Interesting Choices: Interesting Gameplay pt.8

The fourth and final design feature that can support interesting choices or create interesting gameplay is random elements. This one will take some explaining considering that random elements are widely unappreciated and misunderstood among gamers.  

Adding random elements to a game is a simple way of adding unpredictable factors into gameplay. So in a turn-based game of interesting choices (only stressing knowledge skills) even perfectly informed players may be forced to adapt or anticipate changes to the battle conditions depending on what comes out of the radomizer. Some claim that games with random elements cannot be games of perfect information. We'll examine this idea in greater detail below.

When a random outcome can swing the tide of battle in or out of a player's favor, more cautious players tend to play a more reserved. If the random factor is significant enough, players will think twice about even running a dominant strategy. In this way, random elements can limit the effectiveness of dominant strategies and therefore aid in the emergence of interesting choices. 

Some gamers hate random elements. With a negative tone they call it luck or hax (as if some outside person hacked into their game session with the malicious intent on messing things up for them). From the Smash community to many outspoken members of the larger gaming world, many gamers have a very negative view of random. In this case I don't mean random level generation or random bonuses. I mean the kind of random that affects the action-reaction pairs of a game's core interactions. In some ways I don't blame them. When designed poorly random elements can really take the control away from the players. When the control goes so does interesting choices and agency. But I don't worry about such grim possibilities. Outside of casinos and coin flips, I can't think of many games that seem to be skill based but end up being mostly or purely chance based because of how its random elements are designed.

My Random Defense

In general I think the view that random elements are inherently bad is myopic at best and foolish at worst. To present my "random defense" I'll start with this definition from dictionary.com: "proceeding, made, or occurring without definite aim, reason, or pattern." 

This definition is a little squishy in that it can describe events that are not random at all. This is why I like it. Consider a very complex number generator. This machine looks at 100 different individual number values to generate an output number. To us who do not know that the machine is simply adding up the values and taking the first digit, the number it spits out seems completely random. And even if we knew all 100 factors it looks at, we wouldn't be able to do the math fast enough to make any kind of timely, informed decision. So even with complete knowledge of the machine because we can't really make a pattern or determine the clear reason behind the numbers it spits out, we conclude that it's aimless or random. Technically it isn't truly random (for true random go to random.org). All computers simply use a mathematical formula to generate a "random" number. We still call it random. The take away from this example is that what we know and our frame of reference is important to our considerations. 

So I offer these examples to complicate your conception of random in video games:

Example 1: 

Take RPS. When you play against another human player, isn't your success random? When you throw your hand you don't know what the opponent will throw. There's no definite aim or reason to your move other than that the opponent is trying to counter you though not knowing what you're going to throw. Even if you think you've spotted a pattern in the opponent's moves (like avalanche: rock, rock, rock.) how do you really know that the opponent's aim was to throw 3 rocks the whole time? Is 3 rocks any more distinguished as a pattern than throwing rock, paper, scissors? Or paper, paper, scissors? Just because 3 rocks is easier to recognize after the fact doesn't make it more legitimate as a pattern. But you might argue that you're playing a human opponent. That's the whole fun of the game right? Out guessing/playing the opponent. True, the fun of the game may be playing humans, but the game can be played with a computer. 

  

Example 2: 

  

So what if we play RPS against a computer? I should clarify, this computer doesn't cheat in any way. All the rules of RPS against humans applies to this computer opponent. Are the computer's moves random? Will you give up trying to outwit it because it has no wits? Or should I say organic wits? What makes you so sure that human opponents are trying to out play you and this computer opponent isn't? What makes you think that the computer doesn't have a database of strategies or patterns it uses? If you don't know what goes on inside a human opponent's head then it's hard for me to think of the computer opponent differently especially when I don't know what's going on inside of it. 

   

Example 3:

  

What about a fighting game? Like many real-time games, fighting games are filled with double-blind encounters, where the interplay works out just right so that players can gain or lose certain advantages only if they both act simultaneously (or very nearly so: within 10 frames or so). This feature and the RPS-like interplay loop in fighting games is all about making a guess at the right moments. How is this any different from the RPS scenarios previously described? Is one random while the other isn't. Additionally, what about when you play a fighting game against a computer AI (that doesn't cheat using super human reflexes and execution)?

  

It's clear that making guesses is an inherent part of all real-time games. Guessing inherently involves making a partially informed decision. In some cases you can use knowledge skills to make an educated guess to increase your success rate. But in the end, an educated guess is still not a completely informed decision about the scenario. Put another way, an educated guess is still a guess. So if you have a problem with making anything less than a completely informed decision in a game of perfect information, then most games should frustrate you. 

  

Example 4: 

  

If you have hatred of random elements because you think take away control from the player, I say that all games are built around a lack of control. If you could do anything you wanted at any time in any game, we'd have no structure to build counters and consequences around. When you JUMP in Mario, you forfeit a level of control while you move through the air. When you hit B in Smash Brothers and your characters says "FALCON..." you surrender some control until after you hear the "PUNCH." And when you're stuck in a 20+ hit combo in Marvel vs. Capcom 3 (or most combo heavy games), you technically have no control at all. Playing games is all about taking the limited control you do have and making the most of it. 

   

  

It seems to me that this isn't an issue of randomness, but an issue of dealing with blindness and elements outside of one's control. There are lots of factors and forces in games that you cannot control. One obvious example is the opponent. So what we do to compete is to get as much skill as possible so we can have as much control or agency as possible to obtain victory. I don't see a big difference between knowing all the moves an opponent can make from a given position and trying to plan around those possibilities and knowing the multiple ways a random element can result and planning around those possibilities.

  

In games that stress knowledge (where knowledge = power) the more you know the better you can play. I've found that players who get the most upset with random elements think they know a lot about the game at hand, but they really don't. If they took all the data into consideration, it would be clear to them that their strategy was open to a possibility of failure. With this information perhaps they would change their strategy to minimize the random possibilities or change their attitude. Strategy isn't limited to games of guaranteed results. One can plan just as well for branching or emergent possibilities as one can for random outcomes. After all, this is what a matrix is, the highest level of strategy. 

  

  

From left to right: Link 'n' Launch, Tetris DS, BOXLIFE, Base10

    

If you want to play a game where nothing is random (outside of the opponent's actions) there's Chess, Checkers, and many other board games. These are games of perfect information. And if you can't stand the "random" moves your opponent makes in these games, then there are plenty of appropriate puzzle games out there (see image above). Such puzzle games are all about presenting very complex or deep gameplay systems that you can completely solve if you're willing to learn it. You should already know I'm a huge puzzle fan. I love games like this, and I love games with tons of random elements. 

  

Completely solvable games of perfect information makes great puzzle games, but they do not make good competitive multiplayer experiences. I do not play any puzzle game (or game for that matter) competitively if it doesn't have significant real-time, random, or blind design elements. The random items is one of the reasons why I love Nintendo multiplayer games. Most of the time, I can't stand how pre-calculated games like Chess can be. Having the same initial play conditions every time in turn-based games privileges deep knowledge (LTM) more than anything else. To get this knowledge one must play a lot and most likely study even more. Losing because my opponent studied some aspect of the game years ago can be off putting for me. I'd rather have a bit of random to at least make us both adapt. Even in games that almost entirely stress knowledge skills, random elements can help put more stress on adaptation.  

  

  

   

It's funny how the percentage chance a result will occur changes how we feel about our chances and evaluate our actions. Recall an article I wrote titled Random Knowledge Quirks. In it I discuss how people can think a random element is more than it is. The following are examples of how the frequency of success can influences our perception. Use the randomizer widget above to see how frequently you hit or miss. 

• 1% chance actions happen so infrequently that we tend to forget about them and rarely plan for them. Tripping in Brawl is a random event that only happens on 1% of all dashes. Yes, one can avoid all tripping by not dashing, but the chances are so infrequent that we just dash anyway. Tripping always seems to catch me off guard. 
• 5-15% chance actions are still too infrequent to bank on them occurring. However, they happen enough that it's not a big surprise. Most Pokemon attacks have a 6.25 critical-hit percentage. Critical-hits simply double the damage of an attack. Critical-hits (crits) are an abstraction of actual battle conditions where an attack hits an exposed weakspot on the target. In real-time games players take the occasional devastating hit when they're distracted or overwhelmed. Remember that being overwhelm helps make "anything possible." So for turn-based games like Pokemon, crits achieve a similar "anything can happen" feel. Furthermore, crits give a slight advantage to offensive players. This biased design is common in many competitive games. You'll never crit a recovery move like Rest, Roost, or Recover. So if there's a stalemate between a recovering Pokemon and an attacking one, the attacker will probably win because of the eventual critical-hit. 
• 20-35% chance effects are frequent enough that players beginning counting on them. Though they won't happen most of the time, you'd expect after 5 attempts you'll be successful at least once. In some cases this is great! For example having a 1/5-1/3 chance that the opponent will get paralized from your attack is just the right kind of fraction that is easy to work with. 
• 50% chance effects are deceptive to many players. With equal chances to success or fail, one might expect to succeed about half the time. Such results are likely to be the case after thousands of attempts. But on a smaller scale, you can easily lose 6 times in a row. This can seem infuriatingly unfair. 
• 75-95% chance effects make us a bit overconfident. Because these mechanics work most of the time we get comfortable relying on them and making strategies around them hitting. Some of the most powerful attacks in Pokemon have accuracy drawbacks. Take fire blast. With a 85% accuracy, a missed attack can be a harsh reminder of the move's drawbacks. 

In the end, if you're the kind of player who doesn't want to take any chances, then you should pick the strategies that minimize percentage effects and stay away from many action games. If you choose to play just don't get mad when you fight someone bold enough to take the good and the bad to work together a strategy that compensates for any unlucky occurrences. To a good strategist, everything can be calculated. And a bit of random in a game system can separate the real strategists from the random wannabes. 

Comments

[Ray]

I think a similar problem actually applies to lots of real-time strategy games (StarCraft is the one I have in mind). A lot of the strategy that professional players use has been practiced time and time again, and the players aren't actually critically thinking during the game. I wonder if randomness could ever be weaved into an RTS?

[Richard Terrell (KirbyKid)]

@ Ray

True. Pure game/matcup/experience knowledge often creates these barriers that keep lesser players out and greater players closer to routine/standard pro play. Without fog of war, StarCraft simply wouldn't work. Because extra effort has to be put into scouting information and keeping it secret, there's a greater variety of strategies and playstyles possible.

I've been watching pro StarCraft2 matches on youtube for almost a year now. I've seen some pretty routine stuff and some pretty surprising plays. Another great part about real time games is that even when the pros do their routine strategies, sometimes their execution falters so they have to adapt to the random/unpredictable way they messed up.

If you want randomness in your RTS, check out my articles on DigiDrive's multiplayer mode and my series on Pikmin 2 battle. Links below. I think a lot of RTS fans are limiting their genre by demanding that all games be like Star Craft. There's so much more that the genre can handle.

http://critical-gaming.squarespace.com/blog/2011/1/18/appraising-the-art-of-combat-pt8.html

http://critical-gaming.squarespace.com/blog/2009/5/19/critical-gaming-pikmin-course-week-1.html

[CrowbarSka]

"And if you can't stand the "random" moves your opponent makes in these games, then there are plenty of appropriate puzzle games out there"

This sprung to mind when I read that: http://www.kongregate.com/games/MrJinx/talesworth-adventure-ep-1
The game is based on strict rules which allow you to completely solve a room before you even press any buttons, then execute your solution and see if you got it right.

[Richard Terrell (KirbyKid)]

@ CrowbarSka

Ah yes. Taleworth adventure ep 1. I really enjoyed that puzzle game. I was on my way to 100% the game, but I got stuck on some of the "hard solution" extra challenges. I got 21 of 25 gems.

That's a great game. Last I talked to the creator, he was working on episode 2.

You're right about the game. There's nothing that can't be calculated down to a precise singular outcome. This is why many puzzle games are so great. In no other genre does so few rules create such great challenges in so clear a way.

[CrowbarSka]

Yes, and it even becomes rewarding in itself to be able to apply the ruleset mentally and have your prediction confirmed.

Talesworth caught my attention because at the time I was working on Create which has a similar "set things up while paused, then press play to see if it works" format. However, Create was based purely around a physics engine which brings its own kind of randomness (or at least a perceived randomness as the intricacies of physics are too complex to reliably predict to a fine degree). Because even a microscopic change to an object's position or rotation can drastically alter the results the player can easily lose the ability to predict the outcome with confidence. This leads to frustration, especially when you had something set up to work and then the slightest slip of the finger could ruin it all.

Certainly one of the major pitfalls of using physics in games, I have learned.

[Richard Terrell (KirbyKid)]

@ CrowbarSka

Well put. I've always been a fan of the kind of simplified physics that I could keep track of (or mostly keep track of) in my head actively as I play.

[Kris D]

Your entire argument here hinges on the idea that humans are exactly as predictable as a random number generator, and that introducing random number generators are exactly as deep as introducing a human player.

Let's compare examples 1 & 2. Humans can pick between rock paper and scissors. Computers randomly generate between rock paper and scissors. The difference here is that over time, computers will have a 33.3% statistically even spread. There are a variety of strategies you can take in rock paper scissors, and competitive rock paper scissors is a marvel to behold. http://www.youtube.com/watch?v=_eanWnL3FtM Notice the commentary. I don't think I need to tell you what a Tell is, or what a pattern is. Were humans as random as computers, it would be impossible to figure out what your opponents want to play. This is not the case, even in a game as abstract as rock paper scissors.

I play fighting games competitively, and the one I play best is Virtua Fighter, which I'm certain you've at least heard of. It's a game where every move in the game has a hard counter, which means that if you predict what your opponent is going to do, you will be rewarded tremendously. Why is that?

It's because of a certain thing called Valuation. Different players will value different actions and strategies in a game. If you are able to pick up how your opponent values their actions, you will be able to use a counter-strategy. This is isn't pie-in-the-sky magical thinking, this is all practical stuff that I use on a consistent basis to win against people, people who are ultimately predictable, because they're not perfect abstract machines.