Only 1 leader possible from any given elite

[HellboundHeathcliff]

It seems the thing to do would be to put elites that have produced leaders into armies. Especially cavalry, which aren't going to get upgraded. A 20 HP cavalry army is no joke.

 

[Lt. 'Killer' M.]

Originally posted by watorrey
It seems the thing to do would be to put elites that have produced leaders into armies. Especially cavalry, which aren't going to get upgraded. A 20 HP cavalry army is no joke.

Exactly, but it will be quite tedious to keep track of them, keep slots in armies open for them and you`ll lack the army power until you get leaders......

So, this is a measure that`s fine if you happen to have free space in an army - like your first one....


Why not do it this way: elites that produce a leader get marked ELITE instead od Elite. From now on, they have double the military police power of normal units (i.e. two happy faces) or, in Govs without mil police, they cut corruption&waste by 10%, 3 Gold+1Shield minimum?????

That way, there`s still a good use for them.....

 

[sumthinelse]

Originally posted by Killer



From now on, they have double the military police power of normal units (i.e. two happy faces) or, in Govs without mil police, they cut corruption&waste by 10%, 3 Gold+1Shield minimum?????

You could have unique civ-specific elite units. Saddam Hussein's Republican Guards, U. S. Delta Force, Napoleon's Old Guard, and I don't even need to mention the historical German elite units

 

[sumthinelse]

Killer,

I loaded the game, built the herioc epic, attacked the nearest Persian city, won 50 victories with the army, no leader. I can, however, create a leader very quickly with the knights that ain't in an army. With the heroic epic the odds of no leader from a non-army elite victory would be 11/12 (are you listening, theniceone?) and so the odds of 50 tries/no leader would be about 1 in 77. However, I think the truth is that the probability is 0 in this case, that is, attacking with this army in this game.

You have a nice game going. *If* it is impossible for an army to create a leader you have "wasted" only one elite knight (the only elite unit you have put in an army).

Based on the evidence we have seen, my opinion is that Mike B.'s statement that armies don't create leaders is true *most of the time*. However, you may have created a leader with an army in one case somehow and we just don't know how to do it again.

Mike B., is it possible for an army (not the units in it) to be promoted to elite? Never saw one.

All computer problems are easy if you know the answer. :enlighten

 

[Lt. 'Killer' M.]

OK, first round is through, 100 tries, no leader from the army

I`m going to see a doctor


Then, 200 fights with elite knights not in army against Persian cities after breaking ROP.
Leaders on battles:
(none for 38)
39 (none for 2)
42 (none for 32)
75 (none for 8)
84 (none for 4)
89 (none for 1)
91 (none for 4)
96 (none for 20)
117 (none for 5)
123 (none for 2)
126 (none for 4)
131 (none for 12)
144 (none for 8)
153 (none for 7)
161 (none for 1)
163 (none for 16)
179 (none for 20)


so, what does this tell me? I guess the most surprising thing is the high number of very short intervals, like two times only 1 turn in between, three times 4. All in all, 8 instances where the wait was less than 8 turns to the next leader, 1 where it was exactly 8, and 8 where it was more. This sound more like a distribution for 1in8 with extra draught periods


Now i want to test whether the two really long periods at the very beginning might have something to do with breaking the treaty. So I`ll repeat the fights, then do it twice with HE, then do it after waiting for the ROP to run out.

 

[damunzy]

Originally posted by Killer
@ palehorse: Thanx

OK, here`s the game:

test game zipped

Will DL when I get home.

@ sumthinelse: East Coast USA.

 

[sumthinelse]

Originally posted by Killer


This sound more like a distribution for 1in8 with extra draught periods

Uh-oh, you reported a "leader drought" without posting it in the "leader drought" thread. I'll have to report this to the mods

If you built the epic the odds are supposed to be 1/12 and if you count the tests after the first 38 all your tests it comes out 1/9.4 which is not too far off.

Your longest "drought" was 38, but that was before you built the epic. Assuming that the odds of a leader were 1/16 then, the odds of that happening at any given point are 1 in 12, so that is kind of unusual and makes me wonder. But we really need a longer test at 1/16 odds.... 38 isn't very much.

OK, now after the first leader you built the epic I guess, and you had 141 victories at 1/12 odds? The odds of getting 2 leaders in 3 tries are about 1/72 so it's not surprising that happened twice in 141 tries. (If it was 1/16 per try, 2 leaders in 3 tries is about 1/128)

If you built the epic, the chances of getting a leader in 8 tries or less, believe it or not, is about 1/2. You would think the "break even" point would be 12 tries but it's not. So I think you shouldn't be so surprised if sometimes you get a leader in a few tries as often as you did.

Reasoning: each time, the odds of not getting a leader is 11/12, so the odds of not getting a leader 8 times in a row is 11/12 multiplied by itself 8 times = .4985, very close to 1/2. So, you should get a leader in 8 tries or less about half the time. Over a long period of time, of course.

But the average number of tries to get a leader should be 1/12. Sometimes it takes 20, 30, or more tries.

At 1/16 odds, you should get a leader in 11 tries or less about half the time.

 

[Tassadar]

Originally posted by Killer
OK, first round is through, 100 tries, no leader from the army

I`m going to see a doctor

You are not alone Killer, i said i had a leader from barbarian fight and its impossible. I was sure it happened once. I think after playing several night hour we imagine thing and later on we think it realy happen. Incense's fume doesnt help either .

 

[Lt. 'Killer' M.]

Originally posted by Tassadar


You are not alone Killer, i said i had a leader from barbarian fight and its impossible. I was sure it happened once. I think after playing several night hour we imagine thing and later on we think it realy happen. Incense's fume doesnt help either .

nor does work.... 3:0 a.m. here and I could swear I just saw the Perisan turn heel when I started producing Knights in every city... though i know from a lot of testing that the AI doesn`t soon-to-be-finished units into it`s strength calculations..... yaaaawnnnn!

 

[sumthinelse]

Killer,

I assumed that you built the epic wonder after the 1st leader in your detailed leader results above, because it looked like you were getting leaders a lot more often then 1/16 after the 1st one. Then I looked at my records of creating leader when the odds were supposed to be 1/16, and I found that I had created 13 leaders in about 120 tries without the epic wonder. That's more like one leader in 10 than one in 16 So I'm starting to wonder whether those results you posted, where you got one leader in about 10 tries, were done with no epic wonder, so you should have gotten about 1 in 16 but you got 1 leader in about 10 tries overall, which was what I got. Maybe the program knows that knights are "heroic" so they must generate lots of leaders

I also installed the civ3edit multitool by Gramphos to see if I could figure out what internal flag in the save file was for "elite created." If I could figure that out it would help me to tell which elites have been "used." Unfortunately, the file structure for units in the save file is pretty huge, with flags and stuff all over the place, so it will take a while. I'll tell you how to find the unit structure in the file if you are interested. I don't know what most of the stuff means in the unit structure.

 

[Lt. 'Killer' M.]

sumthinelse, that`s exactly what I thought strange: I did not build the Heroic Epic!!!!!

That`s why I said 1 in 8 plus maybe dry speels for treaty breakers. In my second try, I again got lond dry stretches early on! But it definately isn`t 1 in 16!

Now I`ll try a game with seed on and get the results ofr it and I`ll repeat that until I have 200 elite wins with unused elite - just to compare.

But I always said the RNG is too streaky for the game - I had the feeling I had an excellent stretch in my tries. I`ve also seen games where I had horrible stretches for awfully long times..... So I`ll do the test game over and over, also create others and test them over and over until I have an amount of data that is statistically significant.

Thank you for taking all this time for the leader thing!!!!

 

[sumthinelse]

Killer, you and I together have 29 leaders in 300 elite victories before the epic wonder. The odds are supposed to be 1/16 but our combined results are about 1/10.

Then I built the epic wonder, and now leaders are harder to get so far! Out of 198 victories I have 11 leaders, so that's 1/18 when the odds are supposed to be 1/12. What's going on here? Did the programmers get the odds reversed for before and after the epic?

 

[Lt. 'Killer' M.]

Wow, you get that, too? I had 16 out of 220 with HE!

 

[sumthinelse]

Originally posted by Killer
Wow, you get that, too? I had 16 out of 220 with HE!

Killer, are you staying up late again wasting your time on this silly game?

Note to new readers: all fractions below represent number_of_leaders_produced/elite_victories.

After you built the epic, you got 16/220, or about 1/14 -- half way between 1/12 (the alleged odds for after the epic) and 1/16 (the alleged odds for before the epic). That's not so different from 1/12. However, before the epic you got 16/179, or about 1/11.

--------------------------
SUMMARY

If you combine our results before the HE it's 29/300 or about 1/10. Should be about 1/16.

If you add up your results (16/220) and mine (11/198), both after the HE, you get 27/418, or about 1/15. Should be about 1/12.
--------------------------

Isn't this the kind of results we would expect to see if the programmers accidentally reversed the odds?

Mike B., does it look like the odds 1/12 and 1/16 are applied under the correct conditions (before/after Heroic Epic) for generating a leader?

I started some testing under the exact same conditions (same attacker/defender/locations/date) but without the HE. Will tell you what I get.

 

[Lt. 'Killer' M.]

hehe, sure i do stay up, but actually I`m wasting my time on my diploma thesis and the new FAQ thread

I have another baffling sequence for you:

10 leaders in 78 turns with HE, then none for 44. All in all, getting 16 again out of this time 200 battles......


I`m looking forward to hear from Mike!

 

[roadwarrior]

You may have posted this before and I missed it, but what civ are you using? Are they militaristic? I know that makes you get promoted faster and I thought that it made you get more leaders too.

 

[punkbass2000]

Originally posted by roadwarrior
You may have posted this before and I missed it, but what civ are you using? Are they militaristic? I know that makes you get promoted faster and I thought that it made you get more leaders too.

Militaristic does not increase the chances for leaders.

 

[sumthinelse]

Originally posted by punkbass2000


Militaristic does not increase the chances for leaders.

That seems to be the case, but we are trying to figure out what *does* make leaders more likely/less likely. The evidence so far suggests that the Heroic Epic makes leaders less likely (according to the books it's supposed to make them more likely), but after more tests we will have a better idea.

Killer is using the Babylonians (scientific/religious). I think those characteristics apply to Killer personally but he is also militaristic :soldier:

 

[punkbass2000]

I seriously doubt the HE makes leaders less likely. This is just a statistical anomaly, like when a spearman beats a MA. IT will even out eventually, or if include everyone's battles.

 

[sumthinelse]

Originally posted by punkbass2000
I seriously doubt the HE makes leaders less likely. This is just a statistical anomaly, like when a spearman beats a MA. IT will even out eventually, or if include everyone's battles.

It may turn out that you are right, but through emperical means we have already discovered a few surprises about leader generation. Keep an open mind and let the evidence determine the truth.