HoF Map Finder/Generator Utility 2016-10-05

[superslug]

I'm about to head out for a ten hour shift at work, so I'm going to leave this delight of a program running for twelve hours when I'm gone computing maps. I'm going to set it for C3C 1.15 pangaea/warm/wet/5billion and a domination limit of 4001 tiles. I'll post again when I get home as to how many .sav's I have.

Since computer specs will likely have an effect on how fast this program runs, my processor is 1.1-1.3ghz Athlon (don't recall exactly) and 3xx megs of RAM.

 

[Moonsinger]

Originally posted by superslug
I'm about to head out for a ten hour shift at work, so I'm going to leave this delight of a program running for twelve hours when I'm gone computing maps. I'm going to set it for C3C 1.15 pangaea/warm/wet/5billion and a domination limit of 4001 tiles. I'll post again when I get home as to how many .sav's I have.

Since computer specs will likely have an effect on how fast this program runs, my processor is 1.1-1.3ghz Athlon (don't recall exactly) and 3xx megs of RAM.

I think the default values (1000, 9000, 1000, 7000) should work just fine on your system. I set the time delays on my P4 3.4 GHz as (1000, 4000, 1000, 3000). You can pretty much time the number of second between those screens and adjust it accordingly (please note that I will make it to remember your configure so that you don't have to fine tune it each time - on the next version). For example, if it's spending too much time at the "Choose Your World" screen before moving to the "Player Setup" screen, just decrease the time delay for the "New Game" screen from 9000 down to 7000 or something. By adding all those time delays all together, you can pretty much know how long it takes to iterate a game.

Even for the (1000, 9000, 1000, 7000) config for most slow system, it would take about (1+9+1+7) = about 20 second per iteration. If you run it for about 3 hours, 60*60*3 div 20 = 540 iterations right there. I'm sure it will find a few dozen maps out of those 540 games for you to pick.

 

[superslug]

Would there be anything wrong with just setting the values to 0001 to let it blitz through, or would that be going to low?

Also, which value should I leave at default to give myself time to click on the M to shut the program down?

I didn't click on the M earlier and just shut Civ itself down. It was really kind of funny watching your program try and generate maps from Internet Explorer. I should've heeded your earlier warning about making sure to click on the M, but hey, no damage done and now I know how to do it properly.

 

[Moonsinger]

Originally posted by superslug
Would there be anything wrong with just setting the values to 0001 to let it blitz through, or would that be going to low?

You should not set it any lower than 1000 ms or 1 second. Please note that Civ application need to have enough time to do its business or the whole thing will not work. For example, if you tell Civ3 to save a game or load a game, you need to give your hard drive enough IO time, etc. If you actually time the status bar in Civ3 during loading or saving, you can tell that it would take a few second right there defending on the speed of your system and your hard drive. Therefore, you must not blitz it through.

Also, which value should I leave at default to give myself time to click on the M to shut the program down?

No, you can't not adjust that value. It's actually equal to 1000 ms! Basically, I let your system to have about 1 second to relax and to ready for the next run.

I didn't click on the M earlier and just shut Civ itself down. It was really kind of funny watching your program try and generate maps from Internet Explorer. I should've heeded your earlier warning about making sure to click on the M, but hey, no damage done and now I know how to do it properly.

No harm done! You can terminate it anyway you like, but please do try to terminate it gracefully by clicking the M icon.

 

[superslug]

Originally posted by Moonsinger
...but please do try to terminate it gracefully by clicking the M icon.

Well, I'll definitely terminate it through clicking the M from now on. I don't know about the graceful part though, I don't really know anything about terminating things with grace. Anyone who's downloaded my HOF Chieftain game and looked at the last turn in the .sav knows what I'm talking about.

 

[Moonsinger]

It's now officially version 1.0:

http://www.civfanatics.net/uploads6/mapfinder1-0.zip

If you have previously tested the Alpha version, please make sure you clean out old output directory (basically, start out with a fresh and empty Output directory for this one). From now on, it will keep track of its internal counter and previous settings and the Output directory doesn't have to be empty the next time you run. I also put a cap of 256 acceptable games for each batch (just in case someone don't know what they are doing, I don't really want to fill up their hard drive with zillion of save games)

Next version:

1. Optional check for bonus food/shield/river feature
2. HTML report at the end of each batch with the table listing of the result plus runtime and map statistic. You can thanks Dianthus for this feature; I'm really lazy with these report card, but he sweet talked me into it.
?. The rest will be based upon your request. Have fun!

 

[superslug]

Originally posted by Moonsinger
1. Optional check for bonus food/shield/river feature

Thanks for all the hard work! I have more of an idea than an actual request. If it's not too hard, could you add luxuries to the food/shield/river options? I'd love to let my computer sit there and generate maps until it found one with all four!

The 256 batch cap is a great idea...

This is going to save me so much time, I might actually be able to beat one of your scores Moonsinger!...Wait a minute, you don't have any games on Monarch. Nevermind...

 

[Moonsinger]

Originally posted by superslug
If it's not too hard, could you add luxuries to the food/shield/river options? I'd love to let my computer sit there and generate maps until it found one with all four!

With the screenshot that I have provided for the selected map, I think you should have no problem spotting out the luxuries. Anyway, I concur with what you said, it would be nice if we can find a map with all four at the starting location.

In the meantime, I just PMed Aeson if it's ok to include his formula for calculating the possible potential score for the HoF map at 4000BC (just like what he is doing for the GOTM). Imaging knowing what score you are going to get (within +- 2000 points of course) before playing, is it too much of a spoiler? Is it legal for the HoF game?

 

[Dianthus]

Originally posted by Moonsinger
Anyway, I concur with what you said, it would be nice if we can find a map with all four at the starting location.

I'm not sure this is even possible. Doesn't the map generator tend to cluster resources of the same type together and seperate different resources?

 

[Moonsinger]

Originally posted by Dianthus
I'm not sure this is even possible. Doesn't the map generator tend to cluster resources of the same type together and seperate different resources?

I think the "four" things that Superslug said was

1. food
2. shield
3. river
4. luxury

Personally, I don't think luxury at the starting location would help determine the highscore in the end. Sure, it's nice to have, but it isn't that important for a highscore game. On the other hand, the bonus of food/shield/river is very important for us to quickly setup the second and third city and so on. After we have a handful number of towns, we could easily pumping out troops to look for nearby luxuries. Therefore, luxury at the starting location doesn't benefit much in a milk game (well, may be few extra point for those happy faces and some extra gold - not much really).

 

[Svar]

I downloaded this utility yesterday and ran some test worlds all the way down to a domination limit of 3300 but have yet to get any save game files in a Civ3Map folder. At 8:00 AM PST I started another run at a domination limit of 3800 so will be able to see if I have any when I get home for lunch at 11:30 AM PST (I live 7 minutes from work). Where exactly will I find the Civ3Map folder?

 

[Aeson]

In the meantime, I just PMed Aeson if it's ok to include his formula for calculating the possible potential score for the HoF map at 4000BC (just like what he is doing for the GOTM). Imaging knowing what score you are going to get (within +- 2000 points of course) before playing, is it too much of a spoiler? Is it legal for the HoF game?

The GOTM formula requires knowlege of the map and wouldn't be acceptable. Other simpler formulas that are derived from only the domination limit would be ok.

 

[Dianthus]

Originally posted by Svar
... and ran some test worlds all the way down to a domination limit of 3300 but have yet to get any save game files ...

Svar, are you generating huge maps? If not then the domination limit will have to be quite a bit lower!

 

[Grey Fox]

Can someone tell me why this wouldn't be like cheating?

Sure, might eliminate the fact that those who randomly get good maps will win the HOF more easily. (If everyone playing the HOF uses this program, otherwise it will be worse.)

I don't know... I'm still thorn about it.

 

[Moonsinger]

Originally posted by Svar
Where exactly will I find the Civ3Map folder?

Save game and screen shot that meet the domination limit requirement that you are looking for.

 

[Dianthus]

Originally posted by Aeson
The GOTM formula requires knowlege of the map and wouldn't be acceptable. Other simpler formulas that are derived from only the domination limit would be ok.

Are you referring to the need to look at the map to derive some of the inputs to the GOTM formula, or that the score output from the formula would constitute "knowledge of the map"? If the former then maybe a way to programmatically derive those inputs without the aid of a user would be possible, so avoiding the need for the user to gain that knowledge?

 

[Dianthus]

Originally posted by Grey Fox
Can someone tell me why this wouldn't be like cheating?

Because it doesn't do anything you can't do by hand. You can create random maps and reject them based on the start location and mapstats already, this utility just automates the rejection process!

 

[Grey Fox]

Well, I don't think mapstats should be used either.

 

[Moonsinger]

Originally posted by Grey Fox
Can someone tell me why this wouldn't be like cheating?

It's not cheating because this program just simply save you time in looking for a good random map. Sure, it may be looking for river and food bonus at the starting position but that's no different than what you would normally do at 4000BC anyway. The only different is that it can evaluate at least 1000 savegames per night while it would take the human month to do the same thing.

 

[Moonsinger]

Originally posted by Aeson
The GOTM formula requires knowlege of the map and wouldn't be acceptable. Other simpler formulas that are derived from only the domination limit would be ok.

Thanks for your quick answer! Sure, it may be nice to know the potential score by summing up the total of land type based on the mapstat but let's not walk on thin ice.