FreeOrion

Dilvish wrote:While folks continue exploring those ideas, I'd propose that we take an easy step to deal with the main problem the OP cited, being comsat abuse, by taking at least part of that initially proposed measure-- I'd propose increasing the comsat base cost from 1 to 3 (instead of the above proposed 5). I think that would go a long way towards curbing any comsat abuse, but would still be close enough to current values for other applications (outpost bases & just barely close enough for trooper bases I think) that we wouldn't need to make a significant change to AI planning.

I completely agree, I haven't checked troop pod costs when I suggested 5.

Some figures:

Assuming default parameters (RULE_SHIP_PART_COST_FACTOR == 1, etc.) and varying colony base hul cost between 1, 3 and 5 we get:

First colony base costs 128.2, 130.2 and 132.2 (i.e. for hull cost 3 you are increasing cost, over hull cost 1, by a 1.55%, for 5 it's a 3.1%).
Once you got 10 colonies, it costs 193, 195 and 197 (increases of 1% and 2%).
For hundredth colony, it costs 841, 843 and 845 (increases of 0.24 and 0.48%).

A troop drop costs:
For first ship 4.18, 6.18 and 8.18 (so the increases are proportionally huge and most probably are relevant in early game when PPs/turn are in the order of tenths).
For tenth ship 5.8, 7.8 and 9.8 (still big increase in percentage, 34.5% and 69%, although big PPs/turn will feel it just slightly).
For hundredth ship it would be 22, 24 and 26 (so in late game it would have little effect, considering PPs/turn).

Any comsat (first and thousandth) costs 1, 3 o 5 (i.e. increases of 200% and 400%). This will have a great effect when PPs/turn are low but little effect latter on, except for isolated planets. Examples:
- A blockaded planet with 20 PPs per turn could produce in three turns 60, 18 or 12 comsats (so base hull cost 3 achieves a similar effect than 5).
- An Empire with 200 available PPs/turn that wants to produce quickly 100 comstats for cannon fodder in its choke-point system and devote the rest to build military will have for military production 166, 100 or 33 PPs/turn, so the effect is rather strong for 3 and way more for 5.

Watching this, I think there is very little effect on the costs for either 3 or 5 except for troop drops on early game, for which either 3 or 5 will have a noticeable effect on the required PPs, while 3 seems enough to greatly counteract the comsat abuse.

Oberlus wrote:An Empire with 200 available PPs/turn that wants to produce quickly 100 comstats for cannon fodder in its choke-point system and devote the rest to build military

Keep in mind also, any empire that pops out 100 comsats will then have increased their military production costs for their warships by 100%, a pretty significant thing.

Dilvish wrote:

Oberlus wrote:An Empire with 200 available PPs/turn that wants to produce quickly 100 comstats for cannon fodder in its choke-point system and devote the rest to build military

Keep in mind also, any empire that pops out 100 comsats will then have increased their military production costs for their warships by 100%, a pretty significant thing.

i think the production cost hit only applies after being produced. So you can pause your comsats in batches until one turn before you need them. i think tripling the cost is a great idea

How much time does a turn take? Is each turn just minutes or hours?
Or is each turn days, weeks or months. This is actually important to
the whole question of whether comsats could block an enemy invasion.
If a turn is only minutes or even a few hours I might buy it. Days or
weeks and it falls apart. Months and it is totally bogus.
AFAIAC, unarmed comsats stopping an invasion, even temporarily, is
a gross abuse of a hole in the rules. It shouldn't happen.
To fix it I would propose that any ship with a combat value of zero
not participate in combat at all. After all combat is resolved any side
with just zero combat value ships left is just automatically eliminated.

EricF wrote:To fix it I would propose that any ship with a combat value of zero
not participate in combat at all. After all combat is resolved any side
with just zero combat value ships left is just automatically eliminated.

In that case you would have to put mass drivers on your heavy troop ships in order make them count... sounds also not good.

Better idea is the battle scanner (target all armed ships before targetting unarmed ones).

Better idea is also to that unarmed ships are unable to block invasion troops (not sure if they do actually).

Note that the "you can trigger combat with a comsat" is actually a feature ATM. Actually its a workaround for a missing feature.

Ophiuchus wrote:Better idea is also to that unarmed ships are unable to block invasion troops (not sure if they do actually).

They don't.

My/the main concern here is that they work as cheap cannon fodder that, with the correct numbers, cripple the offensive capabilities of an invading fleet by taking "out of the combat" a considerable number of shots that otherwise would have destroyed the defending ships and planetary defenses, which implies that the invading fleet will be receiving more hits and would be connecting less hits than what would have happened if the defending fleet had all the PPs invested in armed ships instead of comsats. Is in this last sentence where the main concern is: you can get more out of your PPs by doing something "stupid" (building useless floating junk) than by building actual armed forces.

EricF wrote:How much time does a turn take? Is each turn just minutes or hours?
Or is each turn days, weeks or months. This is actually important to
the whole question of whether comsats could block an enemy invasion.

Not really. The realism here is neither relevant nor fixable. If combats take minutes, then there should be many many combat turns (thousands, millions) for each turn, given that we are traveling light years and increasing planet populations by millions in a single turn. If combat takes months then the whole combat system is incredible unrealistic and hard to swallow.
This concentration of real-time activities that are packed withing the abstraction "turn" are supposed to be happening at very different paces and we just addapt it to unrealistic dimensions for the sole purpose of pouring some fun into our brains in a way that it can handle it better. The same happens with games like Fallout or Elder Scrolls: you walk like a few hundred steps and you have moved from one sunny, sand country to an iced, snowy mountain range; but try to make the map for real dimensions and you take out all the fun.

I think we should give the tripled base hull costs a thorough evaluation for solving this before proceeding to other measures. I've posted a PR for it now to give anyone else a last chance to comment before it goes into master; I plan to merge it fairly soon.

So has anybody some feedback if comsats are still too usable? Question arises because I wanted to nerf decoys slightly via targetting preferences Combat Preferred Targetting thread and it is not clear if that is necessary.

I can use them effectively. 3 or 5 cost is affordable for chaff when any weapon part is 20.

Oberlus wrote: ↑Fri Aug 31, 2018 8:31 am I can use them effectively. 3 or 5 cost is affordable for chaff when any weapon part is 20.

And at what price would that stop being true for them and the other hulls that currently cost in the 5-6 range?

Honestly, in late game, building decoys is pretty much always an efficient strategy for all-out warfare considering you have and are facing a reasonably sized fleet.

Consider building either a warship or M decoys for the same PP.

Assuming you already have N warships of same type, the fleet strength will be increased by an approximate factor of [(N+1)/N]^2 or (1+2/N) for large N.
Assuming all your decoys are shot down in first combat bout, you reduce damage to your warships by factor M/(N+M) in that bout. Over the entire battle, you will thus reduce damage by factor 1 - (M/3(N+M)) or, equivalently, increase the effective structore of your ships by the reciprocal 1/ (1- M/3(N+M)) which for large N converges to 1 + (M/6)*2/N).
Assuming standard (structure*damage) heuristic to evaluate fleet strength, that factor is what your fleet strength increases by.

So, as long as you get more than 6 decoys instead of a single warship, in an all out fight, you would probably go better with adding some decoys. Maths above is obviously only valid if you haven't any decoys yet in your fleet. As warship value scales with number of decoys, there will generally be a ratio of decoys to warships you want to have which will depend on the cost ratio.


As far as the current combat system is concerned, it can not be avoided that decoys will always have their place. Even if the smallest hulls cost 20PP each, as your warships will easily cost several hundreds of PP in lategame, it will be cost efficient to add decoys.
There is obviously a strategic trade-off of increased fleet strength vs fleet upkeep and "guaranteed" PP loss but at the very least for the decisive battles or short-term war efforts, it will be a good idea to add decoys to the fleet.

Great calculations, Morlic.

So, to forbid decoy/chaff strategies (if that is what you want), targeting (if implemented) should take into account the damage output or the total PP cost of the ships: decoys would have low or no damage output and/or low PP cost compared to the other available targets.

But, if Vezzra's concept of orbital combat is to be accepted as the right one, then decoys should be a valid strategy.

In single player games, you're free to not use decoys if you don't like them. If the AI does, fine, that's still the AI with all its limitations, so no big deal - at worst, see that as an extra bit if challenge.

In multiplayer, everyone can (should) agree on chaff rules at the start of a game (what is considered chaff, are they allowed, what is considered as a legit chaff design if they are, etc.) That was the standard in the Stars! multiplayer community.

So I don't understand why everyone here seems to think solving the decoys/chaff "problem" is such a big issue.

Rather than worrying about exactly what cost to adjust a hull to to make cannon fodder ships less useful, I reiterate the suggestions from the previous page: make combat target selection ignore unarmed ships (ie. those without direct weapons or fighter parts) unless there are no targetable armed ships present. Combat could still be used to take out troop ships (but only after all armed ships defending them are gone). Relatively cheap ships with minimal armament can still be created and function as fodder, but the point is just to eliminate the particularly "abusive" seeming cases with hulls intended for a specific non-combat purpose.

Later, I'd probably consider reverting this rule and implement target prioritization (eg. with parts or leaders) which players could use to make their fleets target armed or unarmed ships first. This would make cannon fodder designs ineffective, but only be needed to respond to an opponent making use of such a strategy.

IMH, Geoff's post settles the question.