Historical Accuracy vs Historical Precision

[unmerged(10630)]

Well i playtested 2 GC:

there are 2 main flawns:

a) ai deviances on the political map

b) ai inactivity after 50-70 years game

Causes:

a) 1. innacurate initial set up
2. wrong choice in events
3. random wars-diploannexation with random alliances aggravating
4. Players choices

b) Engine? Founded super power balance? Cant answear to this question

Solutions:

a) 1. work on set up off course
2. no solution
3. have the courage to set up in the ai files war propensetion near to 0 and asking paradox for a alliance command that works
4. no solution off course

b) have the courage to add war command in every proper event regarding wars.

Don't know if this is against agceep philosophy, but i find it the only way to handle this flawns.
Dont' want to flame again a war around determism. All the proposed solution let a lot of indetermineted, the player will ever made some strange choice, the ai will ever chose some wrong option in events, or start some engine war, but in this way i'm sure we'll enjoy a game nearer to history with still the possibility to change history or with a history with a strange outcome. Further players will have a much more interesting game from the 16th century to the end, that is really not the case now. As stated by many person very few have the interest to go on playing after 15th century, the unansweared question is why?

 

[|AXiN|]

The way you handle the ai is to save and reload every once in a while, not by adding war commands.

 

[Toio]

The HC need to make a decision that since the AGCEEP is a more historical game, then options in events need to be minimized so that the AI and human do not have achoice, I suggest:

1. All good historical events to only have choice "action A"

2. All moderate and bad events to have either A, B or C with the A option being the most historical and the worst with the following choices tapering down to not as bad.

These hard decisions by the HC will stop all the bickering in respect to - a agceep is not historical.

The AGCEEP is being split because of these. An example is in the ottoman thread 1520 to 1820.

 

[Toio]

Well i playtested 2 GC:

there are 2 main flawns:

a) ai deviances on the political map

b) ai inactivity after 50-70 years game

Causes:

a) 1. innacurate initial set up
2. wrong choice in events
3. random wars-diploannexation with random alliances aggravating
4. Players choices

b) Engine? Founded super power balance? Cant answear to this question


b) have the courage to add war command in every proper event regarding wars.

On A1. its historical with the limitations we have in regards to the map.

A2. should only be one choice for every good event , that way you get a more historical game.

A3. Nothing can be done except fix the combat lists for the AI files as these trigger 80% of wars.

A4. Limit the choices for players by only have 2 or 3 choices in the bad events, BUT make sure all choices are bad.

What are superpowers as they change every 50 years in historical terms.

WAR command will not work if , the nations are allied or the declaring nation has a stability of -3 and a few more. It is hard to place WAR in an event.
The only one I did was in a venetian event against the OE , but this was easier to do because the OE and VEN are nearly always at war. Also note that the years the event is to fire will need to be stretched out as you might miss this event completed because of all the parameters that need to be met.

 

[LlywelynII]

Just as a sloppy fix until the Turkish minors thread gets some additions (hint.)

When 1490, 1495 rolls around and AKK is nowhere to be found, just

f12

event 65051 QAR
event 103003 QAR

& then if you see the REBs in control of Tabriz province (the map one, not the AGCEEP one)

event 65052 QAR

Bam. Instant Persia.

If it takes awhile to get some REB control going on, just 65051 a couple more times.

Works for any other uppity tag in the region as well.
jay.

 

[unmerged(10630)]

On A1. its historical with the limitations we have in regards to the map. .

Don't agree, read the posts above about Persia and Akk!

A2. should only be one choice for every good event , that way you get a more historical game. .

Definition of good event?
I'd been found guilty to want a deterministic game, but this is far worst. I think a little choice of deviance should stay.

A3. Nothing can be done except fix the combat lists for the AI files as these trigger 80% of wars. .

Engine wars are triggered by ai agressivity and by ai files. If all the ai files are near to peacefull there will be less chance of a random ai war.

A4. Limit the choices for players by only have 2 or 3 choices in the bad events, BUT make sure all choices are bad. .

Ok, but this chance little in historical outcomes by the ai countries.

WAR command will not work if , the nations are allied or the declaring nation has a stability of -3 and a few more. It is hard to place WAR in an event.
The only one I did was in a venetian event against the OE , but this was easier to do because the OE and VEN are nearly always at war. Also note that the years the event is to fire will need to be stretched out as you might miss this event completed because of all the parameters that need to be met.

It's not so hard as it seems, if we have an alliance command that work and all the alliances are historical over the whole game, events will shot in the right way. The only one that can change this is the human player chosing the wrong alliance.

The way you handle the ai is to save and reload every once in a while, not by adding war commands. .

I suggest you to make some more playtesting, cause this is not working in late game after 1520 in my games. The only result is adding more random wars, in the first 70 years of war.

 

[unmerged(29041)]

b) have the courage to add war command in every proper event regarding wars.

This is utterly nonsense. By forcing wars you do a lot more harm than good. As poor as it is, the AI knows a lot better than a blind event when a war is convenient or not. It is often the case that the supposed winner comes out a big loser because the historic time for the war was not a good game time for a war.

 

[Garbon]

Just as a sloppy fix until the Turkish minors thread gets some additions (hint.)

When 1490, 1495 rolls around and AKK is nowhere to be found, just

Ahem, the correct date would really be 1500/1501.

 

[unmerged(2456)]

This is utterly nonsense. By forcing wars you do a lot more harm than good. As poor as it is, the AI knows a lot better than a blind event when a war is convenient or not. It is often the case that the supposed winner comes out a big loser because the historic time for the war was not a good game time for a war.

Yes for every war this would be rediculous, but in the later games this might be apporpiate where very few non-human initiated ones start.

Anyway i agree we don't need 20 events firing every year (yea its a bit of an exageration), but the problem with EP is that if you go in an unhistorical route, while IDLF does a good job at making sure those historical events that would be rediculous to fire because they'd be inappropriate don't, he doesn't give much to the player in reguards to other events. People don't want events spamming them, but they don't want to play most of a game getting only random events either and he doesn't address that other side well.

But we in AGCEEP haven't addressed the event spamming perfectly nor have we made sure many historical events that seem rediculous in various situations still fire, even when there are triggers to stop them.

 

[LlywelynII]

Ahem, the correct date would really be 1500/1501.

The correct event is for AKK. If I'm in control of the Sufis, I'll have them revolt when I feel like it, ie before QAR already owns 2/3ds of the former Timurid empire.

jay.

 

[Garbon]

The correct event is for AKK. If I'm in control of the Sufis, I'll have them revolt when I feel like it, ie before QAR already owns 2/3ds of the former Timurid empire.

jay.

Oh dear, I forgot they had a tendency to do that.

 

[Toio]

I want to ask everyone.

Why do you want the Ai to be handicapped by not choosing the original "good" event 100% of the time "guatanteed".
That is , only have 1 option.

I find it extremely unfair that a human has a choice while the AI is handicapped by % choices for the AI in it's selection.

Same for all events really.

 

[Garbon]

I want to ask everyone.

Why do you want the Ai to be handicapped by not choosing the original "good" event 100% of the time "guatanteed".
That is , only have 1 option.

I find it extremely unfair that a human has a choice while the AI is handicapped by % choices for the AI in it's selection.

Same for all events really.

I'm not quite sure what you are advocating?

 

[Mad King James]

Unlike in CK, you can't set the AI to pick an option x% of the time, the first option is almost always selected.

 

[binTravkin]

I'm not quite sure what you are advocating?

He's opposing the idea to make AI events linear (thus making it behave more historically) thru having only the action_a, while human events would have alternative choices as well.

 

[binTravkin]

Unlike in CK, you can't set the AI to pick an option x% of the time, the first option is almost always selected.

Guys, do you really don't remember what one of you (HC) has once said:

Even if the percentage that AI will chose action_b is <= 5%, in event chain these percentages are proportionally cumulative, thus the chance that the AI has chosen the hisotrical path is:

chance = 0.85^ev

0.85 is an approximation of the 85% chance to select action_a.
In 2 action events it will be bigger, inb 3 exactly 85%, in 4 action events it will be smaller (as per Havard).

ev is 2+ (2 and more) action event count for the given country.

For example if an event chain has 3 events with 3 actions each, the possibility that the result will be historical is:

0.85^3 = 0,614125 = 61%
which is less than 2/3 and is not far from the critical 50% at which we should not talk about historic accuracy/precision as the chances for a historical result versus ahistorical one are equal.

 

[Mad King James]

I suppose I like a little bit of ahistorical things to happen in a game, as long as it isn't bizarre or impossible. For me, Castille not merging with Aragon is unfortunate but actually kind of interesting. It's Denmark seizing control of Constantinople that drives me nuts, or similar activities.

 

[binTravkin]

It's Denmark seizing control of Constantinople that drives me nuts, or similar activities.

Yeah, and it has nothing to do with events, but rather with the stupid (sorry but I really think so) person who decided to make AIs fleets immune to attrition.

 

[unmerged(29041)]

Yeah, and it has nothing to do with events, but rather with the stupid (sorry but I really think so) person who decided to make AIs fleets immune to attrition.

Programming intelligence is beyond the means of a small game company and is not justified by the return. Fleets not having attrition is a reasonable solution for an AI that is incapable of handling them. Otherwise the AI would be handicapped to the point of not exploring, not colonizing, not moving armies to other continents, etc... I don't think that would be better.

 

[binTravkin]

and is not justified by the return.

Yeah, that's the sad thing.

I watched a trailer for game for new PS3 platform and THAT was exactly what are you telling here.
I was simply disgusted when enemy soldiers stood as bars in bowling, just waiting to be hit.
Everything was perfect - graphics, sound, the legend. Everything but the AI.

I will never buy PS3 unless they fix it, so they can count my money out.
(and Im not the only one).