Psychochild's Blog

A developer's musings on game development and writing.

21 April, 2009

The role of random
Filed under: — Psychochild @ 8:03 PM

I recently got back from IMGDC and had a great time. Since I was traveling, it was a good excuse to bring along my Nintendo DS and play Castlevania: Order of Ecclesia, a game I got for Christmas last year. One interesting thing is that the citizens in town give you “quests”, like in MMOs.

And that’s where it got frustrating. The drops you require are randomly dropped. Ick!

This made me want to post something about the nature of randomness in games.

I’ve talked about random numbers in games before. One design challenge dealt with making randoms more fair in paper RPGs. An earlier one dealt with when to use different types of randoms in games. I recommend reading these posts and the comments to get some background.

Why so random?

What is the purpose of random? Games have included randomness, but why?

Let’s start with the classic “a series of interesting decisions” definition for games. What makes the choices interesting? One option is to have a large number of possibilities. Chess or Go are examples of games which have a large number of possibilities. Because of this, it is unlikely that people will play the exact same game twice. Note that neither of these games have a random element as a core part of the game.

As we reduce the game’s possibility space, choices become limited and therefore more likely to be repeated. Consider movement in a game like Monopoly, which is along a linear path around a board; this path is a much smaller possibility space than Chess a 64 space board in addition to sixteen different pieces of six different varieties. Note that the linear movement in Monopoly doesn’t actually require the player to make any decisions, let alone interesting ones.

So, game designers add randomness to this situation to force people to react to the situation. In Monopoly the dice rolls and decks of cards make the situation less predictable. The interesting decisions don’t involve movement as in Chess, but rather the purchase and trading of properties.

Good randoms

So, randomness is a way to add more possibilities to a limited possibility space. Let’s consider a computer game now; in an ideal raid situation in WoW, for example, a balance Druid will cast only about 5 different spells in a spell rotation. This is a pretty limited situation, even if people used a larger number of abilities.

How do we increase the possibility space with randomness? The Druid from might get a critical hit randomly and do a surprising amount of damage and gain too much aggro. Or, a randomly fired enemy ability might have taken some of the character’s mana and it has less resources to use for casting.

Note that player choices still influence some of these decisions. For example, a player can chose to adjust some stats to alter the chance of landing a critical hit. Or, the if the monster’s ability to drain mana has a range limitation, the player makes a choice if moving (and not being able to cast while moving) is worth not being affected by this ability. I believe this is an important part of good randomness: it still allows for players to make interesting choices.

Bad randoms

What happens when the randoms take away or have no way for players to make meaningful choices? This is where we get bad randomness in games.

The most notable example in MMOs is the random drop for quests. If a quest requires you to get drops from an enemy, it can be frustrating if it seems to take a rather long time to get the item(s) you require. In a typical “random with replacement” system where an enemy has a less than 100% chance to drop the required item, it’s possible (even if not probable) that you may never see the drop. In this case, you have no way to increase the chance of a drop, the player’s only choice is to either cointinue killing the enemies or not, which means either trying to complete the quest or not. Not much of a choice.

This is the same situation I ran into with the Castlevania game for the DS I’ve been playing. The items I need to complete the quests (and open up more healing items and equipment) are randomly generated. Although the player can affect the drop rates somewhat by increasing the “Luck” stat, the same principle applies: it’s possible (even if not probable) that the item may never drop. On top of that, the monsters that drop the item are not specified, so without outside information you may not kill the right monsters.

Design motivations for bad randoms

So, why are bad randoms used? In the case of quest drops, it’s to make the content seem longer than it is. If the quester only had to find a kill one common enemy, the quest would seem short. Even killing a small number of enemies with a 100% quest item drop rate is likely shorter than a lesser drop rate. At 50%, the player has to, on average, kill twice as many enemies to get the same number of items. With randomness, the number killed will vary, making the quest seem less rote. Of course, we can reduce the frustration in this type of random by using a “random without replacement” system to put a cap on the number of enemies you must kill.

The other problem is that designers, like most people, don’t always understand how randoms work. It’s very simplistic to look at “gather 10 items with a 50% drop rate” and mentally average it out to 20 kills even though that’s not what it really means. In a simple “random with replacement” system it means that the player will have to kill between 10 and an infinite number of enemies to get the items required. Of course, this same lack of understanding of probability is the same thing that allows gambling, like the lottery, to get people to play.


Ultimately, the important thing to do is make sure that the player has meaningful input. Any situation where player input is meaningless, whether it involves random or not, can lead to frustration. Adding randomness to a lack of meaningful control is the shortest road to frustration in a game.

What do you think? What is the proper amount of randomness to an MMO game?

« Previous Post:


  1. Bad randoms would IMO include randomly spawning monsters, or similarly disruptive events. There’s nothing more frustrating than not being able to walk an apparently clear route.

    It gets a bit better if there’s a concept of “safe” zones and “unsafe” zones, where only the latter spawn monsters randomly. At least here you’re informed that you have to be on your toes.

    But I still don’t like it.

    Comment by unwesen — 22 April, 2009 @ 12:05 AM

  2. I like random drops of items. loot in general. But I do not like random quest item drops. This is just extending the time you need for the quest. OK, it is fun to get a drop if it is not guaranteed, being lucky. But this kind of randomness is too closely tied to collection quests. Quests that say “bring me 5 Gnoll Paws” (which drop randomly from Gnolls) are just kill quests without a fixed number in disguise.

    Some interesting change to random drops was done in WOTLK. Maybe you did not notice it, but if you did not get a quest item drop from a looted mob, the chance to drop on the next mob is increased more and more by the game… I cannot find the source anymore, I think it was Ghostcrawler who dropped this information.

    I actually liked it, I never despaired because I did not get the final quest item drop for a long time in WOTLK.

    Comment by Longasc — 22 April, 2009 @ 12:16 AM

  3. I think you summed it up pretty well Brain :) Randomness is great when used to expand the range of situations a player may find himself in. Bad when used to determine something that should be determined equitably, rationally, and predictably. I think the knottiest randomness problem that I know if is land generation in the Civilization games. I’ve sort of derived this rule of thumb from that problem: ‘If randomness within a discrete situation produces a significant difference in quantifiable outcome for two players of the same skill, it is bad randomness.’

    Comment by Bret — 22 April, 2009 @ 6:55 AM

  4. Random drop quests need to curl up and die, in my opinion. I’ve played WoW a lot, and some of the quests are ridiculous; in one you have to gather 30 troll tusks, which have a 10% drop rate. That’s 300 trolls killed, on average, or around 75 minutes killing at top speed for a DPS character in levelling gear. To finish one single quest. There’re other quests that require you to get a single very low drop rate item, in the range of 2 – 5%.

    A perfect RPG, in my opinion, would never have any quest in the form of “Kill X boars”, “Kill X more boars”, “Collect X items” etc. where X is any number larger than about 5. If you can kill one boar, you can kill them all. It just takes, nay wastes, time. And it gets old. Instead, I’d have quests like “Retrieve item X from dangerous creature Y. Bring friends.”, “Kill the boss mob of this group of mobs so they’ll stop attacking my farm”, etc. Set goals that will be somewhat difficult but don’t require unreasonable repetition.

    Also, you didn’t touch too much on randomness in combat, but that’s even harder to balance. In WoW they’ve been trying to tone down random burst damage ever since launch – people don’t remember the 50 times they beat an unbuffed warrior 1v1, but they sure as hell remember that one time that a fully geared warrior came in with the old 40% enrage up, plus battleground berserker buff, plus player buffs and recklessness active… and took their head off in two swings. Combat without an element of randomness becomes boring in PvE, but too much random is bad for PvP.

    Comment by Tane — 22 April, 2009 @ 6:42 PM

  5. Dammit Longasc, I wanted to say that decreasing randomness with each kill without quest loot could be interesting, and here you go and say WotLK did it !! :)

    Comment by Modran — 24 April, 2009 @ 2:02 AM

  6. unwesen wrote “Bad randoms would IMO include randomly spawning monsters, or similarly disruptive events. There’s nothing more frustrating than not being able to walk an apparently clear route.”

    Actually, it seems to me that you are mixing up two properties, randomness and not-making-sense, and blaming the wrong property for your frustration.

    One thing that I find irritating about many games, WoW included, is that “boss”-flavored encounters have important elements which don’t make any sense in terms of the rest of the world. Fighting Ed Van Cleef (?) in the Deadmines, his associates spawn as he gets injured. And that makes sense how? And you could anticipate it before the fight how? And it is satisfying, how, to figure out a strategy of killing the adds ASAP so that as VC becomes more injured more you never need to fight too many at once?

    To me the main problem with monsters spawning and lurking in cleared paths is similar, that it doesn’t make sense, and relatedly that you can’t see what’s going on or respond to it in any meaningful way.

    Is randomness the problem for you, or meaningless incoherence? Is a system where monsters predictably materialize on cleared paths less annoying than one where they materialize randomly on cleared paths?

    I think the right way to fix that frustration (while keeping a similar challenge in the great scheme of things) would be not to reduce randomness, but to reduce incoherence. How about a system where monsters wandered somewhat randomly over larger areas than just the path you wanted to use, and where you could see some of what was going on, and do sensible things to evade them? And, ideally, it would make more sense in terms of the game world than random materialization (an ecosystem of monsters at one end, or ghosts-spawn-in-the-graveyard at the other end of the plausibility scale).

    Given gameplay involving evading and picking fights and such, I really don’t think randomness would make it less interesting. Mobs which react by predictable clockwork can be an OK design choice in some cases, and predictability does avoid various implementation difficulties. But for any challenge involving evading mobs or picking fights with the weakest groupings and/or in the best terrain, a lack of randomness tends to make things pretty dull, especially for replay, and it’s not too difficult to implement suitable randomness. I expect to see a lot of aimlessness in a real animal or person searching or wandering, and much less aimlessness in urgent matters like combat. Randomness occurs in combat, too, but random aimlessness seems easy to simulate, while it’s hard to simulate intelligent responses to randomness in intensely goal-directed stuff like combat.

    I notice that in sneaking-related games I’ve played (Deus Ex, e.g.) the mobs seldom had much aimless randomness in their motion, and that’s perhaps evidence that it’s harder than I think to implement it well. But those games didn’t have as much replay value as something like Civilization, where many random things determine the way one’s enemies happen to be distributed, and then one has lots of information (though not complete information) to use to choose how to deal with it. Replay value is often a consideration, especially now that so many game designers are chasing a subscription gold mine, so I think randomness will often be a good idea.

    Comment by William Newman — 28 April, 2009 @ 9:59 AM

  7. I find your blog about randomness interesting and I thought I might comment about my ideas about your question from a different angle. Personally I am not a big MMO player, I normally play FPS or Strategies games which have a high level of randomness to the game play, from experience in the game to luck to the ability of a player to use the special skills of their units. So I am not so intuned in to all the details of MMO games like WoW or Eve.

    I think there is no median level of randomness in a MMO. All random factors can be contributed to the game designer wanting to afflict a particular level of uncertainty in the MMO to players. As to the idea that, for a mission you need to kill X number of creatures to get an item will frustrate players is not the point of MMO. A MMO is designed for you and a group of friends to complete a mission and not meant to be played individually; after all it’s called Massively Multiplayer Online for a reason. Otherwise you could call it a RPG or other types of games designed for individual or a few players.

    Randomness in games is designed to increase the probability of different results each time the game is played, like you said. As I said before the game designers wants to in part a level of uncertainty to the mission, in the case of collecting tusks, it would be also quiet boring to kill one boar and get the one tusk as there is no sense of achievement, your farming like a farmer.

    One reason the game designer might have made missions like collecting tusk, might be so you will ask other players if they have excess items so you can purchase these items from other players (increasing what I think is the main drive of MMOs, the social aspect of the game, communication between players).
    I make I note that I have not commented on your example of Castlevania’s on DS as your question was about MMO.
    If I have offended anyone it was not my intention. :)

    Comment by Andrew C — 5 May, 2009 @ 5:56 AM

Leave a comment

I value your comment and think the discussions are the best part of this blog. However, there's this scourge called comment spam, so I choose to moderate comments rather than giving filthy spammers any advantage.

If this is your first comment, it will be held for moderation and therefore will not show up immediately. I will approve your comment when I can, usually within a day. Comments should eventually be approved if not spam. If your comment doesn't show up and it wasn't spam, send me an email as the spam catchers might have caught it by accident.

Line and paragraph breaks automatic, HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

Email Subscription

Get posts by email:

Recent Comments


Search the Blog


November 2019
« Aug    



Standard Disclaimer

I speak only for myself, not for any company.

My Book


Around the Internet

Game and Online Developers

Game News Sites

Game Ranters and Discussion

Help for Businesses

Other Fun Stuff

Quiet (aka Dead) Sites

Posts Copyright Brian Green, aka Psychochild. Comments belong to their authors.

Support me and my work on