Our fellow 1PGer Kevin Shaud posted this article on his BGG blog (Kevin's Korner) yesterday and, upon reading it, we decided it makes excellent Grumpy material. Even if we love Maximum Apocalypse, we begrudgingly admit his remarks are right.
Variability in a game is a good thing so long as the game is well-designed. Variability increases replay value and keeps the game feeling fresh over time and multiple plays.
Randomness in a game isn't necessarily a bad thing either, but the type and number of sources of randomness certainly need to be considered. When done properly, randomness adds elements of tension and uncertainty.
Where I think game designers sometimes get into trouble is equating variability and randomness. They design a game with a half dozen sources of randomness and tout that the game is highly variable due to multiple random elements. The problem here as I see it is that you begin to lose control as a designer when you depend too heavily on randomness to provide variability. You can't effectively balance scenarios and provide a consistent play experience when the game's difficulty can swing wildly due to multiple sources of randomness. One play might be laughably easy while the next is nigh impossible and the only thing that changed was the outcome of all the sources of randomness.
One game where I found this to be an issue is Salvation Road. It has been a while since I owned/played the game so I won't have the correct terms for some of the game elements. The locations and where they are placed on the board is random so that can affect how much effort it takes to move back and forth. The tribute requirements are randomly chosen and hidden from player view. These are the resources you need to drive to salvation and win the game. There is a location that allows you to reveal one of the tiles (there are three I believe), but it's difficult to spare precious actions using that location and from what I remember, it gets tougher as the game progresses due to threats. There are event cards that you draw at the end of each round (I forget the name - apocalypse cards maybe?) that are brutal and a punishing source of randomness. Generally, it's safer for players to be in the compound, but it's difficult to accomplish much of anything if you're holed up in the compound with any frequency. There is also at least one card that injures anyone at the compound so safety is not guaranteed. You roll dice to accomplish most actions and this often results in injuries to the characters, which slow them down and reduce their effectiveness.
Each of these sources of randomness wouldn't be a problem in isolation, but they are crippling in aggregate. Depending on the outcome of all the randomness, there are likely unwinnable game states where there just aren't enough actions available to win the game no matter what choices you make.
Another glaring example is Maximum Apocalypse. It's a game that I initially enjoyed, but upon another play realized that all of the randomness kills the game because it can result in ridiculous situations developing. For example, I was playing as the Fireman and the Army Ranger. Until midway through the game, the Ranger wasn't able to attack enemies on his tile (in front of him) as his only weapon was long range. How ridiculous is that? An Army Ranger that can't attack enemies right in front of him? He got ravaged each turn by three enemies in front of him and could do nothing about it. He couldn't even get away from them because enemies in front of a character follow them as they move.
The tile set up is randomized (including the objective tile), the monster deck is randomized, each of the three resource decks are randomized, the player decks are randomized, and the monster spawns are random. Your characters only draw four cards to start the game and get one free draw on each turn. So without taking actions to draw cards, it takes pretty much the entire game to draw through your deck. If cards you really need are near the bottom? Oh well...randomness. This game is a prime example of a designer's over-reliance on randomness to generate variability. The objective tile could be right next to your van making the game super-easy or it could be five locations away. It's the same story with the locations where you can gain resources. It might take multiple turns to even get there and then what if the few cards that you need are buried at the bottom of the deck? Too many random elements cause chaos in a game and make the playing experience frustratingly swingy.
Variability needs to be built in carefully by using some random elements, but controlling others within a pre-defined range so that the difficulty arc is kept within reason. It's unfulfilling to play a game that turns out to be way too easy and maddeningly frustrating to play a game that feels all but impossible to win. Balance in a game comes from careful design choices and one of the most important choices is how to build in variability without letting randomness run rampant.