger size (max is Large)
-- once a weapon is selected we try to improve the battle computer where +1 attack = +15% dmg
-- after computers we try to add maneuver specials (teleporter, inertials) where +1 combat speed = +10% dmg
-- if no teleporter, then try to add beam range specials (only 1) where +1 beam range = +5% dmg
-- if no teleporter, then try to upgrade ship maneuverability where +1 maneuver class = +5% dmg

Obviously, the whole thing with non-weapon components is trying to come up with some relative equivalence where you can estimate "this much +attack or +maneuverability or +range is equivalent of this much damage inflicted".

It's all very rough but it doesn't have to be an exact science, and I referenced tables in the OSG where they were appropriate. Since there is no optimal formula, the ultimate point is to make competitive designs and to ensure all of the ship components find a way to get used.

I think RefSteel's algorithm has gone a long way towards focusing my thinking on how to do that. Thanks tremendously! It was a huge help!

[RefSteel]

I'm glad I could help! Since I seem to have no time to do a full write-up for bombers, I'll just note for now that combat speed is hugely important for them (especially the jumps to 2, 3, and 4 spaces per round) - more so than even for Gunners - unless the planned opponent has miserable weapons and excellent ECM/Shields.

[Ray F]

You have been a tremendous help! I will be starting on your Destroyer design soon. I expect the bombers to be relatively easy since, as you said, they are more one-dimensional.

But shouldn't they stack missile defense?

[RefSteel]

But shouldn't they stack missile defense?

In terms of my designs against the original MoO AI, the answer is, "Almost never," and I could practically drop the "almost." The ideal bomber fleet reaches its target and destroys all the bases (usually in two volleys) without making contact with an enemy missile at all, and it's commonly possible to get quite close to, or even actually achieve, this ideal.

In combat against a competent tactical opponent, missile defense might be worthwhile, as repulsor ships (especially in multiple stacks) with good goal-tending tactics can delay the bombers significantly, or even force the bombers to wait for e.g. heavy gunners to clear the repulsors out. Even so, missile defense should not generally be a go-to option for bombers, as it interferes with maximizing damage output and tactical speed.

(Note for instance that a 25 BC bomber with good missile defenses has only 80% of the effective damage output per BC invested in bombers of an otherwise-identical 20 BC bomber with no missile defenses. As such, the former is only worthwhile if the ships are expected to receive so many missile salvos that those survivors that would have died without their missile defenses ultimately deliver as much damage to enemy bases as at least 25% of the rest of the fleet.)

That said, you mentioned that you want to ensure that all ship components find a way to get used, and I find that point compelling. Although anchors (destroyers) can be equipped with missile defense for their role in covering bombers in planetary attacks, the easiest and probably best way to include missile defense systems is aboard bombers - especially larger designs. Just make sure the defenses don't get in the way of the bomber's primary job!

---

On the design notes you listed above for gunners: Flat +% to damage numbers will tend (in most circumstances) to lead Mrrshans to overvalue computers, and lead anyone with Alkari enemies to undervalue them. This might not be a huge problem, but it's something to be aware of. I also suspect that your formula (and mine) may undervalue maneuverability, but I'd have to see how battles between competent opponents actually play out to be sure. (Partly because it depends on whether it's possible, and under what circumstances, to move after firing weapons.)

[EDIT: Also, it is almost always worthwhile to give Large ships the best basic armor available, and often shields as well, even if they're intended to be gunners. Once the cost per ship starts getting up into the hundreds of BCs, reasonably cheap survivability systems tend to pay for themselves handsomely.]

[Ray F]

In combat against a competent tactical opponent, missile defense might be worthwhile, as repulsor ships (especially in multiple stacks) with good goal-tending tactics can delay the bombers significantly, or even force the bombers to wait for e.g. heavy gunners to clear the repulsors out. Even so, missile defense should not generally be a go-to option for bombers, as it interferes with maximizing damage output and tactical speed.

I always want to assume that the competent tactical opponent is the player. It sounds to me like missile defense might be a good investment.

That said, you mentioned that you want to ensure that all ship components find a way to get used, and I find that point compelling.

Yes, there should be no Hyperspace Communications freebie techs to give to the AI.

On the design notes you listed above for gunners: Flat +% to damage numbers will tend (in most circumstances) to lead Mrrshans to overvalue computers, and lead anyone with Alkari enemies to undervalue them. This might not be a huge problem, but it's something to be aware of. I also suspect that your formula (and mine) may undervalue maneuverability, but I'd have to see how battles between competent opponents actually play out to be sure. (Partly because it depends on whether it's possible, and under what circumstances, to move after firing weapons.)

For maneuverability, I was thinking that 2 Maneuver = 1 combat speed, so I halved the benefit. This method gives up some tactical accuracy while alloweing numbers to be easily tweaked. Remember, the goal is not to make the best possible AI, but to make a competent AI that players can't roll over. A good player should still always be able to beat the AI, but at the same time should always lose if he rebels against the majority council vote.

[EDIT: Also, it is almost always worthwhile to give Large ships the best basic armor available, and often shields as well, even if they're intended to be gunners. Once the cost per ship starts getting up into the hundreds of BCs, reasonably cheap survivability systems tend to pay for themselves handsomely.]

Thanks! I'll add that tweak to the never-shrinking list of things to do!

[Psillycyber]

[EDIT: Also, it is almost always worthwhile to give Large ships the best basic armor available, and often shields as well, even if they're intended to be gunners. Once the cost per ship starts getting up into the hundreds of BCs, reasonably cheap survivability systems tend to pay for themselves handsomely.]

I expect that part of the reason why this was always effective was that the player was always going up against a lot of AI-designed ships. If the majority of the AI opponent's fleet is a swarm of obsolete laser fighters, then slapping a class-IV shield onto a cruiser (or heck, even a destroyer) gives an almost magical advantage. But if the AI has upgraded its fleet to mostly fusion beams, then slapping a class-IV shield onto a ship that is primarily designed to be a gunner might not be so cost-effective. It might be better to just focus on killing the enemy more quickly.

[HansLemurson]

[EDIT: Also, it is almost always worthwhile to give Large ships the best basic armor available, and often shields as well, even if they're intended to be gunners. Once the cost per ship starts getting up into the hundreds of BCs, reasonably cheap survivability systems tend to pay for themselves handsomely.]

I expect that part of the reason why this was always effective was that the player was always going up against a lot of AI-designed ships. If the majority of the AI opponent's fleet is a swarm of obsolete laser fighters, then slapping a class-IV shield onto a cruiser (or heck, even a destroyer) gives an almost magical advantage. But if the AI has upgraded its fleet to mostly fusion beams, then slapping a class-IV shield onto a ship that is primarily designed to be a gunner might not be so cost-effective. It might be better to just focus on killing the enemy more quickly.

I made a spreadsheet once calculating the cost-effectiveness of various weapons versus shield-strengths. Even low levels of shielding are surprisingly effective against most weapons. In your example, of Class IV Shields vs. Fusion Beams, the Shields will actually block around 40% of the incoming damage from normal mounts and 25% for heavies.

The overall Combat-Power of a ship is its Durability * Firepower. If adding a shield reduces your guns to 80%, but cuts incoming damage to 70%, then you're better off with the shields.

The only place where Firepower and Durability aren't interchangeable is where the First Strike is particularly devastating. Guns increase your first-strike (assuming initiative) but shields only diminish the returning fire.