I've played Machi Koro a number of times, most often with in a group of 3 players, sometimes 4. So far the prevailing strategy usually ends up being to roll only one die and to:
- Buy as many Convenience Stores as you can.
- Occasionally buy some Cafes and Bakeries, but primarily:
- Save up for a Shopping Mall.
- Construct the Radio Tower, then the Amusement Park, and then the Train Station.
So basically the Shopping Mall makes Convenience Stores powerful. Even though Convenience Stores activate only on your turn, you have a 1/6 chance of hitting them, and there's a sizable payoff if you do. Meanwhile buy other 1-6 cards so that even if the Convenience Stores don't hit, you can have a 100% chance of hitting something on your own turn. (You might not get income if other players have sufficient Cafes, but that's significantly mitigated once you get the Radio Tower.)
My group has tried this where 1 of 3 players sticks with only one die and where 2 of 3 players stick with only one die. So far the players who stick with one die usually end up with more money. Going for two-dice cards requires a greater investment, and since it's harder to get coverage for all possible dice rolls, provides more opportunities to not hit anything on your turn. It seems hard to justify rolling two dice and giving up nearly guaranteed income. And if most of the players aren't rolling 2 dice, then buying Mines and Family Restaurants is a lot less attractive. If nobody has any interest in buying cards >6, then the game seems broken.
Are these experiences atypical? If not, are there any good house rules to combat this?
Update #1: From the comment discussion in one of the answers, it also occurred to me that if I started rolling two dice, other players could just buy Family Restaurants. Assuming that they also have Cafes, if I roll 2 dice instead of 1, I now have a 1/4 chance of paying other people instead of 1/6. The threat of Family Restaurants thus seems like another disincentive to ever bother with 2 dice. Is my analysis incorrect? (If I take expected values into account, Family Restaurants get even more weight, although I'm notably not calculating the expected values for a Cheese or Furniture Factory strategy.)
I suppose I should try to write some Machi Koro bots and run a few thousand simulations...
Update #2: I've had a few ideas about how to address this (which I've posted as separate answers to this question), but I don't actually play Machi Koro enough with the same set of people to have been able to play-test them.