Why am I seeing some of the same leaders/civs so often?

[Scroller]

Of the dozen or more games that I have started recently (and, to be frank, beyond that as well), I have seen several of the same leaders in around half of them. Mansa Musa is one of the most frequent, closely followed by Peter. Of the rest, there are some that I have hardly seen at all - Canada and Arabia being two examples.

If you have a suggestion solely based on 'random' then please don't bother. The game has weights applied to each 'start', so overly-repetitive plays should be anticipated and modified (eg: through logs). Additionally, I am aware that the setup makes a difference. I have varied my game setup but it has little impact in the civs the AI selected.

I am aware that I could pick the civs but this takes away the anticipation and should not be necessary in order to have varied gameplay.

[Mercutio]

I would like an option to not have certain civs in the game.

[BlackSmokeDMax]

Originally posted by Mercutio:

I would like an option to not have certain civs in the game.

Something like the new "select natural wonders" menu option, but for random Civs, would be perfect!

[Mercutio]

exactly

[tempest.of.emptiness]

https://en.wikipedia.org/wiki/Texas_sharpshooter_fallacy

You can search this forum and find similar threads from other players reporting the same situations you are reporting: that certain civs appear in their games with greater frequency than other civs do. The interesting thing about those threads is that they aren't all finding the same civs to have increased frequency.

I suspect that the actual cause is random clustering, with some people who experience clustering perceiving it as a pattern because of the relatively small sample size of games that they have created. If all the games every Civ VI player starts were taken in aggregate there is likely no statistically significant difference in the frequency of appearance of any randomly chosen civ (assuming I am correct).

If you find that this pattern that you are observing bothers you, then one possible solution is to select the civs in your games pseudo-randomly. To do this, after you have selected your own civ, use a random number generator to select each of the other civs that will appear in your game, and then select those manually in the setup interface instead of choosing the random option. If this process selects a civ that you feel is appearing in too many of your games, simply reject it and choose another randomly until you are satisfied with the results.

In doing so you'll lose the element of surprise where you discover which civs are in your game while playing the game, but you'll also eliminate that feeling that the results are not sufficiently random.

[Mercutio]

While I appreciate your input, the idea is certain civs you may be tired of seeing, but wish there was still a randomness to the remaining ones.

[keith_dackson]

I play Trajan almost exclusively, and I have played a not insignificant number of games. I seem to have the following AI civs against me most often: England (both Victoria and Elanor), France (mostly the old queen, but since the new persona she has come up more than once), Brazil (quite often), Frederic (quite often), Egypt (quite often), Hungary ( a fair amount), China (fair amount), Korea (fair amount), Saladin (fairly often), Byzantium (fairly often), Russia (maybe every 5-10 games), Inca (every 5-10 games), Aztec (fairly often), Greeks (fairly often) - you get the picture.

While some of this can be chalked up to the when the other civs have been released, there appears to be a bias - some of it is due to traditional rivalries (but we all know how the Roman/Brazilian wars ended) but otherwise who knows - is it based on the inherent balance between the civ you choose and the aggressive tendencies of the AI civs? Something else?

[Scroller]

A great idea!

Originally posted by Mercutio:

I would like an option to not have certain civs in the game.

[Scroller]

The problem with always relying on the 'random generator' option is that this precludes the possibility of there being anything actually wrong. A cluster is certainly possible and I have not eliminated it as a possibility but such a cluster could have been avoided with better programming of the seed generator.

That said, I do like your idea and I may try picking them with my eyes shut or something. The things we do to have fun playing Civ 6...

EDIT: If we dismiss this simply as the sharpshooter fallacy we also permit and accept the possibility of the same game being generated every single time. This is not likely but it is entirely possible in an unrestrained random system. That would be a very bad game and that is why there should be constrains and balances to prevent so much clustering.

Originally posted by tempest.of.emptiness:

https://en.wikipedia.org/wiki/Texas_sharpshooter_fallacy

You can search this forum and find similar threads from other players reporting the same situations you are reporting: that certain civs appear in their games with greater frequency than other civs do. The interesting thing about those threads is that they aren't all finding the same civs to have increased frequency.

I suspect that the actual cause is random clustering, with some people who experience clustering perceiving it as a pattern because of the relatively small sample size of games that they have created. If all the games every Civ VI player starts were taken in aggregate there is likely no statistically significant difference in the frequency of appearance of any randomly chosen civ (assuming I am correct).

If you find that this pattern that you are observing bothers you, then one possible solution is to select the civs in your games pseudo-randomly. To do this, after you have selected your own civ, use a random number generator to select each of the other civs that will appear in your game, and then select those manually in the setup interface instead of choosing the random option. If this process selects a civ that you feel is appearing in too many of your games, simply reject it and choose another randomly until you are satisfied with the results.

In doing so you'll lose the element of surprise where you discover which civs are in your game while playing the game, but you'll also eliminate that feeling that the results are not sufficiently random.

[Scroller]

I did consider the biases but I'm not sure that is the entire problem. For instance, having played as both Englands, Spain, America and Germany, I have seen Mansa Musa far more than any other. I'm not sure about historical rivalries that would explain that.

It could be that the biases and weights are being amplified in some way but it is nigh on impossible to get that looked at with so much 'its just random' sloganning on the board.

Of those you have seen, I have played against both Frances perhaps once or twice each. I think I have seen Saladin once in about 20-30 games. The same with Brazil.

Originally posted by keith_dackson:

I play Trajan almost exclusively, and I have played a not insignificant number of games. I seem to have the following AI civs against me most often: England (both Victoria and Elanor), France (mostly the old queen, but since the new persona she has come up more than once), Brazil (quite often), Frederic (quite often), Egypt (quite often), Hungary ( a fair amount), China (fair amount), Korea (fair amount), Saladin (fairly often), Byzantium (fairly often), Russia (maybe every 5-10 games), Inca (every 5-10 games), Aztec (fairly often), Greeks (fairly often) - you get the picture.

While some of this can be chalked up to the when the other civs have been released, there appears to be a bias - some of it is due to traditional rivalries (but we all know how the Roman/Brazilian wars ended) but otherwise who knows - is it based on the inherent balance between the civ you choose and the aggressive tendencies of the AI civs? Something else?

[asnjas19]

Small sample sizes.
One persons observations.

Its just random. Play only 20 games and you might get the same leader 4- 5 times.

[dcain3456]

I really don't get you people complaining about random leaders. Pick 'em yourself like I do, no problem!

[nullpo]

I think they need to make the opponent less random so it's more random. When you face AI in one match, basically reduce their chance to be matched with you again next game. In all my game I always see france and georgia, but never even once i see japan and inca in my game

[tempest.of.emptiness]

Originally posted by Mercutio:

While I appreciate your input, the idea is certain civs you may be tired of seeing, but wish there was still a randomness to the remaining ones.

My previous post suggests a solution that allows you to determine the civs in your game randomly while also eliminating any civs that you feel are appearing too frequently or simply do not wish to play against.

[tempest.of.emptiness]

Originally posted by Scroller:

The problem with always relying on the 'random generator' option is that this precludes the possibility of there being anything actually wrong. A cluster is certainly possible and I have not eliminated it as a possibility but such a cluster could have been avoided with better programming of the seed generator.

True randomness includes clustering. However, I get that what you want is not true randomness, and you are right that the developers could have eliminated clusters if they had wanted to do so. I would love to see an option to exclude a civ from random consideration and also an option to turn off clustering in the random civ selections.

However, while we wait for that there are ways to take control of the situation ourselves and produce the results we want. In offering such a method as a suggestion I did not mean to suggest that the developers themselves should not take note of this situation and take action for the sake of player quality of life.

[Scroller]

A failure to read results in a failure to engage with the matter.

I have played many more than that and Musa has appeared in more than half.

Originally posted by asnjas19:

Small sample sizes.
One persons observations.

Its just random. Play only 20 games and you might get the same leader 4- 5 times.