RevolutionDCM for BTS

[glider1]

Eh Os
Thanks for checking these things I'll put you down on the credits for this release for sure because it has made it so much easier to have a second pair of eyes. The pace to release 2.90 is gathering. The AUA I had to extract without complete commenting which made it hard work. Unit suiciding I didn't even know was an AUA option. I'll take a quick look but it probably will have to be released for version 2.91 I'd say at this stage. Updates to 2.90 I expect still to happen for a month or so. I'm in final testing for 2.90 and the stability of the mod and small details are really important to ensure at this stage.
Cheers

 

[os79]

Very well.
Fortunately, I still have RAND to enjoy in meanwhile. I am glad I was of some help .

Yes, Unit Suiciding is in AUA BUG tab page.

Basically it see your set odds (for example mine is 50 for Hunt). Then divide that by the amount of units inside a stack. For my example, I default at 10 units per a stack. So that means the odd of 50 divided by 10 and that equals to 5% odds that the weakest unit in the stack will attack the enemy. Only when the odds arrive at 50% or higher, will the regular combat system kick in.

 

[glider1]

Eh OS
Someone in the RAND forum reported that automate protect CTD's if the protected unit dies leaving the protector with no object to protect. Are you seeing that?
Cheers

EDIT: Gunna have to fix unit suiciding if you think it is broken, because it is there in the options page for AUA.

 

[os79]

1. Thanks on fixing Unit Suiciding. I love it.

2. I don't use Auto-Protect. It wasn't even one of my visualized Automation.
Someone else like the idea of it and Afforess obliged that person. Don't remember who.
But yeah I THINK it makes sense if Auto Protect an unit then that very unit dies, then maybe it crash because that unit is no longer in the game and the code did not compute for that possibility. I dunno, just a guess.

 

[Thomas SG]

I've got a question about AI autoplay. Am I right that a human on autoplay gets the AI bonuses? Doesn't really hurt in singleplayer games but in multiplayer games it may be very unfair.

 

[Voyhkah]

Would it be possible to base my upcoming mod on RevDCM? Crediting you, of course!

 

[os79]

There had been over ten modpacks that are based on this modpack . So yes you can. You just need a good merging experience and the understanding of BUG and WoC Lite functions. Then you are all set.

 

[glider1]

I've got a question about AI autoplay. Am I right that a human on autoplay gets the AI bonuses? Doesn't really hurt in singleplayer games but in multiplayer games it may be very unfair.

Eh Thomas
Will have to look into it good question. Is multiplayer proving stable enough with your mod?

Cheers

 

[Thomas SG]

Eh Thomas
Will have to look into it good question. Is multiplayer proving stable enough with your mod?

Cheers

I don't know. Hotseat is fine for sure. But that's not really MP. And I got only one real MP feedback from sivsta. But as it seems you can play CCV in MP through the internet.

 

[Duneflower]

Came up with a new question to wonder about as I've been updating NwT's extra leaders to use BBAI, particularly "playing to win": The ideal sum for VictoryWeights is 150...what happens when a leader's sum is significantly different from 150? My guess is that high sums make the leader overreach (and thus potentially screw up their chances at [/i]any[/i] victory) and low sums make them "lazy", ie not pursung any victory as aggressively; am I anywhere near the truth? Or am I not even correctly understanding just what those variables do?

 

[glider1]

Eh Dune
The place where the victory weights contact the code is just at:
CvPlayerAI::AI_getDominationVictoryStage() for the case of domination weight
Take a look at those functions for the different weights some time. You'll see that the weight for the leader in the XML just influences the inclination the leader has for a particular victory strategy relative to others and includes a random factor as well. Wild numbers do not change the basics of how the AI operates in that decision process. They just effect the inclination and the relative tendancy for one strategy or the other to gain momentum. If you throw in 999999 as a weight into the xml, even then it does not mean that the leader will only go for that victory strat. Just that his civ will incline to that strategy sooner. That's how I understand it at a quick glance. If you want a deeper understanding of what max values you can use that still make some sense, we can work that out by looking a bit deeper at those functions.
Cheers

 

[os79]

Any progress on merging in Unit Suiciding part of AUA yet? Or you decide for sure to hold that until 2.91?

 

[glider1]

Eh OS
Yeah will get onto that today and it will happen for 2.90 so long as the mother of reality lets me (The BIG MOR). Reason is, that I thought I had caught that code and so the question arises, did I miss anything else in AUA?
Cheers

 

[os79]

Nope, only this one. Already caught the Hide Automation error some time ago.

Only this one, and the tiny code tweak to prevent any CtD from Auto-Protect if there is any, that's it.

You did the great job merging based on code that was not clearly commented. I don't blame Afforess because it is possible he never thought that your team would be interested in AUA in base RevDCM .

 

[glider1]

Interesting bit is that the code is in there for unit suiciding, so the only thing could be either something in the testing or that the bit doesn't get changed from the interface.....
Cheers

 

[os79]

Ah, then maybe interface because I have it on in BUG. And I have TEN! Modern Tanks for example wandering around instead of attacking my enemies. Trust me, there are many, many enemy units available for that stack to attack at 5% odds (remember my explanation of how Unit Suiciding work? Well, that is this .)

 

[Duneflower]

@glider1

Ahh, k...so if I'm reading you right, it seems my intuition was more-or-less correct. Cool.

 

[glider1]

Cool Duneflower

Ok OS, if I have five swordsmen in a stack, four of them on combat 1 the other unpromoted, and if there is an enemy maceman adjacent to them, and I put the stack on auto hunt, allow unit suiciding 0% odds, then the weakest swordsmen should attack the maceman even if his odds of survival are 1.35%. Is that what should be happening?
Cheers

 

[glider1]

Ok OS tested it and it seems to be working. Here is what I did:
X is at war with Y in neutral plains territory:

X stack = 4 x swordsman combat 1 promotions, 1 x swordsman no combat promotion
Y unit = maceman adjacent to X with no promotions

With auto hunt unit suiciding 100% on all units in the stack, the stack retreats from maceman.

With auto hunt unit suiciding 0% on all units in the stack, one of the combat 1 swordsmen attacks maceman kills him and suicides, and the entire stack move to the attack square.

If you turn off auto hunt and just manually attack the maceman with the stack, you get the same result as the unit suicide 0% odds case.

So the check box for unit suiciding is working. The pull down odds drop down is working. The only issue that disagrees with your expectations, is that the strongest unit in the stack makes the attack not the weakest.

Cheers

 

[os79]

If you set it at 0%, all Hunting units will attack anytime, no matter what odds.

Better example.
You have Hunt odd at 50%
Unit Suiciding is on

OK, here is the two stories:
You have one unit (NOT STACKED!). Now It see an enemy unit and it is 51% odds, then yes that unit will attack. If it is 49% or less, then NO. Understand, Unit Suiciding is not at work here because you just have one unit in the stack.

Now to next story:

• You have 5 spearmen. Four of them are promoted. One is not and so is the weakest. You set Hunt odds at 50% remember? Ok, The enemy unit is by your stack (understand it is stacked!) The odds is at 49% Now remember if you had only one unit, it WON'T attack! But remember you have Unit Suiciding on, right? OK that change things now. Because you have the stack of five Spearmen. Now the number of units in stack is important.
• 50% odds divided by 5 units inside the stack.
• 50/5=10%
• What that mean is if the the odds is above 10%, the weakest unit will attack.
• Remember it said 49% right? Now the weakest spearman (the one with no promotion) attack the enemy and weakening the enemy. It dies.
• Now the odd change from 49% to maybe 61% because the enemy is weaker.
• NOW the normal combat system come in. The appropriate spearman in that stack will now attack and hopefully win the battle.