AI enhancement (test games)

[kyrub]

Thank you both for your support!
Just remember that I could not have done 1% of this without the savegame analysis of sargon0.

@RefSteel
Some of the changes may actually make AI easier to handle: the feared SoDs were down to the action_radius bug, as it seems. But they were slow and ineffective. So we lose one AI factor while gaining something else, the mobility and the capacity to defend borders. It's hard to tell the result. (I will try to let the sub-fleets merge more than originally intended, to create bigger AI fleets overall).

@jmas
Here you are: The space_screen_bombing part of the code is a horrible work (they had a hangover or something?). There are two bugs and more conceptual flaws, some of them unrepairable.
Beams_stronger_than_bombs bug - The fonction multiplies the bombing damage weapon by weapon based on the ammo count: Bomb=10, larger_missile_rack=5, smaller_missile_rack=2. When it comes to torpedoes, they wanted to make them quite effective because of unlimited ammo, the multiplier is 30. Unfortunately, the beams are considered by the program as having unlimited ammo! So the beams are 3 times stronger "bombing" weapon than bombs themselves...
(The other bug halves damage for every missile instead of every torpedo, just for your info.)

This explains why the space_bombing occured always so overpowered in the game.
(I do not dare to mention the other flaws, since they are bad exploits.)

[jmas]

kyrub, excellent job analyzing the code, thank you for yet another insight into the game. Of course I have more questions

By "space_screen_bombing" do you mean the "Bomb the Enemy Planet? Yes/Cancel" part of the turn? On first reading your post I thought you were referring to bombing on the combat (or some say "tactical") screen, and I don't recall seeing problems with bombs, beams, or missiles there--torpedoes have been more rare in my games.

Assuming we are talking about the "Bomb the Enemy Planet? Yes/Cancel" option: The strategy guide says that that bombardment is supposed to yield "the results of 10 rounds' worth of attacks against that planet", so multiplying the damage of a single round times the ammo count of the weapon seems OK (since no limited-fire weapon has > 10 rounds), except of course for the x30 multiplier for beams (that is what you meant by "3 times stronger 'bombing' weapon than bombs themselves", right?), the halving of missiles instead of torps, and (I would say) the x30 for torps. Can you explain why you think it was intended they be stronger due to their unlimited ammo? Going by the "10 rounds" concept they should just get 5 shots since they fire every other turn, with the damage from each of the 5 shots being halved--so assuming the halving would be factored in for the "representative" combat round that is then multiplied, the multiplier for torp damage should be 5 (otherwise 2.5). Granted that the strategy guide is far from infallible--sargon, for one, has clearly shown otherwise--I'm just not clear what your reasoning is in speaking about the intent regarding torps. Sorry to carry on with this many words, just trying to explain my ideas. Great work!

[kyrub]

By "space_screen_bombing" do you mean the "Bomb the Enemy Planet? Yes/Cancel" part of the turn?

Yes - and it is valid for AI players as well.

The strategy guide says that that bombardment is supposed to yield "the results of 10 rounds' worth of attacks against that planet", so multiplying the damage of a single round times the ammo count of the weapon seems OK ........... Can you explain why you think it was intended they be stronger due to their unlimited ammo? --------I'm just not clear what your reasoning is in speaking about the intent regarding torps.

Thanks for mentioning what the OSG is saying. I don't have it so I'm rellying my judgement purely on what I see in the code and I am trying to guess the intention. I agree that '10 rounds' conception cannot explain why unlimited ammo (-1 for the code) is exchanged with 30. In my eyes, it is logical that the carpet bombing is dependent on the attackers' ammuninition, how often they have to rearm. In this respect, torpedoes factor (30) is understandable.

BTW, if you have the OSG, what exactly does it say about
- the usage of battle computer during this type of bombing
- the ship movement through nebulae, and more specifically about moving through it with warp 1 engines?

[jmas]

I agree that '10 rounds' conception cannot explain why unlimited ammo (-1 for the code) is exchanged with 30. In my eyes, it is logical that the carpet bombing is dependent on the attackers' ammuninition, how often they have to rearm. In this respect, torpedoes factor (30) is understandable.

I guess what I don't follow is, if you are OK with torpedoes being more effective (having a higher multiplier) than one would expect, and the reason you give is that they have unlimited ammo, then why do you object to a higher multiplier applying for beams (which are also unlimited)?

I agree that having beams be more effective than bombs seems a little silly, I'm just not sure I follow you on torps; why give them a "buff"? (unless you are trying to make them a better weapon in general; I know there is a thread on here a ways back about their relative ineffectiveness given the tech levels at which they occur, ship space consumed by them and the alternatives (usually) available)

BTW, if you have the OSG, what exactly does it say about
- the usage of battle computer during this type of bombing
- the ship movement through nebulae, and more specifically about moving through it with warp 1 engines?

I'm not at home so don't have the guide with me right now, but I don't recall it saying anything about the battle computer applying during the optional bombardment portion of the turn. I do know that the tables in the back for damage dealt to planets by weapon type have a column for a defense level of 0, which applies when there are no missile bases (either not present or destroyed). I always thought of that in relation to tactical, on-screen combat--I don't think it ever occurred to me before to wonder about whether the battle computer influences the off-screen bombardment. If I find something in the guide about that I'll post it.

If I recall correctly the guide says that warp speeds greater than 1 are reduced by 1 in nebulae (so warp 2=1, 3=2, 4=3 etc.), however I think that others may have posted here on RB that it actually works a little differently. I just looked at the thread sargon started on movement--didn't see nebulae addressed there, but I could swear I have seen this topic discussed in some thread.

[Cyneheard]

In the coding, does the space-screen bombardment take into account the planet's deflector shield? I think that it does, since the last time I built a shield while a planet was being bombed, the damage dealt dropped by 90%, but I'm not 100% certain that this is true. If the deflector shield doesn't matter, then it should be (relatively) trivial to work around the halving-the-wrong-stuff problem by changing the ammo multipliers to give the correct results (Unlimited does 15x, Missiles 10x and 4x, Bombs still 10x). If not, then I have no idea how to fix it.

Another problem: Torpedoes fire every other round. Beams fire every round. So is 30x even a good multiplier for torpedoes? Maybe fix the halving-the-wrong-stuff bug, and changing the multiplier?

[kyrub]

My point on torpedoes / beams bonus is twofold:

a) I just cannot accept that the beams should be considered a mass destruction weapon, while the torpedoes surely belong in this bracket. It is so unrealistic, that the ships with weapons with no area effect are able to do substantial damage or even destroy whole industry and, behold, kill all the population hidden in the bunkers! How much damage would a 10.000 fighters fleet do on a bombing raid? How much a 1.000 fleet of bombers? How much a fleet of 100 flyers, continuously firing torpedoes? The answers seem simple to me.

b) The in-game logic is the second part: if all you need to glace a planet is beams n' beams, why should anybody use the bombs? They could be always substituted with something else (shield-halving beams or rockets), but I thought them more effective, at least. They are not.

But, Cyneheard, you are right: the shield can influence the equation. If the beams cannot get through the shield, they do nothing at all. And the halving beams get 0 advantage. Is this what they wanted??

[jmas]

--Do you plan to reduce the effectiveness of beams then? That part sounds fine to me. I'm not as much "on board" with the torpedoes, though I can see your point.

--My mistake: the guide says "There is a warp 1 speed limit for ships while crossing through nebulas." So, warp speeds would be reduced TO just one parsec per turn, not BY one parsec per turn. I still think that others may have posted here on RB that it's different than what the guide claims.

--I don't see anything in the guide about whether battle computer levels are considered during bombardment. You could try conducting in-game tests with ships with identical weapons and different battle computer levels. Planets subject to bombardment have a defense of zero, and only the first five levels of attack_level superiority make a difference (hit percentages max at 95% at (attack_level-defense_level) = 5), so any difference should be observable within the window of attack_level = 1 to 5 (not from, say, 5 to 10 ).

[kyrub]

jmas, I intend to change as little as necessary, so unless I see that there is a bug or a clear design flaw, I just pass on it. I am unsure over the beams' thing right now (because they have to punch through 2*planet_shield to do any damage, thanks Cyneheard) and I will most probably let torpedoes do what they did.

There will be two versions of the patch, one extremely conservative, to allow anybody to play the old game bug-free.
The other one, the "Twillight of Orion" version, will - on the top of that - try to boost up the AI thinking routines and clear the design flaws and slightly improve the interface.
So, no fancy mod stuff, I dislike it. I have a list of about extra 7 changes to the game I want to put in the twillight patch, though.

Everything else stays as it is. The Moo I is an excellent game.

[jmas]

I agree with you, it is an excellent game in my opinion too, and I want to say THANK YOU for doing this work.

One thing I am curious about, having had this discussion: will you stop the halving of missile damage, since you mentioned it as a "bug"?

[kyrub]

The current status of yes/no_bombing function in my head

1.31 patch
missiles normal / torpedoes halved.

twillight patch
missiles normal / torpedoes halved,
improve the bombing / invasion decision for Bulrathis,
...
+ is doubtful to happen, but still considering:
?) to decrease the average damage done to POP, thus changing the ratio of damage dealt to POP/ind to about 1/5 or even 1/6 (now 1/4) - to make the conventional bombing effect clearly distinct from Bio bombing and to 'counterbalance' the increase for the missiles
??) to halve the beam damage multiplier 30 > 15, at least
???) to alter the strange role of computers in the bombing result