7/7, 6 hours, 90 AI Progress and this CPA ? wtf

[Bognor]

Wait? CPAs still use the flat ship count? I thought they used the Relative Strength like waves/etc. do?

I think they have to because the announcement always specifies a number of ships, not a strength of ships.

[TechSY730]

Yea the flat ship count system kind gives an interesting dynamic.
There is the risk that a lot of low cap ships will be freed, making the CPA quite dangerous. In return, the AI looses quite a bit of defensive ability because they had to "pay" a lot for those low cap ships when they were initially spawned for defense, making future attacks easier (assuming you can survive it ).

But if they chose a bunch of high cap ships, then the CPA won't be very threatening, but at the same time, the AI hasn't lost all that much in terms of defensive strength.

Basically, if the AI got a bunch of low cap ships, its a higher risk, higher reward type thing (more likely to actually hurt you, but if they fail or don't do all that much, the AI will lose out on more)

[tadrinth]

Ouch, didn't realize CPAs don't take caps into account when picking ships.  That's potentially really nasty.  I can see why it's beneficial to have the number of ships freed be exactly the number advertised in the warning, but I'd rather have the strength of the CPA be predictable than the number of ships. 

This also means it is advantageous to selectively corrupt designs for low-cap ships, so they don't show up in CPAs and wreck face.   If CPAs are what's killing you, anyway.

And it means that I am terrfied of the CPAs generated by the Spireling AI type, since they have tons of low-cap ship types.  Two Spireling/Chivalric AIs would generate some seeeeriously scary CPAs. 

[Kahuna]

Seems like AIs pay too little for low cap ships. (Especially for Starships in Exos)

What and how many ships did the CPA have in addition to the SSBs?

[TechSY730]

Seems like AIs pay too little for low cap ships. (Especially for Starships in Exos)

What and how many ships did the CPA have in addition to the SSBs?

Except for exos (which are known to be a bit...off in terms of cost balance), do you have any evidance for that, or just a "feeling"? (Keep in mind, number on one planet does not count, rate is what matters, do they get more than they seem to be able to reasonably "pay" for per given amount of time). From what I have seen, waves, reinforcements, and special forces' math seem to be working properly.

I have mentioned in the past that I do think one factor not put into consideration for the AI "costs" is build time and resource costs (the AI can pop out a SSB and its equivalent "cost" in laser gattlings in the same amount of time, even though there would be a good chance, for humans, it would likely take you longer to get that SSB out), but that only really matters in the short, and to a lesser extent, mid-term. In the long term, any skew by initial build time and costs begins to fade in terms of importance (provided you can keep them alive!), so I don't think it is that huge of a deal (unless you like to play more aggressively and faster, in which case, the AI being able to reinforce low caps at the same rate as the "equivalent" number of high caps, while you very well might not be able to, could start becoming noticeable).

I will say though, now that the "preferred ships for defense" per planet logic, there is an increased chance that the AI will build up a near un-counterable number of low cap ships on any one planet (yes, this can happen with high cap ships too, but generally, it takes much more, even if you scale based on the cap differences, to reach that point for high cap ships than low cap ones), so I wonder if the "ultra low" cap ships need a "max per planet" logic reinstated for them, now that the anomalous "skew" for them is now much more possible.

Basically, this may very well be a problem with some of the low cap ships themselves and/or some of the high cap ships. Like if a low cap ship has 1/20 the cap of the "standard" cap, but has, say, a single of that low cap has 24x the strength of a single of one of those mid cap ships (thus making a cap of that high cap ship 1.2x as good as the mid cap ship cap), then of course a ship cap ratio based cost will be off, because the stats themselves are even given their caps (aka, that lower cap ship is OP even given its low cap).



Finally, for starships, ever since their buff, I do think that at least reviewing how the math for how the AI gets them in waves deserves at least a lookover.

[LordSloth]

Strictly speaking about the exo-waves: the costs for those were rescaled rather arbitrarily. They aren't scaled by formula, but assigned a cost by best guess according to certain tiers. Keith asked for feedback and received some, took that into consideration, and did fairly well, imo.

As far as I remember the feedback he got was pretty targeted, not comprehensive. Now that there's the larger body of post-alpha experience to draw from, there might be a point in thread necromancy.

[Kahuna]

@techsy
i did have some evidence some months ago when i complained about "ai not caring about ship caps". most of the ai planet had more low cap ships than 96 ship cap ships and something..

so I wonder if the "ultra low" cap ships need a "max per planet" logic reinstated for them

i don't think that's necessary. ships just need to be balanced

Basically, this may very well be a problem with some of the low cap ships themselves and/or some of the high cap ships. Like if a low cap ship has 1/20 the cap of the "standard" cap, but has, say, a single of that low cap has 24x the strength of a single of one of those mid cap ships (thus making a cap of that high cap ship 1.2x as good as the mid cap ship cap), then of course a ship cap ratio based cost will be off, because the stats themselves are even given their caps (aka, that lower cap ship is OP even given its low cap).

even if the low ship cap ship and the high ships cap ship have the same cap health and dps.. the low ship cap ships win. that's because the high ship cap ships have less health and die quickly which will result in less dps. for this same reason snipers are so powerful (much more powerful than people think (they stay away far away from the fights so their dps never drops)).


lets say mid ship cap ships have 100% cap health and 100% cap dps.
low ship cap ships could have 90% cap health and 100% dps or 100% cap health and 90% dps
high ship cap ships could have 110% cap health and 100% dps or 100% cap health and 110% dps


EDIT:

LordSloth: Strictly speaking about the exo-waves: the costs for those were rescaled rather arbitrarily. They aren't scaled by formula, but assigned a cost by best guess according to certain tiers. Keith asked for feedback and received some, took that into consideration, and did fairly well, imo.

As far as I remember the feedback he got was pretty targeted, not comprehensive. Now that there's the larger body of post-alpha experience to draw from, there might be a point in thread necromancy.

i've said many times that the ais pay too little for starships/golems and a bit too much for hunter killers. after a couple of changes, tweaks, patches and stuff..... i've been able to barely stop h/k exos a couple of times... but starship/golem exos... not even once. h/k exos have much less firepower than starship/golem exos.

[LordSloth]

Yep - but honestly, I forgot previous discussion of starships in particular.

Point still stands, I hadn't had as much time as I'd expected free, and there's a fresh new wave of people/returned players with each sale. Now that people have caught up in playtime you might find a bit more consensus. Though I don't remember how that thread ended exactly, so I might forming the wrong conclusion.

Point I was making, it seems a more viable thing to revisit than my little idea of changing the way CPAs function since it is a list of costs and not something involving coding.

Changing the reinforce costs might also be something to consider, with my 7 difficulty experience it doesn't seem very off target - although I've had extremely unpleasant experiences with Stealth Master complicated with Fallen Spire and Medium Broken Golems messing up the careful balance.

I am running an extremely vanilla 7 game with no expansions enabled and human colony rebellions, resistance, and marauders turned on at the moment. About 8 hours in. I got distracted a bit during the week by Dominions 4 and Total War 2, but I'll be getting back to completing a completed game this weekend, as much fun as checking those two out was.

Joy of joys, I even ran some CAT5 so I might even get back into co-op after all this time.

Tangent aside, I don't even remember what I was writing about just now.

Biscuits, right? I am a bit hungry, I should post again when I can actually concentrate on a single line of thought... Perhaps pulled pork sandwich?

[TIE Viper]

The pig is such a wonderfully delicious animal. 

[TechSY730]

even if the low ship cap ship and the high ships cap ship have the same cap health and dps.. the low ship cap ships win. that's because the high ship cap ships have less health and die quickly which will result in less dps. for this same reason snipers are so powerful (much more powerful than people think (they stay away far away from the fights so their dps never drops)).

I wouldn't nessecarily say that's allways the case. Let's look at one extreme example, a cap of Zenith Bombard vs cap of Laser Gattlings. Yes, the zombard can easily one shot the laser gattlings, and can take plenty of punishment from the laser gattlings too. But there are so many laser gattlings, each zombard can only fire one every 30 seconds, the zombards won't be able to one-shot the laser gattlings fast enough, and will actually lose (assuming the gattlings can "corner" the zombards so they can't kite effectively anymore, kiting will make the analysis much more difficult). Yea, zombards are not an ultra low-cap ship type, but a "merely" "just" a low cap, but still, it is a counter-example.


But I get what you are saying. The effects of cap (aka, army size), rate of fire, dps, range, and individual health interplay in very complex ways to determine "net utility", in ways that merely trying to normalize cap stats cannot always compensate for, and the current balance of the ships has (no offense Keith) failed to take into account. That has thrown off quite a bit of the low-cap vs. high-cap ship interactions, different stats "contribute" to utility in different ways for high-cap ships vs low-cap ships.

[KDR_11k]

Don't most games actually give swarmers pretty high DPS per cost (or high HP per cost for ones that are meant to act as meat shields)? AIW's balancing doesn't count swarming as a disadvantage while in other games swarms are considered weaker due to AOE. Now not every swarmer has to be as nasty as Younglings but you'd think laser gatlings would be glass cannons that could completely chew up a force if left unchecked.

[Bognor]

AOE (area of effect weapons) are rarer in AI War than many games, though.  Though I guess they're more common now than they used to be, thanks to Plasma Siege Starships and others that use that mechanic, and the reworked Spire Starships.