Original post (with over 100% more pictures) http://3dtotalgames.com/critical-success/
The game has been going badly. Your opponent has conquered most of the known universe, taking all of the high resource planets, equipped with a vastly army and all of the tech upgrades. You're down to your last few guys in crappy fighter class ships and one resource barren planet. So you throw everything at your opponent's flag ship, a monstrosity that could be mistaken for a small planet in of itself, which carries a weapon that could annihilate everything you have left in a single shot if you were in reach. The fight goes poorly and you lose almost everything, but one fighter lands a shot on the thing. You roll a 6, followed by a 6, followed by a 6. Theological ramifications aside, the thing explodes and you snatch victory from the jaws of defeat. Welcome to the critical hit!
The inclusion or exclusion of a critical hit mechanic is something that can have a huge impact on how a game feels to play. Mechanically oriented strategic gamers can be frustrated by them, as there is often no way to plan or strategise around them and they feel like a source of randomness that detracts from the depth of the game. On the other hand players who are behind don't like to feel out of the running, so giving them a chance - even a slim one - can improve their enjoyment of the game. Narrativists often like criticals too, as they can make for good stories, especially if everything goes wrong.
From a design perspective criticals can also be a source of incomparables as they allow the inclusion of abilities and items that interact with them. They also give the designer potential to include effects that they find are fun, but too powerful for regular play. The downside to this is that the designer must consider the possibility that a player will try to collect all of the effects that work with criticals and combine them into a game breaking dominant strategy. It is a sad fact of gaming that there are some players who will play the most effective strategy even if it is boring and frustrating for everyone involved, so it is up to the designer to make sure that the most effective strategy is also fun for everyone. Playing the same strategy every game rarely is.
This time it'll work.
So the ideal critical system allows for a player to strategise around it somehow, has critical powers and effects that are not so individually weak that nobody would use them, but the combination of them must not be so powerful that it would be the "one true way" to play. Of course not every player wants this, some types of gamer are happy with a game that's quite highly random and never bother themselves with dominant strategies, but it's more interesting to think about how to solve problems for games that they are applicable to rather than saying "You can just aim for an audience that doesn't care about that.".
Three solutions occur to me:
The first is to make the critical effects not stack with each other; perhaps a player might only be permitted to use one effect that makes criticals more likely and one that makes them more effective. For this to work the effects that make criticals more likely need to be potent and situational. Forming a plan around having a 1 in 10 chance of a crit instead of 1 in 20 is pretty tough. Forming a plan around having a 1 in 2 chance when you're flanking or 1 in 20 otherwise is much more viable
The second is to build a game around combos. Add a lot of effects to the game that can be potentially devastating when combined properly and make the challenge identifying and using those combinations. If you're feeling really clever you could try to make acquiring the parts of and executing combos one valid strategy of several which would generate a really versatile game.
The third is to build a game in which the players have limited control over what powers they get, perhaps by having random loot or skills. The focus of such a game could instead be on how to best utilise what you do get rather than making the perfect build from a free choice of options.
As with anything all of these options have pros and cons and each will be better suited to some games more than others. I do enjoy the variety of approaches available when building games and like to think about how different options will influence the feel of a game. At some point I should probably talk about the difference between systems that produce random outcomes and those that merely feel very random, but now I'm digressing which is a sure sign I should finish up the post. Have fun gaming and may all your hits be crits.