1-Turn Win, and Zero City Challenge

[Posidonius]

This is the funniest thing i ever did to Civ1, love bending games but did not think i could twist Civ1 this far out of shape. Didn't use fast-Settler or shift-56 or sentry-unsentry, only used the save-quit-reload cheat. A few hundred times, but you never know the limits unless you push them. Here's a demographics screenshot just before i won the game.

Yep, the Zero City Challenge is now a real thing. I just won the game with no cities, so we know it's possible. Didn't crash or nuthin, no sir. In theory, it should be possible with no cheats, but maybe only on Warlord level. On Chieftan, you can't ditch your last city.

Killed my last rival civ in 3,980 BC, but the game hiccuped and skipped 3,960 as if it didn't happen, and ran me though the win-game process in 3,940 BC. But it credited me with the win in 3,960 BC, according to the civscore. Close enough for horseshoes.

Another funny thing, here's the game's power graph:

There's just nothing there. Nobody had cities for more than a few seconds, in geologic time. Blessed with geography, the starting squares for all 7 civs were on one huge continent. When you know how to manipulate where rival civs respawn, you can control all 6 rivals right from the get-go.

This is not a stunning achievement, it was only an exercise to prove two theoretical possibilities:

1. you can win with no cities, and

2. you can beat 14 enemy civilizations by 3980 BC.

Yes, i said 14. The Babylonians spawned a free Settler as soon as i killed off Babylon itself in 4,000 BC so had to kill them twice, and although i killed an Indian Settler in 4,000 BC, there was the city of Delhi in 3,980 BC, which i also killed, and then killed the respawn of Mongol Settlers. So killed 12 rivals, but had to kill a couple of them twice. All by 3,980 BC.

It's the 1-Turn Win on Emperor with 7 civs. Yes, it's not a true "win" because it pounds on one cheat heavily, but it does prove that it is possible. It got 4th in my HOF, and that taught an important lesson: the highest possible score in a conquest win is dwarfed by the high scores possible when you guide a whole planet to a space win.

 

[Posidonius]

"the highest possible score in a conquest win is dwarfed by the high scores possible when you guide a whole planet to a space win."

Really, i've wanted to prove this for a long time. Still just jaw-dropped that i really pulled it off. It took every trick i knew about managing rival civs. About a dozen years ago in this forum, there was a discussion about it, and some civvers were certain that an early conquest would always beat a spaceship win. Their only quandary was how to get the fastest conquest win. It just didn't ring true to me, that all the work the designers put into it, making an Alpha Centauri win possible, would end up worth less than a hack n' slash rampage across the landscape. Myth now exploded, so now we know. If you want the truly high scores, you have to do the work.

 

[tuga2112]

"the highest possible score in a conquest win is dwarfed by the high scores possible when you guide a whole planet to a space win."

Really, i've wanted to prove this for a long time. Still just jaw-dropped that i really pulled it off. It took every trick i knew about managing rival civs. About a dozen years ago in this forum, there was a discussion about it, and some civvers were certain that an early conquest would always beat a spaceship win. Their only quandary was how to get the fastest conquest win. It just didn't ring true to me, that all the work the designers put into it, making an Alpha Centauri win possible, would end up worth less than a hack n' slash rampage across the landscape. Myth now exploded, so now we know. If you want the truly high scores, you have to do the work.

i gotta say. im really pleased to see confirmation of that theory. (even tho i dont understand the appeal of constantly reloading the game for hours as previously discussed)
i just always assumed a good player would easily outweight the passage of time with population boom boosted with future tech and space race. you seem to prove that to the most extreme.

id have to double check my current score in king, but i suspect my raw score is higher than the zero turn conquest score you got... and i havent "won" officially yet.

 

[Posidonius]

i gotta say. im really pleased to see confirmation of that theory. (even tho i dont understand the appeal of constantly reloading the game for hours as previously discussed)

There is no appeal, it's drudgery. The only reason to do it was to see if it was possible to kill every rival in 3,980 BC. Some months back, posted here about it being possible in theory, but shelved the idea because it would be difficult to try even when the attempt failed. Someone gave me a good idea, try it with less than 7 civs, but there wasn't much more control over the game. Then realized that although a 3980 win was difficult, 3960 should be the opposite, almost guaranteed. If you buy a Trireme in 3,980 BC, you can go anywhere. The save-quit-reload cheat makes the Trireme into an ocean-worthy vessel. And because rivals will respawn immediately in 3960, you can catch 'em all. It would only add a hundred cycles to the game.

But didn't even have to do that, ended up with a ZCC win as a bonus, because all 6 rival civs spawned on my own continent in 4,000 BC. What a stroke of luck, right? Actually not so stroky. At the start of the game chose Emperor and 7 civs, and ran about 30 gamestarts until i got one with Map Making in 4,000 BC. If you're keen on stats, two of the gamestarts were double yolks, with 2 starting Settlers. Four of them had second-tier techs to start, but that was not a stroke of luck. I made that happen. Here's how...

If you don't customize the planet's parameters before playing, you'll end up with a decently productive landscape with plenty of spots the AI can plop down 7 nations in 4,001 BC. We know that the number of starting techs for your civ is related to the potential productivity of your startsquare. So if you want the best odds of getting a 2nd-tier tech like Maps, you make the landscape inhospitable.

Obviously picked "Large" for landmass, since i wanted max huts and max number of rivals i can catch in 4,000 BC. But for the climate, i chose "Cool" and "Arid" because that gives two things: better chance of multiple starting techs, and hobbles the rivals just as much as me. Finally, chose the Young Planet option, 3 bln years. Because a young planet tends to have larger clumps of similar terrain types, the lesser variation in each area makes the possible civ spawning points more discrete, because the inhospitable areas also come in larger contiguous clumps. You don't have to plod a Cavalry through a huge mountain range to know that a rival never respawned there. So, you can watch more potential respawn squares with fewer sentinel units.

Large, Cool, Arid, Young. LCAY in shorthand. Not a planet you would normally design for a game, i usually play LWWY games for the chance at vast lush areas. In one LWW3 game, had a 200-square peninsula entirely covered with rivers. It was the most beautiful thing i ever saw in a random map. But for this purpose, the LCA3 map seemed like a good idea, and it paid off splendidly. Had a thorny problem with one civ respawning on an island, but having read darkpanda's treatise on randomness in Civ1, was able to identify the problem and craft a workaround. God bless darkpanda, that work opened a lot of crazy ways to bend Civ1.

i just always assumed a good player would easily outweight the passage of time with population boom boosted with future tech and space race. you seem to prove that to the most extreme

All i did was cement the upper limit of a conquest win: a 209% civ rating. The theoretical limit of the civ rating is 655% if my memory serves aright. That's where the future lays, how do you get obscenely high scores? All we proved here is that a conquest win is not the way.

 

[tuga2112]

All i did was cement the upper limit of a conquest win: a 209% civ rating. The theoretical limit of the civ rating is 655% if my memory serves aright. That's where the future lays, how do you get obscenely high scores? All we proved here is that a conquest win is not the way.

is the next iteration is to try a conquest game ending at the last year of the game now?
i would expect score can be higher that way.

 

[Posidonius]

is the next iteration is to try a conquest game ending at the last year of the game now?

Every conquest win is, by definition, accomplished on the last turn of the game. A conquest win at the last possible turn would be the minimum possible score. I just set the upper limit, wonder how low it can go in a winning game? The only source of civscore points in a conquest is the year in which it happens. You get no points for Wonders, not for your citz, and obviously 0 civscore bonus for "Peace"

So what would be the minimum winning civscore and rating? I guess it would be winning on Chieftain level in 2,100 AD, the level with the latest playable turn. The game gives you 1000 civscore points for a win in the year 2,000 AD, and there's a +2 bonus for each turn before Y2K. That's how i got 2094 civscore in this 1-turn game: the standard 1000 and there are 550 turns from 4,000 BC and 2,000 AD. Double that 550 and add 1000, for a 2100 maximum score, for a conquest win. Everyone else died by 3,980 BC, but the game registered the win in 3,960 BC, so i got 2094 civscore.

Assuming that the game subtracts 2 civscore points for a conquest win in 2,001 AD, but don't know for sure. If that's true, then a Chieftain conquest in 2,100 AD would mean 800 civscore points. To get the civ's rating, multiply by 0.20 because it's Chieftain level = 160. Then simply divide by 10 to get the civrating: 16%

But because of what i did here, in the game explored at the top of this thread, it seems possible to win with a civscore of 798, and if the algos for civrating all round down, then maybe you can win with a Civilization Rating of 15% That would cement the lower limits of civscore and civ rating. Maybe i'd do it someday. Or you can do it, seems like it wouldn't take as long as it sounds.

650 turns to play, to get to 2,100 AD. But the vast majority of them would be a single click (or keystroke). If i were to try, it'd be a Chieftain game with 1 rival civ, and they would have to be 2nd on the turn order. Would want to find them on my home continent, and their respawn too, so would customize the planet for large landmass. As Chieftain you don't need any advantages, so might as well try and hobble your lone opponent, so would customize the planet for Cool and Arid and Young (3 bln years old). LCAY, your best bet for quickly getting control of your rival.

Found a couple cities, explore fast, reach for The Wheel soon. Found 3-4 more cities and put a half-dozen Chariots in the field. Prune back the rival civ to 1 city in a cocoon, occupy its forests and pillage its mines. Kill anything that comes out of it, and give it a slap now and then with a Chariot.

Then flood the world with Diplomats. Place them in a network where you can see all possible respawn spots by moving them all whichways in one big turn. Kill the rival city, catch the respawn the same turn. Herd the Settler to a harmless spot, and get a Chariot over there. Instead of backing the larval civ up against a coastline, which is an easier trap, you probably want to herd this one up into inland mountains, to protect it from the seaborne Red Wolves.

At this point, ensure the common defense of your cities. Might be wiser to disband all your own cities except the capital. Less complications, for the remaining 575 turns. Don't need gold or techs, but you'll get them anyway from huts, so just turn the sentinel army of Diplomats into a harassing force against Barbarians. Buy them before they get too close to your capital, and turn the zombies into more guards. Set your Sci rate low so you're bothered less as the millennia roll by.

About 30 Barbarian incursions later, it's 2,100 AD. Slay the rival civ you've been keeping as a meek bunny. Then end your turn. One of three things will happen:

1. The game will end in a loss because it's 2,101 AD.

2. The game will end in a win because all rivals are dead, civscore 798.

3. The game will not end because your rival respawned as themselves, but will cease keeping score.

Don't know what will happen. Try it and let us know. If you can win with a 15% civ rating, it would be a feather in your cap. As for me, going in the other direction. Got the max possible conquest win, now trying the max spaceship win. But even that doesn't have to happen on the last possible date. You can calculate how long it will take to complete the other civscore sources, and launch a 15-year spaceship beforehand. You just go when you're ready, there is no set date.

 

[Posidonius]

There's a new oddity discovered in Civ1, from doing this 1-Turn Win. Dozens of times killing off all 6 rivals at once, then watching where they respawn. Realized that i was wrong in the earlier '4KB Aha' post: rivals do respawn immediately when killed in 4,000 BC, but only if you kill them as a city, not as a larval Settler. When you kill a rival as a larva, there is a chance they'll respawn as their own nationality, still the same name, not just the same color. Have seen that puzzler over the years. But this has nothing to do with whether the rival is a city or a Settler.

The odd thing is that there is a variation in the level of randomness, in terms of where in the world the respawn happens. This variation has a pattern. The pattern depends on the rival civ's relative position in the turn order, relative to your own civ's position. The civs just after my position had the widest geographical distribution of respawns, and were the most likely to re-respawn as the same nation. Oh, in case you don't know, the turn order is in the Intelligence Report screen.

All the civs showed a descending amount of variation in respawn spots, up to the very last one, the civ just above me on the turn order. That one civ always respawned in the same exact square, over thirty times. Spooky, right? It gets more. I learned how to exploit this quirk. There was one civ which was a city in 4,000 BC, and every time, twenty times or more, when i sacked it i could not find the respawn. But in CivWin you can see the omalos, where the next civ after you on the turn order is located. And the civ just after me in the turn order also has the greatest variability in where they respawn, right?

So duh, just kill that one off repeatedly and look for the omalos. Sure enough, there was some sort of land to my North where the other, troublesome, civ must have been respawning, every stinking time. Every time, because it was 4th after me on the turn order, and thus had a low geographical variation in respawn spots. Now recall that darkpanda showed that the game's "dice" get re-rolled when you open a city-view screen, but what if you have no cities? Ah yes, darkpanda showed that the weighted dice of randomness in Civ1 are also rolled at the start of a turn.

And that's all i had to do. Just had to shake up the "random" pattern of respawn spots into another "random" pattern. Killed everyone except the 1st-gen city of the irksome color, then went into 3,980 BC, killed all the respawns, and then the annoying city, and voila! they respawned right in front of me. Learned that the respawn spots are determined at the start of every turn, with a decreasing level of randomness as one proceeds down the turn order list, and the effect wraps around the list, starting and ending with your position.

 

[tuga2112]

A conquest win at the last possible turn would be the minimum possible score.

i was not aware the conquest score is 100% dependant on date. that explains why everyone keeps on bragging about conquest games based on the date. since its not relevant if the conquest is done with zero or 3000 billion population

 

[Lord.L.]

What's the save-reload cheat?

 

[sjongejonge]

What's the save-reload cheat?

Move all your soldiers and settlers except for 1 unit --> save the game --> load the game, so they can all move again in the same year.

 

[Posidonius]

Move all your soldiers and settlers except for 1 unit --> save the game --> load the game, so they can all move again in the same year.

No need to reserve one unit. Simply turn off the "Auto End Turn" option, so when you move your last unit there is a prompt to end your turn. Do not click the prompt, do the save-quit-reload cycle.

 

[Lord.L.]

OMG, how many hours have you spent on this?

 

[Mize]

An important point to the save/load cheat is that you need to rename the .sve and .map files so your save turns up in an autosave slot when you load it. Loading it from a regular slot will not give you your units' movement points back. Slots civil0 to civil3 are the four regular ones, which you can access when you save. Slots from 4 to 9 are the six autosave slots you want.

 

[Posidonius]

An important point to the save/load cheat is that you need to rename the .sve and .map files so your save turns up in an autosave slot when you load it. Loading it from a regular slot will not give you your units' movement points back. Slots civil0 to civil3 are the four regular ones, which you can access when you save. Slots from 4 to 9 are the six autosave slots you want.

This is true in CivDOS, but in CivDOS a 1-Turn win is not possible. Best you can do is a 2-Turn win with civscore 2092. Because you can't save the game in 4,000 BC. Doing the save-quit-reload cycle is not possible if you have the Auto Save Game option turned on. Why? Because you start with a single unit, your Settler, and you need to save the game after it moves, but before your turn ends.

Another advantage to CivWIN for this kind of work is that the sve/map files are each a unified file with a .SAV suffix. Opening one of these always starts off at the beginning of your phase of the current game turn, so you always have full movement. And CivWIN doesn't care what you name the files, as long as the format is XXX####X.sav, where the X equals a letter and # is a digit. The first 3 characters are the first three letters of your player name. The four digits are the year, and the last X is either an A or B (for AD or BC). Doesn't matter what you name the savegame file, because it contains the current game year inside. Here's the flow of it:

1. turn off the Auto Save Game "feature"
2. save the game as xxx4000B.sav
3. move your starting Settler, looking for a hut
4. save the game as xxx4001B.sav
5. quit
6. open the 4001B savegame
7. move the Settler
8. save as xxx4002B.sav
9. quit
10. et cetera

Once you find a hut and milk a Cavalry out of it, the turn cycles take a little longer, with 2 units to move. But it is infinitely quicker if you don't have to rename savegame files after creating them, and instead of counting years downwards like we are used to in the BC era, simply count upwards, where you have 5,999 "slots" for saved games.

OMG, how many hours have you spent on this?

The game described at the top of this thread took two civving sessions, about 5 hours each, including the time typing out the CFF post and getting screenshots for it. Most of the time spent solving the problem of China respawning as America on a different landmass, re-reading darkpanda's posts about Civ1 randomness to find the loophole. But also grasped that the pattern of respawn dispersion "randomness" is based on a civ's position in the turn order, which darkpanda has not yet investigated, and then the rest of the game became straightforward.

Once you know what you're trying to do, a save-quit-reload cycle takes little time. Even with a dozen units in play, most cycles are under 90 seconds, from splash-screen to splash-screen, if you avoid using the mouse. So it took 10 hours the first time, but with using the new knowledge, could reproduce the 2,094 civscore result, reliably, in just a few hours playing.