GK2- The Training Day Experiment

I still vote for a MW Army. It will be fun to play with in the upcomming battle, and they are always usefull. Even in the late industrial era a 12-15 hp unit is great to have around, if for nothing else pilaging tiles during a war. I've had crusader Armies in the Modern Era that I still forund usefull. Keep in mind I think that we can add a unit once we build the pentagon, if we do enough war mongering to get 3 Armies and then we could always add a calvary unit to it. We will have plenty of oppurtunities to get another leader in the rest of the Zulu and Persian wars. If we have to build the FP brick by brick so be it. Rush a courthouse and start building.

My .02 cents

Alerum
 
A MW is 2/1/2, right? IRRC, Armies in Vanilla don't get an extra movement point. At this point, MWs will be great against spears and immortlas, but we're crossing into the era of pikemen and knights. A MW army will, IMHO, have the shortest useful life. If we were fighting against overwhelming odds, I'd say go for it. As it is, we have far more power and support then we'll need to crush the Persians. An army would be a waste of a leader.
 
Building a MW army will be useful for about 20-30 turns, if that. We have pretty much killed off all our oppopents on this contient. If we build an army, and fill it up, it will be stuck on this contient, doing nothing but MP duty, or border patroling. Also, it seems like you guys keep confusing what an army can do in C3C and what it can do in Vanilla. we can pillage, but it'll cost use a movement point, unlike in C3C where you can pillage and move, pillage and move....

Now, rushing the Great Lighthouse is a good option, but Bugsy gave us a good heads up that we may have discovered a channel to the other contient. This game hasn't needed one wonder to be built, and I doubt that MB would make the channels so they'd need the GL. They weren't counting on us waiting til Carvels to meet the other side.... they wanted suicide galleys to show us a human strentgh that the AI doesn't have.

Waiting on the Forbidden Palace... Like we've discussed, we aren't millitristic, which means if we pull even one more GL we'll be lucky, and that's taking in to account using the GL up ASAP. Now, we're playing Vanilla which means building the FP means a second core, not just reduced corruption in one single city. We won't be able to use the GL for about... 15 turns or so, at which point we should have control of Susa and have the Persians subdued if not elimited.

I honestly think, for the overall picture that rushing the FP is the best use of our GL. Sure, we can create an army, and load some MWs into it, and be able to build the Heroic Epic, but what good will that do us? I don't see to many more battles down the road... not enough to warrant building the Heroic Epic... trust me, I always used to build that damn thing, and unless I was Germany, never got a thing out of it.

We're heading for the Modern Ages... we're not searching for an early conquest by Knights game... The weakness of armies on Civ3 compared to C3C is dramatic... Building an Army, especially a MW army would be a huge waste... in time and in shields.
 
I believe that Coletite was voting for the FP as well... SK hasn't been heard from in about a week... I had to skip him in another SG. He has alot going on right now though, so don't be suprised if he just pops up in a day or so.;)
 
So...many...ideas! :crazyeye: Good discussion, gang :)

@ TimBentley: Thanks for the catch. I'm playing C3C when I'm not looking at this game, and my rules are getting all mixed up as a result. Even so, though, unless we do get an extra movement point, it would still be more valuable to have three individual MWs for skirmishing than one army that can attack twice.

Can units in an army promote individually and produce GLs? I think the answer is "yes" to promotion and "no" to GLs, but I'm not certain at all.

--

I tried to do a statistical analysis of how likely we might be to end up with another leader out of the Persian war under certain conditions. I have a spreadsheet full of numbers, but I'm still lost trying to put it all together and figure out what it means :)

Originally I thought our chances would be pretty crappy, but now I think they might be okay, as long as we are careful about racking up victories with elite units. By the 22nd victory with an elite unit, the probability of having at least one great leader come from elite combat is 75%.

The problem is that we don't have any elite MWs yet, so we will have to get some before we can expect them to promote. (Diverting swordsmen from the other side of the map is bound to take too long to make much difference.) Our chances here are mostly based on how many troops Xerxes has for us to beat up.

The spreadsheet (which is actually kind of nifty, because it's the first really practical application I've seen of the Binomial Theorem - but I digress) can tell us a couple of things. It first confirms that if we want a certain number of elite promotions, we're most likely to get it by fighting (8)(that number of promotions) battles. That makes sense given that the chance of a veteran->elite promotion is 1 in 8.

If we can estimate Xerxes' military size, we can also use the sheet to figure out when it would be best to stop fighting with veteran units and let the elites shoulder the load.

If memory serves me right, we can send that scout parked south of Panama into Xerxes' territory, and, since the scout is non-combat, not have to leave when asked. That might be one way. Or we could investigate a city or two and guesstimate. Either way, the data is pretty cool when you have a number of fights you're expecting to win :D

Anyone who wants to see the spreadsheet or the positively psychedelic line graph it produces is welcome. I'll post a screenshot if there're any requests.
(edit2: actually the line graph looks kinda lame now, but at least it's useful this time. had the wrong series in the wrong places)

--

Edit: Crossposted with a whole bunch of people :) My sense from these numbers is that Alerum is right and it's not very likely that we will get an extra leader.

My vote is indeed locked in for the Forbidden Palace.
 
Alerum: There may be a suicide path to the other continent which is great for contact, but if we are going to be sending units over there, we don't want to risk them to the open seas. To think about it another way: if someone "over there" builds the lighthouse and decides they don't like us, they will be able to send units over to us, but we won't be able to, safely, get to them.

Another point- I wouldn't get caught up in trying to second guess what MB had in mind with the map. We need to learn to play the game as though we had NO information.

True, the FP is powerful, but I'm willing to bet on us getting a leader from the Persian war. [edit] We still have four ZUlu cities to capture as well which means more fishing opportunities.

My vote: Lighthouse
 
Leader Vote

Mistfit: army/or GLight I just don't want to be fighting with no possibliity of getting another Great leader
Alerum: FP
GJ: FP
Coletite: GLIght
SK: is SK still with us?
Edit: changed some stuff/ added commentary
Well At least we all agree on this one :lol:
 
Coletite was the first to pick up on the fact that a victorious army will allow you to build the Military Academy, and he (correctly) notes that an Army costs about as much as a small wonder. He also recognizes that for a spaceship win you need "lots of productive cities..."

This is going to sound like an odd question, but how might a Mounted Warrior Army help you win the Space Race? :mischief:

@Team: Please do not infer that I think building the FP with your MGL is not a strong move; I think that argument is compelling. I simply want you guys to understand everything you are giving up... and there is one thing I see that you are not considering. (If you consider it and discount/disregard it, that's fine... as long as you consider it.)

I'll make this easier for you:

What is the cost of an Army compared to the most expensive Spaceship Part, and how can you use this to advantage?

----------
This next part is mostly for GJ...or anyone who loves stats....

@GJ: I would very much like to see your spreadsheet, but I have some cautionary notes.

First - make sure you account for the combat calculator basics: The odds of getting a MGL from an elite unit are 1:16 when the elite wins. An elite MW versus a veteran spear (fortified in a town with no other defensive bonuses) will win 69.2% of the time. Thus your overall odds of getting a MGL leader from that particular combat are something like 4.3%...

These odds would imply that you could expect a MGL from roughly 23 Elite MWs... but...this represents the "central tendency", and is by no means iron-clad. If you were to press 46 Elite MWs into combat against Vet spears, you could expect 2 MGLs. You may very well get more, or you may very well get none. As you repeat the events more and more times the outcomes will resemble "regression towards the mean" and you will observe a distribution of outcomes that resembles the odds...

...but just as you cannot bet the bank on odds that reflect "central tendencies", you must also be careful of the "Random Walk". The "killer spearman" smiley is not merely a CivIII "urban legend". If you attack enough spears with enough tanks, you will eventually lose a tank to a spearman; I have.

Some more food for thought... (and why a model based on binomial theorem may not quite work...)

The RNG in CivIII seems to do it's level best to keep each probability event "isolated"...

Consider gambling, the probability of someone rolling a "7" on a craps table seven times in a row is something like .00035 (or three one-hundredths of a percent). Slim odds. Ususally the chips start flowing towards the "Don't Pass" on about the 3rd or 4th roll. The fallacy here (as with most gamblers who claim to have a "system") is that each event (roll of the dice) is an isolated, random event.

In other words, the odds of that next roll being a 7 are still 16.667%, no mater how many "Sevens" preceded that roll. (Having said that, I recognize that the RNG in any computer program cannot be purely random....)

Edit - I must have confused Binomial Theorem with something I tinkered with using either a causal model or related events (time-series analysis). Binomial Theorem should be able to describe isolated "random" events...I've been out of school for a few years, and I'm a little rusty on some of this stuff...
 
scoutsout said:
----------
This next part is mostly for GJ...or anyone who loves stats....

@GJ: I would very much like to see your spreadsheet, but I have some cautionary notes.

First - make sure you account for the combat calculator basics: The odds of getting a MGL from an elite unit are 1:16 when the elite wins. An elite MW versus a veteran spear (fortified in a town with no other defensive bonuses) will win 69.2% of the time. Thus your overall odds of getting a MGL leader from that particular combat are something like 4.3%...

These odds would imply that you could expect a MGL from roughly 23 Elite MWs... but...this represents the "central tendency", and is by no means iron-clad. If you were to press 46 Elite MWs into combat against Vet spears, you could expect 2 MGLs. You may very well get more, or you may very well get none. As you repeat the events more and more times the outcomes will resemble "regression towards the mean" and you will observe a distribution of outcomes that resembles the odds...

...but just as you cannot bet the bank on odds that reflect "central tendencies", you must also be careful of the "Random Walk". The "killer spearman" smiley is not merely a CivIII "urban legend". If you attack enough spears with enough tanks, you will eventually lose a tank to a spearman; I have.

Some more food for thought... (and why a model based on binomial theorem may not quite work...)

The RNG in CivIII seems to do it's level best to keep each probability event "isolated"... (snipped)
Click to expand...

All fine points :)

The distribution is actually quite attractive to look at, assuming you like to look at tables full of numbers.

The distinction here is the number of units versus the number of victories. I found the percent chance of having at least one leader given a certain number of victories; by the time I figured out how to do that I needed to go to bed :) I'll do some extra analysis this afternoon to figure out how much we might expect out of a certain number of UNITS. Hopefully that'll be less of a pain.

Especially relevant is the "random walk" point - we shouldn't forget runs like that...statistics are likely to go out the window in practice :) But I am willing to make and use any analysis that will help us come to a decision.

Math point for scouts: I'm with you all the way on the independent events, and that's the analysis we would use to find, say, the probability that we would get NO leaders, or that we would get at least one or two or however many leaders. But I think the binomial theorem is necessary to get *exactly* a certain number of leaders. (It took me a while to figure out I would need that analysis for the numbers I wanted.)

I'll post the spreadsheet's information as soon as I finish tooling with it and get some screenshots.

(Edit: Cross-posted with scouts' edit. Don't ask me why the Binomial Theorem works here, I only know that it does :D that unit on sequences and series seems so far away...)
 
Scout: Are you thinking in terms of disbanding the army in a city building an SS part? I can see the value of that in having two cities working on the wonder (one building armies).

Maybe the real question is, what is our long term strategy? I know that you want us to fight a lot of different wars to get familiar with the different units, so there should be plenty of opportunities for fishing. In fact, I think we should go just to the domination limit by the time we have Modern Armor.

Another reason to use leader to rush the lighthouse: Right now, we are just moving ahead in terms of production. By the Modern Age we should have a considerable lead and not have to worry about rushing wonders.
 
Apologies for doubling up, but I didn't want throw all this into my last post :)

Here is a picture of the data analysis page that shows the likelihood of getting N elites from M victories.

vet_promo_1.jpg


Apologies for the small print.
The column on the left is the number of victories you expect to have. The row along the top is the number of elite units you desire. Each individual cell gives the probability that, given the number of victories, you will get that particular number of veteran-to-elite promotions.

(Math note: The #NUM errors come from trying to take the factorial of a negative number. PM me if you really want to know :))

The rows add up to a probability of 1, as they should, which makes me feel better about the math involved.

--

I did this first because I was worried about the number of veteran units we had in Panama. The formula is easy to change, though - replace 1/8 by 1/16 and 7/8 by 15/16.

Here is a link to the original spreadsheet file, in case you want to play with it. Disregard the stuff on Sheet2 and Sheet3; it doesn't do anything interesting - yet...
 
@GJ: The stats are a fun little intellectual excercise... just remember, in the game you don't get "n" trials, you get a finite number of units and a tactical problem. ;)

@Team: Don't forget to read SirPleb's article on Creating and Using Leaders.

@Coletite: I'm not talking about disbanding anything - I'm talking about prebuilding. When you get to the Modern Era, you don't have very many options for high-shield cost builds. Armies make nice pre-builds for stuff.
 
coletite said:
Scout: Are you thinking in terms of disbanding the army in a city building an SS part? I can see the value of that in having two cities working on the wonder (one building armies).
Click to expand...

No, I think he is referring to the fact that a city with the Military Academy can use a new army build as a pre-build. This is a nice way of allowing a strong shield producing capitol to prebuild items, since it can not prebuild the Palace.

EDIT: cross posted with Scouts reply
 
scoutsout said:
@GJ: The stats are a fun little intellectual excercise... just remember, in the game you don't get "n" trials, you get a finite number of units and a tactical problem. ;)
Click to expand...

Sad but true ;)

Not only will we need elite units, we'll need to be winning with them. It will take a lot of wins just to get a couple of elite units (if the probabilities hold or come close to it, anyway), and then those elite units are going to need a fair number of victories...I'm just not convinced that Persia has enough units for us to beat, or that we even want to spend the resources necessary to get that number of wins.

So I'd just save the leader for the FP.

(Or I could find something lucky before taking my turns. hmm cd of Mahler's Second ;))
 
D'oh!! I used prebuilds in CGOTM for SS parts, too! It's saturday, leave me alone! :crazyeye:

So yes, Military Academy does have its uses, but I think that is too far away to think about. I am willing to bet we will get at least one more leader before then.
 
coletite said:
It's saturday, leave me alone!
Click to expand...
Now that you guys understand my original question, I think I will give you a break. At least until I come up with another question....

Bridge_of_Death.jpg


Seriously - there are a lot of tradeoffs to be made in any Civ game. Like I said before, I think the case for using the Leader to rush the FP is a strong one. Building the Great Lighthouse also has some advantages. I simply wanted you guys to see all of the implications of not building an Army with that first Great Leader.

If the thought process made your heads hurt a little, tough cookies. :p
 
Mistfit said:
Oh good lord Algebra... I think I'll go take a nap (just like I did in school)

Please remember in the up comming war with persia we will be getting into our Golden age so self building the F.P. may not be as bad as you think.
Click to expand...

I just got up from my nap - It doesn't help - still too many numbers and I'm a CPA.

One thing that hasn't been discussed recently is building the FP in your core (especially during the GA) and using the Leader to rush a new palace. There is slightly more area to settle in the old ZUL holdings and some cities yet to be captured could be razed to help with RCP placement. I think the added corruption in our first core would be a minimal increase.

coletite said:
I am willing to bet we will get at least one more leader before then.
Click to expand...

Don't always plan on another one. Algebra is bad but statistics are worse. I have actually played entire epic games without getting a GLeader. In GOTM29 playing as Egpyt, on a Pangea map and warring slowly throughout the game, I did not get my first GLeader until 1842AD.

TDG_First_GL.jpg
 
You must log in or register to reply here.

Similar threads

LKendter
LK173, Civ=Scandinavia, CCM3, Island Hopping
21 22 23
Replies
451
Views
18K
LKendter
LKendter
jorissimo
Jorissimo's opening training shadow thread
3 4 5
Replies
95
Views
5K
jorissimo
jorissimo
P
4.17 + 4UC Arabia Emperor
Replies
11
Views
579
azum4roll
A
Brew God
Alliances?
Replies
2
Views
595
MACKA0ili
M
Nizef
TSG 265 Announcement
Replies
6
Views
999
Ecocrexis
E
Back
Top Bottom