If this was the bulk of your tourism for a culture victory, then yes, the variance on them would be insane.
I was saying that for any given expected value of tourism they can have more or less variance - i have no opinion on how impactful rock bands actually are. Although, some of those 1000 tourism promotions seem kinda strong. I'm not complaining about tourism, I'm complaining about the
fun from using the feature, particularly for players who are playing without number crunching.
I don't think it's going to cause me to rage quit
The starts was an extreme example. I wouldn't expect people to stop playing a game after band fails- they just won't enjoy the feature. The key difference with spies is that you send them on a mission and then you have to wait a while. With this, it's instant feedback. This is important - since the time to complete missions helps us decouple expected outcome and reality and act "better." Rock bands don't have that. Giving a band a really high fail rate like 62% is just too much. When i was referring to the "set of outcomes where you get a bunch of failures in a row" I was referring to this:
If one googles "wolfram alpha" and types in "binomialdistribution(4,0.375)" this can be replicated. The first argument (4) is the number of trials - aka coin flips or number of rock band units you make- and the second argument, (0.375) is the chance of success in a coin flip, or 100% minus the chance of failure.
The little table is detailing how likely each outcome is (0-4 refers to how many successful concerts you have) and the second column refers to the chance of getting that outcome. What this means is that if you send out 4 separate rock bands with a base fail rate of 62%, 15.26% (about 1 out of 6.5ish) times you end up with
four failures after they all play once. That means a full wipe. 52% of the time,
or half, you lose
at least 3 of the units. There's only a 15% chance that at least 3 of the rockers survive to play again.
The chance for any particular unit to successfully play a certain number of times is captured as (chance of success)^(# of successful concerts). Even a band with a promotion like some of the ones we saw in the video that had a 30ish% fail chance, only has a ~30% chance to actually play 4 concerts. Since a rock band at least can play one concert every time since it'll do something even if they die, you need to add that failed concert in with the total count. Here's a table describing those 62% failure rate bands looks like:
Basically, very few bands can actually do 2-3 concerts. If you had 8 rock bands touring, you'd end up with something like this:
5 total failures - 2 one success, then fail - 1 two or more successes, but almost always just two then fail.
The other issue is that they seem to have coupled concert tourism and failure rate. This creates a compound distribution for the expected tourism value which will be heavily skewed by the fact that failures are giving pretty much straight 0 star concerts, and successes can be 1-5 stars. Not good. FXS, unless you have a statistician on your squad, simplify the math here. It'll be brutally hard to balance. And you run the risk of having games where a player (think of multiplayer) gets that lucky roll where they go gangbusters and just rake in successful concerts and possibly win in a matter of turns. Even the lucky stars outcome has a 2% chance of actually happening, in a small map (6 players) setting it'll happen in 11% of games.
Dangerous.