I was the person AlanH was talking about. All I can say is that Alan's spreadsheet worked perfectly for me. It made the prediction of the exact year of my 20k victory. So as far as I am concerned it worked perfectly. We even backtracked and check a couple of old games.
Cultural 20K Calculator (uses excel)
- Thread starter Greebley
- Start date
I've checked my current game up to the end of Spoiler 2 and Alan's spreadsheet is the one that was right. It was spot on 
, once I discovered the length of my Anarchy period (must keep a note of that next time!).
, once I discovered the length of my Anarchy period (must keep a note of that next time!).
Thanks for the confirmation. Glad to hear it worked!Originally posted by Dianthus
I've checked my current game up to the end of Spoiler 2 and Alan's spreadsheet is the one that was right. It was spot on
Alan,
I finally got a chance to test it on my completed game and it was completely accurate for me too.
I really think you should post your version in the utilities directory. Its a great utility. There is one joke wonder in there (based on a post by peanut I think it was). I would change it to other or custom and perhaps move it to the bottom before submitting it formally.
It might also be worth mentioning in the help that using F5 lists all culture items built so far and the date that it was really built. It is a good way to find the dates you need for already built items.
I am glad you had the time to do this and remove the clunkiness of my version.
I finally got a chance to test it on my completed game and it was completely accurate for me too.
I really think you should post your version in the utilities directory. Its a great utility. There is one joke wonder in there (based on a post by peanut I think it was). I would change it to other or custom and perhaps move it to the bottom before submitting it formally.
It might also be worth mentioning in the help that using F5 lists all culture items built so far and the date that it was really built. It is a good way to find the dates you need for already built items.
I am glad you had the time to do this and remove the clunkiness of my version.
Thanks for your corroboration. I guess we have enough evidence that it works now. I'll make the changes you suggest and post it there. (I do hope Sir Pleb doesn't get upset - given his Gotm performances it'd be very scary to get on the wrong side of him).
I would emphasise that I only tidied up your idea, and I'm better at implementation than having the ideas in the first place. So if you have any more ideas, I'll be happy to have a crack at them
SirPleb
Shaken, not stirred.
!I think its great - your spreadsheet doesn't need macros enabled, has a different feel, the more tools the merrier!
Thanks for your support, Sir Pleb. Rather than start a new thread with a similar name I've appended it to your thread in Utilities.Originally posted by SirPleb
I think its great - your spreadsheet doesn't need macros enabled, has a different feel, the more tools the merrier!
BTW, I think the discrepency that Dianthus raised may just be the one you mention in the notes at the bottom of your spreadsheet. Your calculation has compensated for the fact that a building starts accumulating culture in the between-turns when it's completion is announced, so the player must enter the date of the following turn in your tool. In mine I decided it would be better to require the previous turn date that Civ3 records in F5. If you enter the same date in both tools, as Dianthus did, you will get different answers.
I've changed the version to 6, but the changes are only in the Help, to add some color to the lists, and to provide a complete shield cost and culture per shield listing. I'll replace the links to older versions in my previous posts to point at my new post.
Mystery solved, I think. In Sir Pleb's spreadsheet he expects you to enter the date after your building completes, so doubling occurs earlier to compensate for the fact that Civ3 counts 1000 years from the previous turn. In my spreadsheet you have to enter the turn dates as recorded by Civ3. So if you enter the same date into both spreadsheets you are telling them that the improvement completed in different turns. Hence the different answers.
I prefer my approach because you can get the dates directly out of Civ3's F5 screen and in any case it's more consistent with the way Civ3 handles dates and is therefore easier to get right.
Note that Civ3 is actually self-consistent. A building, and any other task, completes at the end of your turn, not at the beginning of your next turn. When you have "1 turn left" before completion of a wonder at the start of your turn you know you will snag it because you will complete it at the end of the same turn, and before any other civ's turn comes around. This is also the reason you get the benefit of a terrain improvement between turns.
Alan,
I tried looking for your culture calculator (version 6) and it was not there (I used the link in your sig). This is the second broken link I have found of a file posted recently. I wonder if a bunch of files have gone missing...
Looking at site feedback this is in fact the case. Files were failing to load with no feedback on the problem (They were all size 0 according to the thread). You will probably need to upload it again now the problem is fixed. (Might want to double check it though).
I tried looking for your culture calculator (version 6) and it was not there (I used the link in your sig). This is the second broken link I have found of a file posted recently. I wonder if a bunch of files have gone missing...
Looking at site feedback this is in fact the case. Files were failing to load with no feedback on the problem (They were all size 0 according to the thread). You will probably need to upload it again now the problem is fixed. (Might want to double check it though).
SirPleb
Shaken, not stirred.
Ok, thanks.
In case sometimes people don't read past the first post in the thread I've modified that post to alert them to there being two spreadsheets there
SirPleb
Shaken, not stirred.
It sure is a confusing thing in the game. It always feels to me like improvements are completed at the start of the next turn, because the completion messages are displayed after everyone else moves, i.e. after our turn ends, they get their turns, then we get our turn and improvements complete.
To add to the confusion, if you finish e.g. a spaceship part or the UN in turn X, the win date is recored as X+1, reinforcing the feeling that builds happen at start of turn vs. end of turn.
Of course you are right in your description of how builds do happen at end of turn in many important respects.
Anyway, I thought people might find it more natural to enter completion dates of when improvements feel like they complete vs. what is shown in the F5 screen, so I made the opposite choice. A shame it isn't completely obvious what the sequence is in all respects!
@Sir Pleb (I can't believe I'm debating Civ3 points with the legend himself 
)
I think the order is actually:
As our turn ends the improvements complete, they get their turns, then we get our turn and we then have to choose what to work on in the cities where they need new build orders.
It's a subtle difference that isn't made clear in the dialogs. The spaceship and UN dates can be explained with a little guesswork:
When you complete the last spaceship part it has to get to the Capital to be assembled and launched. I guess that happens at the start of the following turn. After you complete the UN you are asked if you want to hold the vote at the start of your next turn. I guess it has to convene first, and the AI attitutes have to update in response to your actions during the completion turn. It wouldn't really make sense for you to be able to complete the UN, nuke every civ and win a vote, all in the same turn
No doubt the problem will continue to surface, and we'll be able to make a career out of dealing with the difference in our turn counting methods
Also, see below - I've had to add another post to your thread with the file attached. So it's now the 6th post, not the 5th.
@Greebley: Good catch. Thanks.
The file was, indeed there when I posted - I always copy/paste the url from the upload directory listing, and I happened to still have that browser window open today when I saw your post. The file was zero K long then and has since disappeared
. I then couldn't upload it at all using "Upload File", so I've made a new post and attached it (you can't add an attachment by editing an existing file). The new post and attachment seems to work and I've amended my sig to match.
)I think the order is actually:
As our turn ends the improvements complete, they get their turns, then we get our turn and we then have to choose what to work on in the cities where they need new build orders.
It's a subtle difference that isn't made clear in the dialogs. The spaceship and UN dates can be explained with a little guesswork:
When you complete the last spaceship part it has to get to the Capital to be assembled and launched. I guess that happens at the start of the following turn. After you complete the UN you are asked if you want to hold the vote at the start of your next turn. I guess it has to convene first, and the AI attitutes have to update in response to your actions during the completion turn. It wouldn't really make sense for you to be able to complete the UN, nuke every civ and win a vote, all in the same turn
No doubt the problem will continue to surface, and we'll be able to make a career out of dealing with the difference in our turn counting methods
Also, see below - I've had to add another post to your thread with the file attached. So it's now the 6th post, not the 5th.
@Greebley: Good catch. Thanks.
The file was, indeed there when I posted - I always copy/paste the url from the upload directory listing, and I happened to still have that browser window open today when I saw your post. The file was zero K long then and has since disappeared
. I then couldn't upload it at all using "Upload File", so I've made a new post and attached it (you can't add an attachment by editing an existing file). The new post and attachment seems to work and I've amended my sig to match.
SirPleb
Shaken, not stirred.
I think it is more complicated than that, that the play/production sequence within a single turn-date is:
1) We play our turn.
2) The AIs play their turns.
3) Our production occurs in our cities.
4) The AI production occurs in their cities.
And then the turn counter is incremented, and then there's a series of checks to see whether anyone has won yet.
(And in fact we're not necessarily even first in this whole sequence. We normally are, but if you create a scenario manually where the human is not the first player then I'd expect the sequence to change a bit to match that.)
You can tell that something like the above sequence is going on because after ending our turn we see the following things happen in this sequence:
a) We see rivals' units move.
b) We get prompted for the next thing to build in each city where current construction has completed.
c) Note that during (b), you can zoom to the city which just finished producing something, then use the left and right arrow keys to visit your other cities. If you have another city which will be completing production of something this same turn, you could change its production to something else which requires the same or less shields, exit the city screen, and you'll see that the second city produces the new thing you just changed it to. So this tells us that our cities definitely do not complete their improvements at the end of our turn before the rivals get their turns. (Note: I personally think that changing production in cities inter-turn is a bit exploitive and I wish Firaxis would make this impossible. But it can be done
)d) Sometimes after you've been prompted for the next thing to construct in each city, a popup will appear for a rival completing a wonder. This suggests that after our production for the turn is applied, then the AI production is applied.
Anyway, having said all that, I don't disagree that the production is internally happening in turn X, not turn X+1. I was just saying in my other note that the game's "feel" to me is as if production is completing in X+1, even though that's not the case. It has that feel to me because the production occurs after I see everyone else take their turn. Maybe that's just me? I don't know how it feels to other people. For a long time in my Civ3 play I just assumed that production was happening at X+1 because that's how it felt. And it didn't matter for most purposes, the distinction only matters in some very particular cases, like the culture calculator
You are right - it certainly is more complicated than a single sequence of player turns.
The thing that first got me thinking about all this was that when I first started writing timelines for QSC I was writing down everything that happened between turns under the heading of the next date, echoing your approach. Then I found I needed to write down the AI activities under an "IBT" (in-between-turns") heading. Then I found that in order to keep the sequence the same as the appearance on screen I had to move my completion events into the IBT sections.
The IBT period undoubtedly contains some elements like improvement completions that are "end-of-my-previous-turn" events and others that are "start-of-my-next-turn" events. When I truly understand the sequence it's my guess that I shall be able to lose the IBT heading and put everything in my timelines, including AI activities, under the date before or the date after.
All this might sound like counting angels on pinheads, but there are also wider implications.
Like Moonsinger, I feel if I can get inside the heads of the Civ3 programmers I figure I'll have a better chance to beat it, and I suspect that understanding this IBT event sequencing may help me with that process.
And what date should a GOTM culture victory score use? I recorded my GOTM21 win date as the turn after it was announced, but I suspect it should really be the date of the turn before, as that was when the culture score racked up to 100K.
The thing that first got me thinking about all this was that when I first started writing timelines for QSC I was writing down everything that happened between turns under the heading of the next date, echoing your approach. Then I found I needed to write down the AI activities under an "IBT" (in-between-turns") heading. Then I found that in order to keep the sequence the same as the appearance on screen I had to move my completion events into the IBT sections.
The IBT period undoubtedly contains some elements like improvement completions that are "end-of-my-previous-turn" events and others that are "start-of-my-next-turn" events. When I truly understand the sequence it's my guess that I shall be able to lose the IBT heading and put everything in my timelines, including AI activities, under the date before or the date after.
All this might sound like counting angels on pinheads, but there are also wider implications.
Like Moonsinger, I feel if I can get inside the heads of the Civ3 programmers I figure I'll have a better chance to beat it, and I suspect that understanding this IBT event sequencing may help me with that process.
And what date should a GOTM culture victory score use? I recorded my GOTM21 win date as the turn after it was announced, but I suspect it should really be the date of the turn before, as that was when the culture score racked up to 100K.
Similar threads
- Sticky
