AI enhancement (test games)

[sargon0]

Thanks for the info kyrub.

On Trade I thought higher trade does give bigger DP gains but I rarely use more than minimum except with Humans who get double positive DP gains anyway. The main problem I have with the 'official' and 'actual' trade DP benefits is none really explain how you regularly see no gain or loss on turns during trade. The OSG implies even minimal trade gains +1 DP per turn until you reach +34 when the modifiers could turn this to a fraction rounded down to zero. The actual code suggests a gain of at least +1 per turn until you reach +25. So how do you regularly see zero gain (excepting Xenophobes who get half any DP gain)? Is the gain dependent upon a random or other condition or is gravitation impacting even when DPs are being gained? If there really are turns with no trade DP gain why does gravitation not take effect?

On the modifiers I have seen double benefit for some positive gains and going beyond +10. For example voting for someone at DP = -57 got +48 (double) gain. The +10 limit would seem more appropriate to temp modifiers rather than permanent modifiers. Also what is the permanent modifier table when losing DPs? The one in the OSG is:
DP = -99 to -67 - 1/3 penalty
DP = -66 to -34 - 2/3 penalty
DP = -33 to +50 - full penalty
DP = +51 to +100 - double penalty
Again this is certainly wrong or not universal. Even accounting for greater losses for Honorables and Xenophobes, I believe I have seen different ranges and multipliers when testing spying. Also the penalty for Xenophobes appears to be only 50% (not 100% as OSG) although Honorables appear to be expected 100%.

[kyrub]

Excellent observations. You forced me to run through that part again, which proved to be quite fruitful.

Thanks for the info kyrub.
how you regularly see no gain or loss on turns during trade.

All regular increases from treaties and trade are under 1 random 50% check.
The gravitation does not appear because it has
0) inc/dec relations check
1) d105 > abs (cur_rel) check ............sure to happen on 0 cur_rel
2) d2 - 1 "intensity" .......................... = 0 in 50% cases.
Well, it's not that strong, after all. It's rather pathetic, especially if you're on affable: 0.4 * 0,5 = 20% of possibility, regardless of inc/dec condition. If you include inc/dec probability, you get 10% - e.g. the gravitation happens every 10th turn.

On the modifiers I have seen double benefit for some positive gains and going beyond +10. For example voting for someone at DP = -57 got +48 (double) gain.

OK, my bad.
DP = -100 to -1 - double benefit
DP = 0 to +24 - full benefit
...
The case you mentioned was probably being Human race, no?
Here comes the flaw no.1: the human 2x benefit is counted after the +10 maximum check.
So you get: 24 (council vote for) * 2 (for having negative relations) (check for +10 max : OK) * 2 (for humans) = +96 increase.
It almost fits your case, that is the best I can do.

Also what is the permanent modifier table when losing DPs?
Also the penalty for Xenophobes appears to be only 50% (not 100% as OSG) although Honorables appear to be expected 100%.

- See the positive gain vice-versa.
- The rest... well, you told all I have to say on the subject.

the bug no.2: The "war damage" increase to temp_peace you get is mistakenly used even in times of peace, when you do not commit a war_offense. For instance, when you vote against an AI you have a NAP with, your temp_peace increases while your cur_relations worsen.

[sargon0]

Thats perfect kyrub. Nice to know the trade/treaties increase is conditional and gravitation is, as you say, pathetic. My mistake on the positive gains thinking your +10 limit was on the increase not the final amount. My vote was not humans but with your double benefit for negative relations giving +24 * 2 to -9 (max +10 ok) is perfect.

Nice to know, thanks again.

[kyrub]

One more note to clarify my stance on the gravitation flaw (intensity put aside) :

1) d105 > abs (cur_rel)

I am quite confident this was unintended as well, it lacks any logic, should have been probably: d105 < abs (cur_rel), which would make the extreme feelings (total love / total hate) for both human and AI players less sustainable than the mild ones.

Right now, once you attain an alliance, you don't have to worry about relations until you're too big for everyone...

[sargon0]

Maybe. My first reaction was you must be correct, the more extreme relations should be more volatile to encourage a general return to neutral. However, on reflection, Love & Hate are strong emotions not easily swayed whereas Neutrality changes very easily so maybe the game formula is correct. Although Alliances are possibly too stable, the difficulty is in obtaining them (without exploits like gifting) not preserving them. Similarly having driven someone to Hate this should not dissipate without good reason. This is not to say the game balances this ideally but perhaps it has a point.

[kyrub]

Fair point. On a second look, it explains as well why they let the zero as the point of reference instead of the base_relation point, because it would weaken the gravitation even more. I am still disappointed by this fact, though. The inter-racial relations are the flair feature of Moo; so the Alkari-Mrrshan basic mistrust should influence their gravitation and make it more troublesome.

------------
Another very important OSG quest:
Are there any smallest hints on how the war damage done to the AIs projects into the relations?

The code has sky-high numbers, like DP_loss = 5 * POP_bombed (e.g. -500 DPs for a 100 pop bombing raid).
It seems they forgot the denominator, but I don't know the general scale. The only hint in the code is that the max_for_fleet_war_damage = 50 / (?) DPs. (The denominator could be 5 or 10?). The problem translates itself into the enemy_of_my_enemy effect... DP_loss is DP_gain... you get the idea.

[sargon0]

OSG
Ship destroyed -1/-2/-3/-4 DP for small/medium/large/huge
Factory destroyed -1 DP
MB destroyed -3 DP
Pop destroyed conventional -5 DP
Pop destroyed biological -10 DP
10+ pop destroyed - DOW
colony lost - DOW
Enemy of my enemy - other players at war with damaged player will gain DPs for you of 1/4 of total amount

Observations
War damage also recorded in temporary modifiers ie. reduce treaty, trade & exchange modifiers by same amount of DP loss but increase peace modifier by 1/4 of amount. I seem to recall testing damage to a planet once but only seeing Pop loss reflected in change so cannot confirm if this is working as designed.

[RefSteel]

Does enemy-of-my-enemy bombing mistakenly yield diplo bonuses (with the target's enemies) equal to the diplo penalties with the target itself? This would perhaps explain things like Bulrathi praising me, on the turn I bombed Ursa in Imperium 10, for bombing various Human worlds slightly harder....

For what it's worth, I think the "sky high" penalties for bombing are appropriate (though the enemy-of-my-enemy bonuses do seem high, which would make sense if the above is true...) - one round of actually bombing a planet's population, even if "only slightly," should undo a lot of goodwill, it seems to me! One other thing I note though is that if the OSG (as quoted by Sargon) is correct, destroying lots of little fighters can have a massive diplomatic effect in comparison with their value. I seem to recall making more enemy-of-my-enemy headway by blowing up Huge ships actually, so I wonder if the code matches the OSG....

[kyrub]

Are they joking?? The design howler of the year! Whoever coded this had no idea about the proportions in the game. And I thought it should have been divided by 10, incredible.

So, next game I'll park 40 ships with Spores over a small planet and collect +100 DPs per turn. (Or should I say +400, because that 1/4 part does not work, though it hardly makes any difference). Ridiculous. And the worst part of the thing is it works ONLY for the human player, must be the worse Moo exploit, and unavoidable one.

(EDIT: now seeing your post RefSteel, spot-on with the bug, but I really cannot agree on the 'appropriate' part. Surely, it works with damaging the relations, but how on earth can you forge an undying friendship with a sworn enemy by bombing one 150 pop planet? It's beyond logic, it denies any experience.)

-------------
As far as I see:
- Only the biggest damage counts (cf. your observation, sargon)
- Factories and MBs part is not in the code.
- When biobombing, you lose automatically -10 Dps with anybody else (but still gain that zillion if he's in war with the victim).

[RefSteel]

(EDIT: now seeing your post RefSteel, spot-on with the bug, but I really cannot agree on the 'appropriate' part. Surely, it works with damaging the relations, but how on earth can you forge an undying friendship with a sworn enemy by bombing one 150 pop planet? It's beyond logic, it denies any experience.)

Hmm. Good point - I didn't think the full implications through. Of course the fact that they don't divide by four is a huge problem, as is the bug I believe you mentioned where you get diplo benefits from dropping spores even beyond the point where they have any effect at all (is this right?) ... but the fact that enemy-of-my-enemy sporing yields double the diplo benefits of bombing would be a problem even if the game reflected the OSG.

Good point about too-rapid diplo turnaround too. It might be best if there's some sort of overall limit (or diminishing returns?) to positive diplo effects like these in a given turn, but I don't know if this can be implemented easily. Also: I think I would love to see diplo effects apply equally to player and AI, but I worry that it would just result in AIs perpetually at war (or bouncing back and forth between war and alliance due to their automatic periodic positive diplo arrangements) since they don't have the ability to not (say) send transports to a non-ally friend's new world when they get a scout in orbit, or to weigh the value of death spores against the diplomatic costs (sighhhhh ... if there were to be costs that don't disappear in the sea of enemy-of-my-enemy, that is).

There's obviously good reason many of us avoid using Death Spores or exploiting enemy-of-my-enemy diplo bonuses intentionally, to avoid abusing the game. (The too-great effectiveness of e-o-m-e has long been known, but I hoped it was just the 1/4 problem. You've convinced me even that isn't enough. Actually, for e-o-m-e purposes, it occurs to me that killing population should be less effective than destroying ships....)