New Version - February 7th (2/7)

[ashendashin]

They'll continue to be aggressive, but domination is time-consuming. If they fear that they can't get all the capitals before someone else wins, they'll shift gears.

G

Since I play with 22 civs I have to agree...but for standard games I assumed it can't be that bad of an idea to push it. Liberal bombardment and tank construction should do wonders for any warmonger ever but I haven't played in a long time...

 

[Gazebo]

Since I play with 22 civs I have to agree...but for standard games I assumed it can't be that bad of an idea to push it. Liberal bombardment and tank construction should do wonders for any warmonger ever but I haven't played in a long time...

The diminishment would only matter if it drops domination below, say, science. If the civ is woefully behind for any other victory, they'll go for domination. If, however, a pivot to science or diplo is marginally possible, they'll try to switch gears.

G

 

[ElliotS]

It depends! The short answer is that, the longer a game goes, the less likely a civ is to attempt domination if they feel they are too far behind in their efforts - other VCs, like science, become more tempting. If you want to spoil your game, you can look at the grandstrategyAI log to see preferences.

G

2 Things:
1- Can the AI tell that when it's 5 techs and 2 policies behind the leader going into industrial, it needs to rush war to avoid falling farther and farther behind? This is the biggest missing piece from what I see, but I don't know the code so maybe they do consider it.
2- I take it this is all based on data, where slightly behind AI win more often under this behavior over slightly behind AI that jump into poorly advised war.

Also thanks!

 

[Gazebo]

2 Things:
1- Can the AI tell that when it's 5 techs and 2 policies behind the leader going into industrial, it needs to rush war to avoid falling farther and farther behind? This is the biggest missing piece from what I see, but I don't know the code so maybe they do consider it.
2- I take it this is all based on data, where slightly behind AI win more often under this behavior over slightly behind AI that jump into poorly advised war.

Also thanks!

Yep. The code is fairly simple for domination VC criteria.

Factors:
Base leader flavors and personality traits considered.
# of nearby civs.
'Cramped'ness of civ.
Policy/religion/UA/UB/UU/UI priority modeling
Current Era
Nuke unit access
Number of capitals controlled above 1
Military strength relative to other players

These are all weighted and weighed.

G

 

[ElliotS]

Yep. The code is fairly simple for domination VC criteria.

Factors:
Base leader flavors and personality traits considered.
# of nearby civs.
'Cramped'ness of civ.
Policy/religion/UA/UB/UU/UI priority modeling
Current Era
Nuke unit access
Number of capitals controlled above 1
Military strength relative to other players

These are all weighted and weighed.

G

Does that mean there is no 'despair' factor? AKA - I'm going to lose if I don't do anything, so I might as well war and try to get a 1% chance at success?

There's no bigger cause of war in a science victory game than "This other person might get to X VC first."

 

[Gazebo]

Does that mean there is no 'despair' factor? AKA - I'm going to lose if I don't do anything, so I might as well war and try to get a 1% chance at success?

There's no bigger cause of war in a science victory game than "This other person might get to X VC first."

Not really, no. I could probably add one fairly simply though.

G

 

[Txurce]

Not really, no. I could probably add one fairly simply though.

Is there a danger of artificially increasing the odds of a Domination Victory with kamikaze attacks?

I think the AI is better off sticking with whatever VC has th best odds at that time. No reason to attack with a 1% chance if an SV is possible at 2%.

The exception would be a collective AI effort: we are all going to lose if we don't act collectively, so let's gang up on the human. That makes sense, but then would make all end games pretty much gauaranteed bloodbaths. And in the case of, say, a small civ on the verge of a CV, a collective war against them may well work too much of the time.

 

[ElliotS]

Not really, no. I could probably add one fairly simply though.

G

I think that it could backfire, but is worth a test.

I know that pretty much every time I'm going to lose, I declare war on the leader to see if I can salvage it.

Even the game I'm in as Maya. I went Progress -> Rationalism. I want to make spaceship go vroom vroom.

Nonetheless I felt like I would get edged out by my battle for religion and politics with the nearby Ottomans being a thorn in my side. I was ahead in techs, but losing on policies and some other stuff, so I DoWed and kicked him in the teeth. I might need to again, but I went to war before it was too late and I feel much more stable now. (Also I was able to use his 'abuse' of Defensive Pacts against him by declaring war on his ally across the ocean, costing me a few boats to avoid a war on 2 fronts with his in-Continent ally and bypassing our DoF.)

 

[ElliotS]

Yeah the biggest problem I see with the 'Despair Factor' DoW is making the game unfun. Not sure what the right balance is, but I think a certain level of it could lead to a bit more warmongering, which I would like to see.

Edit to clarify: Too much 'Despair Factor' and every AI suicides the moment is gets a certain amount behind, which leads to more suicides and insanity. None is obviously better, but a certain amount could make AI more realistic and better.

 

[Txurce]

I think that it could backfire, but is worth a test.

I know that pretty much every time I'm going to lose, I declare war on the leader to see if I can salvage it.

Even the game I'm in as Maya. I went Progress -> Rationalism. I want to make spaceship go vroom vroom.

Nonetheless I felt like I would get edged out by my battle for religion and politics with the nearby Ottomans being a thorn in my side. I was ahead in techs, but losing on policies and some other stuff, so I DoWed and kicked him in the teeth. I might need to again, but I went to war before it was too late and I feel much more stable now. (Also I was able to use his 'abuse' of Defensive Pacts against him by declaring war on his ally across the ocean, costing me a few boats to avoid a war on 2 fronts with his in-Continent ally and bypassing our DoF.)

Good move!

 

[ashendashin]

Yeah the biggest problem I see with the 'Despair Factor' DoW is making the game unfun. Not sure what the right balance is, but I think a certain level of it could lead to a bit more warmongering, which I would like to see.

Edit to clarify: Too much 'Despair Factor' and every AI suicides the moment is gets a certain amount behind, which leads to more suicides and insanity. None is obviously better, but a certain amount could make AI more realistic and better.

If you want to avoid it being suicide then pushing back the timing of the despair factor would make sense. IMO the most dangerous period is at the beginning of ideologies, when leads can turn into unstoppable snowballs.

 

[ElliotS]

If you want to avoid it being suicide then pushing back the timing of the despair factor would make sense. IMO the most dangerous period is at the beginning of ideologies, when leads can turn into unstoppable snowballs.

To be clear you're saying that the ideal time for despair factor wars is around when ideologies start getting adopted? Or are you saying that by that point it's too late?

 

[ashendashin]

To be clear you're saying that the ideal time for despair factor wars is around when ideologies start getting adopted? Or are you saying that by that point it's too late?

Around the time, yes. Leaning towards very early or even a bit before. Any major policy leads need a swift kick as soon as possible, since that means they'll become the permanent lead in everything their ideology grants. Takes some time for everything to get online as they go through the first two tiers, but once ideologies hit they're already ahead in where it matters.

 

[Gazebo]

What a beautiful naval invasion. Bravo, Darius.

G

 

[Gazebo]

Here's what I'm going to add:

bool bDesperate = GetPlayer()->GetPlayerPolicies()->GetLateGamePolicyTree() != NO_POLICY_BRANCH_TYPE && !GetPlayer()->GetDiplomacyAI()->IsCloseToDominationVictory() && !GetPlayer()->GetDiplomacyAI()->IsCloseToCultureVictory() && !GetPlayer()->GetDiplomacyAI()->IsCloseToDiploVictory() && !GetPlayer()->GetDiplomacyAI()->IsCloseToSSVictory();
    int iTotalNumDangerPlayers = 0;

    // Count the number of Majors we know
    for (int iMajorLoop = 0; iMajorLoop < MAX_MAJOR_CIVS; iMajorLoop++)
    {
        ePlayer = (PlayerTypes)iMajorLoop;

        if (GET_PLAYER(ePlayer).isAlive() && iMajorLoop != GetPlayer()->GetID())
        {
            if (pTeam.isHasMet(GET_PLAYER(ePlayer).getTeam()))
            {
                iNumPlayersMet++;
                if (GetPlayer()->GetPlayerPolicies()->GetLateGamePolicyTree() != NO_POLICY_BRANCH_TYPE)
                {
                    if (GET_PLAYER(ePlayer).GetDiplomacyAI()->IsCloseToCultureVictory() || GET_PLAYER(ePlayer).GetDiplomacyAI()->IsCloseToDiploVictory() || GET_PLAYER(ePlayer).GetDiplomacyAI()->IsCloseToSSVictory())
                    {
                        //Close to nothing, and someone else is? Eek!
                        if (bDesperate)
                            iTotalNumDangerPlayers += 25;
                        //someone else is close, but we are too? Let's keep an eye on domination as an option.
                        else
                            iTotalNumDangerPlayers += 5;
                    }
                    //They have an ideology, and we don't? Uh-oh.
                    else if (bDesperate)
                        iTotalNumDangerPlayers += 5;
                }
            }
        }
    }

The 'iTotalNumDangerPlayers' value is multiplied by the current era value, and added to the priority model of domination. So, essentially, players who are falling behind will be more inclined to view domination as a potential avenue of winning (or at least slowing things down).

G

 

[ashendashin]

}
The 'iTotalNumDangerPlayers' value is multiplied by the current era value, and added to the priority model of domination. So, essentially, players who are falling behind will be more inclined to view domination as a potential avenue of winning (or at least slowing things down).

I'm honestly surprised that such desperation factors didn't already exist. Maybe I have some trauma from having everyone turned against whenever I took over the world congress as Germany...

Will we be seeing this as a late experimental push in the next patch or is this what you're testing beforehand?

 

[Gazebo]

I'm honestly surprised that such desperation factors didn't already exist. Maybe I have some trauma from having everyone turned against whenever I took over the world congress as Germany...

Will we be seeing this as a late experimental push in the next patch or is this what you're testing beforehand?

There are 'crisis mode' elements elsewhere, but none of them intrinsically push the civ towards domination...they simply make the AI hate you, which often leads to war.

It's not a huge change, so yeah, it'll be added as experimental.

G

 

[Blue Ghost]

Any estimates on when the next version will be released?

 

[Gazebo]

Any estimates on when the next version will be released?

Every time someone asks, I add 5 minutes.

 

[ryanmusante]

Every time someone asks, I add 5 minutes.

I love it!