The elusive "Common Sense" component

this has come up before, you could search the forum. I think there's an "overkill" setting that could be tweeked. or put different reload rates. You could mod in different ranges/reload rates for slightly diffent versions of the same weapon.

Oh I just had an idea... Use mounts to change the range, by just a little. then use the same weapon with a couple diffeent mounts.

SEV, more than a feeling.

Which version are you using.
1.44 with BM 1.09 my space bases Multifire at a range of targets.
Also check what strategy you have assigned the ship, and if it has a 'Damage % targetted' in the list of priorities, if not or if it's low on the list, add it in at number 1 instead.

Nevyn -- I like the sound of this strategy tweak, it seems like a natural for a PD ship, but can you explain in more detail? I gotta admit, I'm just not sure what moving the "damage % targeted" setting in the priorities list would do exactly.

The total targeted damage % should compare how much damage the target can take to how much damage the weapons targeting it can inflict.

Setting this number lower will make the weapons reach the % setting using less weapons. Giving it a higher priority will make it "kick in" more often.

For example, if closest (absolute) were higher priority than ttd

So you're saying that you can avoid total overkill in the ships strategies? Just making sure I'm following you right. Would that not still only target one ship/fighter to fire all weapons at? I don't know from experience, so I'm just asking some questions.

I don't claim you can avoid it altogether, but it can be reduced. You'd have to set ttd to the lowest setting and give it top priority if you wanted to avoid all overkill, and that would make your weapons reluctant to kill unless no other targets were present.

You'd still get some overkill due to weapons on other ships choosing the same target, and this would make seekers choose different targets, reducing their chance to overload PD's.

Strats need to be balanced and matched to designs. There is no single perfect strat for all occasions.

Since seekers do greater damage per shot, and need numbers to overcome PD's, their ttd needs to be higher than that of direct-fire weapons.

I suppose one could pause combat and manually target the weapons. I've never bothered, and there may be bugs which would interfere.

Whatever you choose, if your results are different than what I said, please let me know. Strats are the main focus of my play. The effect of multiple ships in the same fleet or TF is iffy. It's hard to monitor this in combat, and my memory is unreliable at times.

Edit: I usually use mounts on my designs, and PD's with differing ranges; so the problem is reduced for me even if my strats were crap.

Edit again: there's another issue. When designs use direct-fire weapons which are capable of firing at fighters, one must decide whether or not they should even bother. Setting these weapons to "Do not Engage" vs. fighters will save ammo/supplies, and their chance of hitting is low. Of course they do more damage than PD's, so it's a trade-off. It really depends on how well your PD's are able to cope with the enemy. Do they need help or not?

Thanks for all the information Dvoongar, I need to find more time, to play more games, to fine tune what my ships are actually doing in combat. May your empire be as the sands on the beach.

If I could program, i'd submit code for a multi targetting system. However I can't

The theory involves something like a sonar ping up to a max range variable. The max range is set beyond the max range of the weapon. This allows retargetting before the guns fire.

Then the number of ping replies (ie, those in range) are then divided by the number of guns available to fire.
A threshold is set before battle, defaulting to 1 (ie, 1 gun per target)
So as soon as 1 or more targets get in range, those guns that break that rule are forced to retarget something else until, only doubling up when there are no more targets within ping range.

Your lord and master (below Foamy) LordHavoc

You're on a possible right track LordHavoc, but your routine as-is would end up spreading fire across too many targets all too often, not significantly damaging any of them.

What *should* be happening already, but apparently isn't, is that a ship uses the targeting priority orders to pick its first target (the Nearest, Largest, Ship Type, etc. order settings), and assigns available weapons that can affect it and range on it until either it has assigned everything it has, or it has reached the Total Targeted Damage value set by the orders (default is usually 150% to allow for misses, point defense, etc.; individual orders should adjust this).

It should then mask out any possible targets that it no longer has weapons that can affect (e.g. if it only has missile launchers left unallocated, ignore fighters, sats, etc.), and pick the next most priority target, and repeat the allocation process.

All of the above is supposed to be in the game already, and is supported by the orders definitions, even if the user interface leaves a lot to be desired; and if it worked properly would go a long way toward making hands-off combat more sensible. For instance, someone ran a test a while back where they found that large numbers of identical weapons were not properly using past the first several of them, as apparently the check for whether there were weapons whose recharge cycle had completed and were ready to fire couldn't "see" past the first few. (So, it would fire the first several, and then wait for them to recharge, rather than firing the remaining unfired weapons much sooner.) The Total Targeted Damage setting seems to be quite broken, if it's implemented at all; and so on.

Once it is all working, one of the (hopefully) easiest improvements to allow better control would be to increase the number of Targeting Priority slots in an order; I often find that 5 is not sufficient to describe what I want the ship to do properly. Making it an (mostly) arbitrary number would be great; e.g. adding a "Number of Targeting Priority Settings := ##" field and parsing it similarly to the Target Type settings, for instance; if that's not practical for some reason even a few more slots would help.

A more advanced targeting system improvement, to closer approximate clever human target allocation, would try to not always fire the best weapons at the main targets first, but to allocate more holistically. In general, build a list of targets that can be reached and affected by any of your weapons, and assign fire such as to generate the most number of highest priority targets served with Total Targeted Damage.

For example, let's say you have a mixed beams and missiles ship, facing a higher-priority enemy ship close in, and a lower-priority enemy ship far enough away to be outside of your beam range (but still inside your missile range). If your missiles are generally considered "better" weapons (more damage individually, for instance), in a simple strategy they might all get assigned against the closer ship, with a few beam weapons to finish it off; leaving most of your beam weapons unable to be used against anything. A better strategy might be to assign the missiles to the further (although lower priority) target, and full beams to the closer one.

Accomplishing the above would be tricky; my guess is that a hybrid implementation that estimates the results of a few semi-predefined targeting arrangements and assigns a score to each, and then actually carries out the one with the highest theoretical score. The simplest take on this that might generate usefully better results might be to generate starting from the shortest-ranged weapons to the longest, and then generate from the longest-ranged weapons to the shortest; and then do a weighted sum of the priority target scores that have been serviced with enough theoretical damage, and pick the one that scores higher.

Careful use of things like "Has Undamaged Weapons" and related order components would be needed; a ship at 50% overall damage might be a limping wreck or a deadly threat. For instance, let's say you're faced with 3 identical, fresh ships; the total damage needed to destroy one completely (from undamaged with full shields & armor to explosion graphic) we'll call 100%. If your ship can dish out 100% damage in some period of time, splitting it evenly (33% each) is a really bad idea; a ship at 66% remaining is quite possibly still fully or mostly functional, having merely lost shields and armor. On the other hand, assigning all of it to one ship (using the typical SE:V default of 150%) is also inefficient; you end up overkilling one ship and leaving two completely untouched. On the other hand, if you've got the ability to do 200%, spreading it evenly (66% each) might actually give you the best results, as a ship at 33% might be damaged enough that it's not a serious threat anymore.

What you'd really want in this situation is to pick one ship and fire at it until it has lost the ability to significantly hurt you, then switch to the next ship. All major weapons gone is a rough approximation, but for instance if you're fighting at missile ranges and have a reasonable expectation of being able to hold the range, destroying just the target's missile weapons would be the point at which it's better to switch fire to the next ship.

Another consideration currently lacking is some idea of the "desperation" of the fight; there are times when you want mostly want to move away without taking any damage, and any damage you do to the enemy for "free" is nice but not essential; there are times when you want to do as much damage to the enemy as you can while retaining the ability to retreat, regroup and/or repair in good order; and there are times when it is a desperate battle in which the idea is to do as much damage to the enemy as possible without regards to the cost.

Similarly, there are times when supplies and/or ordnance are not a concern, and times when they are. If a long-distance ship is operating on the edge of known space mostly shooting up unarmed or much lighter ships, I'd like it to avoid using up ordnance for any fight it can win cleanly without those weapons, to save the missiles for priority targets or serious fights. These aren't really orders, as they're more situational; perhaps what is needed is multiple attack commands... "Attack frugally" (minimal use of ordnance, lower supply usage, try to avoid damage if possible, accepting doing less / slower damage to the enemy), "Attack normally" (the default, probably a bit more conservative than the current... moderate use of ordnance, don't worry about supply usage, try to do significant damage to the enemy but don't be suicidal), "Attack desperately" (no holds barred, expect to die like moths in a flame, but try to do as much damage to the enemy as you can before you go down; if ramming will do more damage than weapons fire do it).

One (comparatively) simple way to interface the above with the existing system would simply to allow a ship (or fleet) to have three orders strategies instead of the current one, each linked to one of the above attack commands. A better way would be for the computer to be sophisticated enough to meaningfully manipulate a single order set to accomplish the above intentions; I'd not expect that anytime soon

Between years of play in earlier versions of Space Empires, Master of Orion, and the like, many humans have gotten quite good at doing the above sorts of target juggling in their head. We have literally years (decades for some of us of practice... teaching a computer to perform even nearly as well is remarkably tricky. Given a few more years of More's Law buffing the hardware, tricks like genetic algorithm - driven fuzzy optimization seeking show considerable promise; some of the stuff I deal with in my day job could probably do about as well as a human on this with the right training (which in itself would take several programmer-months), except that on today's hardware it would take several minutes of thinking per second of combat.

Outstanding points, jdunson. Another tool I'd like would be to differentiate between targeting for maneuvering purposes and targeting weapons. For example, missile ships are usually set to maintain max range vs. 'closest (absolute)'. This is to keep them from getting hurt. But I'd like to have them be capable of shooting ships other than the closest ship.

Other than that, I avoid use of the 'closest (absolute)' because the odds of a tie are so small that it effectively nullifies any subsequent priorities one may assign.

We need more tools, and we need the 'available' tools to work as advertised. And for our part, we must utilize all available tools to the best of our ability.

My impression of ttd has been that it works unless placed after 'closest (absolute)' or something else that might prevent it from functioning. I'll watch this closely in the near future. I shouldn't be too surprised if I'm mistaken about it.

With proper tools and clever implementation, it is possible to emulate a good amount of common sense. At present, the greatest horror related to this issue is the stock game's exclusive use of 'optimum range' for all AI ship designs.

Another thing that targeting %hull points doesn't take into account is the varying amounts of shields and armor a ship can have. An 1800 point baseship can have 2000 points of shields and 500 points of armor. Thats 2500 points to cut through before doing any hull damage. Counting misses and/or missiles shot down, 150% might not be enough for larger ships.

I've verified ttd% does work in stock 1.44. Had a frigate with 2 CSM's split fire between a weak moving enemy and one that couldn't move. Killed the sitting duck and did it again with a 3rd ship!

And glockgemini, I think it counts all the ships points, not just 'hull'. Underkill hasn't been a problem very much from what I've seen and what others have said. But you can crank ttd clear up to 200% if you need to.