1. We have added a Gift Upgrades feature that allows you to gift an account upgrade to another member, just in time for the holiday season. You can see the gift option when going to the Account Upgrades screen, or on any user profile screen.
    Dismiss Notice

Metavariables and odds of winning a map

Discussion in 'Civ4 - Strategy & Tips' started by futurehermit, Dec 11, 2017.

  1. futurehermit

    futurehermit Chieftain

    Apr 3, 2006
    This is something that has been percolating in my head for a couple years now. Over the years I have mostly played standard continents maps on normal speed. Last few years have been on Emperor. The reason I've been thinking about this is that -- assuming the same leader -- there are some maps I can win handily and some maps that I fail at miserably. Now, all things being equal -- player, leader, difficulty level and other settings -- what creates these swings from winning easily to failing miserably? I'm a long-time player well versed in a variety of playstyles and strategies/openings, although I tend toward favouring domination victory conditions. In the last few years, I have tended toward using cottages frequently but am comfortable running specialists and taking other approaches when the map calls for it.

    Anyway -- I've arrived at an idea of metavariables and their relationship to the odds of winning a map. This is surely based on many threads I've read here over the decade+ I've frequented this site, so I don't claim any originality here, although I do wonder if this idea has been put forward in this way before (If so, I'd be interested to read the threads).

    I'm using the term metavariables to differentiate these more global factors from basic variables like number of neighbours, amount of food, number of cities available without war, etc. etc. Basically, metavariables would be clusters of these variables, but very specific clusters. Let me give you the main examples I'm thinking of:

    1) Quality of Land (QoL). Without a doubt, I would say this is the strongest predictor of whether or not I am winning or not. If I'm nestled in the greenbelt (equator) with lots of snaky rivers, then I feel good about my chances of winning. If I'm coastal with lots of water and lots of desert and/or tundra nearby, I feel good about my chances of losing.

    I would be interested in exploring possible formulae for describing this metavariable. I think it would include components such as:

    Quality of Capital Land (QoCL). How much food? How much production? How much commerce? How many forests? How many resources? This could be narrowed down even further to how much of each of these things in the 4-5 best tiles (assuming happy cap of 5 early on)? So, you could have Quality of Capital Land best 5 tiles (QoCL5) and QoCL overall. Strong capital? I like my odds. Weak capital? Tougher game.

    However, the capital isn't the be all end all. I've had times where I've had a food lite but production decent capital and wonderful surrounding terrain, so I just rexed into the rest of the land and smooth sailing. So, next would be:

    Quality of Surrounding Land (QoSL). I just played a start that got me thinking about this whole metavariable idea again. I had a coastal start surrounded by desert. Decent enough capital but crap surrounding land. Issy was to my south sitting in the green belt and outgrew me in economy and military easy while I was beating down Joao to secure my marginal peninsula. Whether or not I should have managed diplo with Joao better or not early and hit Issy early is for later in this thread, but if I just consider my surrounding land, Joao's surrounding land, and Issy's wonderful surrounding land (she had lots of green, lots of rivers, multiple gold pits, lots of surplus food, and lots of wine and other resources), I would be betting on Issy hands down at the outset if I was a neutral observer with perfect information.

    I'm wondering what might count as "surrounding" land. Maybe 10 tiles in every direction, not including the capital BFC? More? Less?

    Basically, if I can get myself into 8-15 cities with a nice economy by 1500 AD, I like my chances of massing rifles and clearing my continent. If I've got the larger of the two continents, I feel really good about winning the game by domination or space in particular.

    From there, how to score the land. 0 for desert/mountain, 1 for ocean and other marginal tiles, move up to something like 3 or 4 for grassland adding +1 for river and -1 if jungle on it, more still for resources. I'd be interested in an exploration of the specifics here, but it is the general concept that really makes me mull this.

    2) Competition for Land (CfL). This would be a blend of things like #/type/proximity of neighbours, as well as availability/type/proximity of strategic resources and available production. If I've got copper in my capital BFC and Gandhi on my doorstep with no strategic resources and a juicy capital, well that game is looking pretty much done before it started. However, if I have no copper or horses and Shaka building up strength a bit of a ways away from me, or Alex or Monty on my doorstep, then I don't care how lush my surrounding land is, I'm in for a dogfight that will grind my economy to a halt at best and annihilation at worst. Interestingly enough, I would also say that having NO neighbours -- i.e., isolated start -- would be a negative here. So, maybe CfL isn't the best descriptor for this metavariable. In other words, this would also take into account trade partners. If I've got nobody to help me keep on pace with tech into the renaissance relative to the other continent, then I don't like my chances, especially if there's a lovefest between financial civs on the other continent.

    Possible subvariables here would be things like:

    1) # of neighbours
    2) Disposition of neighbours
    3) Distance of neighbours
    4) Strategic resources (type/location/me vs. neighbours/etc.)
    5) Trade opportunities (Toku vs. Mansa anyone?)
    6) etc.

    Again, happy to hammer out the specifics, but it is the general concept here that I find interesting.

    Anyway, those are the two main metavariables that I would say predict my outcomes. If QoL and CfL are favourable, good to go. If they are unfavourable, likely to lose. Or, to put it another way, these are the factors that heavily modify difficulty level -- all other things being equal.

    I'm interested to hear your thoughts on this and other metavariables you think might be important for influencing the outcomes of your games.

  2. Pedro78

    Pedro78 Chieftain

    Oct 20, 2016
    I don't think quality of land determines whether you're going to win a game or not. With sufficient skill I think that there is only a few types of maps that cannot be won consistently (on deity and not-too-funky settings):
    • Several "dangerous" close neighbours : double early DoW = game over in most cases
    • Boxed in at <4 cities without a strategic resource. Archerpult is an option but is very hard and not always achievable
    • Isolated with crap land AND no access to iron: if the game's tech pace is too fast (i.e. every AI gets Rifles or even Grenadiers before you can pull off a significant non-cannon attack) you're screwed. With good land you can go all the way to Artillery before attacking.
    Of course there are other types of "hard" maps, but imo a very good player should be able to win every map but the ones stated above very consistently. If I keep playing Civ 4, that would be my goal as a player :)
  3. elmurcis

    elmurcis Chieftain

    Feb 16, 2011
    I would put "good games with nice finish date and... whatever" in one pocket.
    And "winable in any way and any time" in other pocket.. 2nd might be real "nightmare" time to time but can remember some crazy-difficult games here... where a lot of Nukes vs Culture win versus different AI and cities was only chance to stay in game until human player win :D
  4. futurehermit

    futurehermit Chieftain

    Apr 3, 2006
    Well, fair enough. Let's not think of it as win or lose, but rather easier or harder. Some games I "lose" I might win in the long run, but may have ground into an uninteresting slog that I don't care to play out, etc.

    My main interest is what are the metavariables that moderate the difficulty of a level, other than game settings. I think quality of land is actually a big deal. I'm not saying that bad land = unwinnable, but surely bad land is harder than good land. The three scenarios you describe I would put under competition for land as I detailed in my original post. As I said, maybe competition for land isn't the best descriptor, but definitely dangerous close neighbours or land-grabby neighbours that box you in or being isolated with crap land are grave scenarios.

    I was just really struck in my last game how the quality of the land really impacted my start vs. how AI Issy was able to start running away with things. When I opened up world builder and compared our land, I was just left SMH. Now, could I have won that game with better play? Yep, probably. If I found a way to keep the peace with Joao (tough cuz I was naturally placed to box him in) and attack Issy early (tough because I would've had to go through jungle to hit her at some distance), I probably could've hurt her and claimed some of that prime real estate earlier -- at the very least, kept her from peacefully growing into it uncontested. Would've been tough, but maybe doable. But my point I guess is precisely that -- it would've been tough, due to the relative imbalance in the quality of land in our starting areas. Then throw in the competition for land with rexer Joao on my doorstep.

    Anyway, I just love exploring elements of this game, even after all these years. It is my #1 game all time. Never gets old, really. :)
  5. Kaitzilla

    Kaitzilla Lord Croissant

    Jun 21, 2008
    I like this quality of land idea.
    It for sure makes the game harder or easier.
  6. Snowbird

    Snowbird Chieftain

    Feb 2, 2017
    Quality of land ofcourse matters, but my vote would go for ability to control situation. That is why iso is problematic, for good part of the game you have just a vague idea that is going on elsewhere. The more AI you know and the more you can abuse them in diplo/war the easier game happens to be.
  7. futurehermit

    futurehermit Chieftain

    Apr 3, 2006
    Yeah, I like "ability to control situation" better than "competition for land." If you are isolated, you have very little control. If you have two aggressors who hit you early, loss of control. etc. And, I don't think it just has to be one metavariable that explains everything. There could be a few or more.
  8. Fippy

    Fippy Micro Junkie Queen

    Mar 17, 2013
    I would add location of starting posi (middle or corner), middle and surrounded by dangerous AIs is much more difficult than in a corner (same AIs). See thread https://forums.civfanatics.com/threads/deity-game-playthrough-1.624959/

    While there's nice land (and Ivory), deity AIs could steal all of that before you can even get out your first settler ~~
    And then start plotting war, so those maps have a high trouble potential.

    Chat for Pedro, there are some maps that could never be won if we would strictly play all starts on various map scripts :)
    I have seen some really silly ones when looking for Boreal maps, 1 tile island with only ice behind the water was best.
  9. SittinDown

    SittinDown Chieftain

    Jun 29, 2016
    True story. I got surrounded and beat despite my best efforts and great advice from some great players.
  10. Macksideshow

    Macksideshow Chieftain

    Jan 28, 2013
    As someone who rerolls maps like a fiend this is something I consider a lot although not in such a structured way as the OP.

    My objective is early win dates. My method is to reroll starts until a good (not necessarily great) starting location then play the map until early strategic resources are revealed (Copper and/or Horses, maybe to Iron in some cases) before deciding whether to continue.

    With that method I am exposed to the metavariables you list however; Competition for land, number of neighbours, disposition of neighbours/trade opportunities are not always completely known.

    I find that QoCL alone is not a good predictor of game difficulty which is why I play the map until Copper/Horses. Good QoCL is important but great QoCL will not make up for poor values in the other variables. Synergy between meta-variables is what counts the most. No use having; great QoCL and poor QoSL, a good start for an early rush but no good target (or vice versa) or a start that has lots of rivers, lots of food but no early luxuries etc. Also, a minimum amount of food/hammers/commerce needs to available early or progress will be delayed even if there is lots of the others available.

    What counts as good for each category is dependant on Civ/leader choice, rivers are better for FIN, food for PHI, hammers/start resources for civs with early rush unique units, etc.

    I try and look at the overall situation once copper/horses are revealed and imagine hOW the map will play out when deciding. On different maps the relative value of each metavariable can vary but the underlying theme is typically one or two 2 great ones to leverage and very few poors to weigh down on progress.

    A couple of metavariables that I consider that have not been specifically mentioned are: Number of forests in capital BFC and nearby (subset of QoCL/QoSL) and feasibilty of leveraging (this is distinct from simply being able to build it) powerful early wonder, eg Pyramids, Great Lighthouse

Share This Page