Blog #8 – Luck and Variance in Game Design Part 2

Hey all and welcome back for part 2 of Luck and Variance in Games Design.

Last month I did a brief introduction to randomisation, some of the misconceptions and some of the benefits. In this post I’m going to briefly cover some of the issues of variance introduces and then go over a small system I’ve been faffing with that (hopefully!) helps balance out variance.

(In keeping with the first post, despite my own personal definition differences between Luck and Random Elements for the sake of legibility I’ll just be sticking to calling it luck for this post)


Luck – Issues

So what’s one of the most apparent issues with random elements? Balancing; as soon as you introduce variance the outcome become less predictable and all your planning and balancing becomes much trickier. Less highlight how rapidly things can escalate with a basic example:

In a turn based game you have a hero who can attack for 3 damage per turn, you know now for certain that after 4 turns, the hero will have done 12 damage.

Let’s throw in some variance into this; your hero now only deals 2 damage per turn, but now has a 1 in 4 chance per attack to hit for 6 instead of 2 (conveniently this still gives us the same average damage of 3 per hit). So how much damage has our hero now done by turn 4?

Well given that our hero can hit for 2 per turn at the lowest that leaves us with 8 as our bottom limit. Conversely our hero could (however unlikely) critically hit for all four hits meaning they do 6 per hit, this leaves us with 24 as our upper limit. Overall the possible damages are: 8, 12, 16, 20 and 24.

So where’s the issue with balance? Surely, despite all this, given that the average damage per turn is the same it’s not too dissimilar to balancing the static damage. Over a large volume of hits the damage may well be similar, however due to this variance we have open up scenarios and options that actually could not have existed without it.

Keeping with our hero example, let’s pretend they are fighting a monster with 12 health. Our static damage hero will, and only will, defeat this monster on their fourth turn. However with our critical strike damage hero all these possible scenarios are available after four turns:

  • The monster is defeated in 2 turns (two critical hits for 6)
  • The monster is defeated in 3 turns (a normal and a critical hit for the first and second hits with a critical hits for the third/final hit)
  • The monster is defeated in 4 turns (some mixture of 1 critical hit and 3 normal hits)
  • The monster survives (0 critical hits and 4 normal hits)

Now by adding a fairly simple bit of random element, we’ve added whole 3 new scenarios for our player (and thus us as designers whilst attempting to balance the game) to content with!


Now I could spend loads of time going on about how variance adds more scenarios, and how balancing around the average number can be a bit of a trap (due to thresholds, ‘wasted luck’, such as critical hitting the monster when they are 2 to continue our example, and so forth), but I did promise to only cover these briefly so before I get distracted again, let’s move onto other issues brought around by luck.

Last time when discussing the perks of luck, we touched upon some of the psychological aspects such as ego and accessibility, this time we’re going to cover some of the human issues of luck. Most of the advantages worked towards players who were less familiar with the game and towards levelling the playing field between players of different skills levels, whilst these are good for the benefiting party they are obviously negative for the party that loses out. This can lead to moments where a player who was previously winning can end up in a losing position due to something outside of their control, leading to them feeling cheated. It is very important when introducing chance and variance into a game environment to ensure that the players are informed of this and they are able to play to it (such as trying to work out if an action is worth the potential risk, given the possibility for an outcome influenced by luck).

If the inclusion of luck isn’t well managed (notably if it is used too heavily) players can feel that their actions are meaningless as they feel that luck contributes more than their decisions. Whilst players may play games like this for a short while, it is very unlikely they become too heavily invested, which in turn reduces the likelihood of the game every becoming competitive (although there are notable exceptions such as Poker, though this more down to the level of skill involved alongside the luck). It’s also worth noting that more luck heavy games may be seen as having less mastery involved, or at least reduce the players desire to master the game. Lastly, before I move onto less obvious forms of luck, for some players winning due to luck can be a cheapening experience that reduces the value of their efforts and the overall game experience.


Luck – Additional Forms

Originally I was going to discuss overt randomness and non-overt randomness, looking at some of the less obvious forms of luck in game. Given how large a topic this can be, and my fondness of going on tangents and ramblings I thought it best to provide some links (as extra reading!) to people who have covered this far better than I could:

Richard Garfield:

Mike Stein (on Garfield’s Video, notably see his 3rd bullet point):


Variety Fair – Balancing Randomisation & Fortunes

So now that I have escaped the responsibility of explaining overt and non-overt randomness (and why Chess can be considered to have luck), I’ll move on to how some game systems have approached chance-related events and how some have attempted to make them more fair before going over a little system I’ve been messing with.

In Settler of Catan, players gain resources each turn based upon the result of a dice roll (if the number rolled matches the number on a tile next to one of their bases, they gain that resource). Given that the resources are the backbone of the game (essentially dictating what a player can do) this can lead to situations where you are simply randomised out of the game. Whilst in reality this rarely happens, one of the expansions introduced a deck of cards which could be used to replace the dice roll. These cards are number such that they match the distribution of dice the rolls (Catan uses two dice per roll, so the deck contained one 12, two 11’s, three 10’s, etc…) however, to prevent card counting to a degree, after the deck is shuffled the last three cards are removed (without being revealed) to keep a degree of uncertainty.

Situation draws (such as who should go first, what wins in a fight of equal strength, etc…) are sometimes resolved via randomisation (a common example is a coin flip for who goes first). Sometimes this randomisation is fairly obvious such as a dice roll or a coin flip. Other times it resolution mechanic itself can be fixed, yet the player who benefits from it may do in such a way that can be deemed as lucky. One of Lords of Xidit’s (a fantasy reskin of the more thematically subdued trading game Himalaya) tie breaking mechanic is based upon how far away the players (involved in the tie) are from the player who went first that turn; the further away you are the better (so the player who went last that turn will win any tie). The Game of Thrones board game take it one step further as it’s draws are based upon the players’ position along various influence tracks (one for military, one for political/game state and one for… other). A player’s position on these tracks is based upon a voting system where players bid a certain number of tokens, all players determine how much they want to bid secretly and reveal their tokens at once (so whilst you get to influence the outcome, you can be lucky/unlucky based upon how much other players have bid).

Well as promised from the first blog post, I’ll now go over my own little framework system; Fortunes. A fairly simple system, yet one that requires a bit of adapting to the game you are playing, Fortunes revolves around the idea of giving something to the players who were unlucky (to soften the blow of losing out to random chance). Whenever a player (or players) loses out in a situation due to luck/random chance, they are given a Fortune coin. Throughout the game the players are able to use these coins to influence their luck in minor ways. Now exactly what these ways are is entirely dependent upon the game you are playing (as well as how influential you want them to be, ultimately they need to be less strong than having been in the lucky position otherwise people will prefer to be unlucky in order to obtain a coin).

Here are some possible options of how they could be used:

  • Dice
    • Single dice re-roll
      • Only affects the player who spent the coins, low cost in coins
      • Effects multiple players, but in a non-negative or competitive manner, medium cost
      • Is a competitive roll (such that it could affect other players negatively), high cost.
    • +/- 1 to result
      • Similar costs to above, but likely a little lower
    • Add a dice to the roll, pick one to ignore at the end
      • Due to the control added by this option the costs across all situation are likely to be high to very high
  • Cards (both traditional playing cards and game specific cards)
    • Draw an additional card, shuffle one back into the deck
      • Much like it’s dice variant above, this one needs to have a very high cost
    • Influence a numerical value on the card by +/- 1
      • This would need to be priced based upon the system it is within (and how influential the numbers are)
  • Draws
    • In the next draw-situation which would be resolved by random chance, the player can spend Fortune coin to be the winner (in this situation do not give fortune coins to the losers…)


Whilst there is a whole host of ways you could include these, I would recommend trying to limit them to aspects that contain luck/random variance (in order to limit what variance can influence). If you have a group of consistent players their coins could persist across games (though I would recommend some form of limit in both supplies stocked and numbers spend per game!)

Right I’ve rambled on for quite a good chunk on this post so until next time, have fun and be lucky!

Please Share our Development Blogs with the Community.

Leave a Reply

Your email address will not be published. Required fields are marked *