promotion probabilities

[WarlordMatt]

I wonder if the probability of a unit being promoted may also be affected by the power of a unit, like if a spearman defeats a swordsman would it have a greater chance of being promoted than if it beats a warrior?

 

[Lt. 'Killer' M.]

the hut is opened BEFORE the city is there. You can see that in the graphics, too!

 

[sumthinelse]

Originally posted by Mike B. FIRAXIS


It's not really that it's difficult; it's just that it was implemented by an intern . Here are the conditions:

................................

I know it's just a game, but I am very pleased that you were nice enough to provide this info!

I was thinking of a metaphor in which your post was a goody hut, and I opened it up and it said. "The Firaxis Tribe has taught us Goodyhutology."

Wow, I am just amazed at your post in several ways.

Have you ever met Sid M.?

 

[sumthinelse]

Originally posted by Mike B. FIRAXIS


Settler:

*Player must not have a settler (active or in production) or any unit with the Settle AI strategy.
*Player must have less cities than (TotalCities / NumActivePlayers).

Does TotalCities include cities the player cannot see?

Does NumActivePlayers mean total civs or only the ones the player has met?

OK, suppose I have an expansionist civ and the difficulty level is deity. I think that if I am still in the ancient tech era the probability of getting a tech from a hut is about 25%. But if I advance to the middle ages and cannot get techs, will I get 25% more "deserted village" results or will the 25% get divided among the remaining possibilities (deserted, gold, settler, warrior, map) with 5% to each or some other division?

 

[Mike B.]

Originally posted by sumthinelse
Does TotalCities include cities the player cannot see?

TotalCities is the total number of cities currently on the map, whether you can see them or not.

Does NumActivePlayers mean total civs or only the ones the player has met?

NumActivePlayers is the total number of civs that have not been eliminated, whether you have met them or not.

OK, suppose I have an expansionist civ and the difficulty level is deity. I think that if I am still in the ancient tech era the probability of getting a tech from a hut is about 25%. But if I advance to the middle ages and cannot get techs, will I get 25% more "deserted village" results or will the 25% get divided among the remaining possibilities (deserted, gold, settler, warrior, map) with 5% to each or some other division?

The chances for each possibility do not change. However, if something is selected that does not meet its requirements, a new selection will be made until something that meets its requirements is selected. For example, if you are in the Middle Ages, the chances for a tech to be in a goody hut are exactly the same as they were in the Ancient Times. If a tech is selected, however, a new selection will be made. That's why the probabilities aren't as easy to determine.

 

[Evincar]

Hey Mike, what you say is equivalent as dividing the 25 % chances of getting a tech among the remaining posibilities proporcionally to their former chances. For example:

If you have 25 % tech 50% deserted, and 25% gold, and tech is not available, the new chances are 66% deserted, 33% gold.

Maths are nice

 

[sumthinelse]

Originally posted by Evincar
Hey Mike, what you say is equivalent as dividing the 25 % chances of getting a tech among the remaining posibilities proporcionally to their former chances. For example:

If you have 25 % tech 50% deserted, and 25% gold, and tech is not available, the new chances are 66% deserted, 33% gold.

Maths are nice

Yes, suppose that there is a baseball game. For simplicity, I will eliminate balks and the pitcher hitting the batter with a pitch.

The batter has 3 balls and 2 strikes. This pitcher and batter are extremely consistent (the pitcher never gets tired) and the probabilities are:

strike: 30%

ball: 20%

foul ball: 50%

And of course, if the result is a foul ball, they do it again. If the result is a ball, the batter walks, and if a strike, the batter is out. In this example, a foul ball would be like an "impossible" goody hut result like barbs when you are expansionist.

The probability of an eventual strike are:

.3+(.3*.5)+(.3*.25)+(.3*.125)... which is .3*2=.6

An eventual ball: .2+(.2*.5)+(.2*.25)... = .4

In this case, a strike has 60% of the possible outcomes (30% is 60% of 50%).

So whatever a given outcome's share of the possible outcomes on one trial, that outcome should get that share of 100%.

 

[sumthinelse]

Originally posted by Mike B. FIRAXIS


The chances for each possibility do not change. However, if something is selected that does not meet its requirements, a new selection will be made until something that meets its requirements is selected. For example, if you are in the Middle Ages, the chances for a tech to be in a goody hut are exactly the same as they were in the Ancient Times. If a tech is selected, however, a new selection will be made. That's why the probabilities aren't as easy to determine.

Mike, thanks. this is very useful.

I think this explains why expansionists do so well with goody huts on deity level. I think the barb probability for non-expansionist is about 50% on deity, so not only does that double the probability of getting a tech for expansionist, but you would think it would make the non-expansionists less eager to open a hut, so more huts will be left for the expansionists. I don't know if the AI is programmed to be afraid of goody huts on higher levels, but human players are more reluctant to open them on higher levels.

 

[sumthinelse]

Originally posted by Mike B. FIRAXIS



Settler:
*Player must not have a settler (active or in production) or any unit with the Settle AI strategy.
*Player must have less cities than (TotalCities / NumActivePlayers).

What if totalcities=12, numactiveplayers=8, and I have only one city. Do you round the ratio up to 2 or drop the fraction to rusult in a ratio of 1?

 

[kring]

Bump, since sumthinelse had a good question, that hasn't been answered.

 

[pompeynunn]

What I found really interesting is the fact that you have to have a military unit for barbs to appear. Therefore, by only making settlers and workers at first, you can open those goody huts without worrying (leaves you defensively vulnerable, so not sure if that's a viable strategy or not).

Also found it interesting that a city one tile from a goody hut will not give barbs - means that settling next to the goody hut becomes a viable strategy.

 

[Mike B.]

Originally posted by sumthinelse


What if totalcities=12, numactiveplayers=8, and I have only one city. Do you round the ratio up to 2 or drop the fraction to rusult in a ratio of 1?

The fraction is dropped so the ratio would be 1. This would make settlers a viable option for this example. (The player has to have less than or equal to the ratio -- I failed to mention the equal part in my original post).

 

[kring]

Thank you Mike.

 

[alpha wolf 64]

Originally posted by Mike B. FIRAXIS
Oh yeah, that's one little tidbit I forgot to mention. If a victorious unit fails to get promoted, it will always be promoted upon surviving any other battle in the same turn. This gives an obvious advantage to defenders as well as an incentive to use the same offensive units for additional battles whenever possible.

Define same turn. If i survive an attack but no promotion, then attack with that unit, shouldnt it get promoted if it wins the attack? I havent been able to figure out the attacked and attacked with rule.

 

[Hurricane]

Originally posted by alpha wolf 64
Define same turn. If i survive an attack but no promotion, then attack with that unit, shouldnt it get promoted if it wins the attack? I havent been able to figure out the attacked and attacked with rule.

Good question. The intuitive answer would be no, since the human player always moves first (and it would thus be a new turn before you get the chance to attack), but my experience is that these units very often actually will get promoted.

 

[kalkas]

Originally posted by Mike B. FIRAXIS


Settler:
*Player must not have a settler (active or in production) or any unit with the Settle AI strategy.
*Number of player's cities must be <= (TotalCities / NumActivePlayers).

Could you please tell me which units have the Settle AI strategy? Thanks for your helpful post.

Regards,
Kalkas

 

[Isak]

In the standard game, only Settlers have the Settle AI strategy. I think Mike thought of modders when he created this "guide".