Friday, October 12, 2012

Innovation, Part 2

Earlier this week I was pretty vague and general regarding what it means to be innovative.  I'd like to fix that today by going into more details when it comes to different gaming genres.

There are several key things you must do and keep in mind if you want to start advancing your skill in a clever and interesting way.

1) State your goals and objectives.  Be as clear as possible when doing so, and strip away ALL assumptions.  Slowly work your way back from your ultimate goal and try to create as small a "tree" of goals as possible.  Then,

We often include extra steps that don't really matter, inhibit ourselves based on assumptions, and stay trapped by how the game is "supposed" to be played.  If you can remove those assumptions and use only the game's hard objectives as your guide, then you will be on the path to coming up with clever ideas.


Consider the (poorly named) MOBA genre.  They are 5v5 affairs and people in many of these communities are very hard-set on how to play the game, on how to compose the team and what to pursue and when.  But many of the things people blindly stick to are not actual victory conditions.  The common meta-game in League of Legends, for instance, is to have a tank in top lane, a mage in mid-lane, a tank-type in the jungle, and a ranged physical damage dealer with a support in bottom lane.  But what are the RULES of League of Legends?

To destroy the opponent's Nexus.  To do that, you must push and take towers.  To do that, you usually need to fight your opponents and defeat them or push them back long enough to take objectives.  To defeat them or push them back, you need to be stronger, and so on.  As you specify the major goal, smaller sub-goals arise.  Do not make the mistake of adding goals that are not explicitly necessary, or you will impede your thinking.

The lowest sub-goal that one can count as an explicit necessity towards victory is becoming stronger than the opponent.  If your team is capable of acquiring more money, higher champion levels, and then succeed at taking towers, it does not MATTER if you have a jungler or a ranged carry or a mage, provided you are able to defeat the other team in battles and then take their towers and eventually their nexus.  Everything else must serve this idea.  As it turns out, having a balanced mixture of damage types, multiple solo champions that don't split experience, and a solid tanky front line capable of initiating is a very strong setup!  But it's not the actual victory condition itself, so slavishly obeying it won't help you learn more.  How can players break from the "normal" way of play and explore more interesting ideas?

Ways people achieve this is to pick champions that don't fit the "role" of their lane, but counter a character that is ordinarily sent against them.  That way the opponent can be shut down and the unconventional character can be stronger than normal, allowing the team to win fights before the other team can recover.  The rest of the team is then structured to cover for the unconventional character's weakness or emphasize their strengths, and thus have a net advantage over the other team.

The real goals are: Win fights and/or push while you're stronger than the opponent, and win the game by destroying the towers/nexus.  If you want to innovate, don't define yourself by any goals beyond that.

2) Stats matter.  If you want to develop an idea, it's not enough for it to be original and surprising, it needs to be SOUND.  And the soundness of an idea can usually be traced to its numbers, its statistics, and its attributes.  It needs to have core elements that outperform other ideas, or can be applied in instances where the others cannot.

It's also worth mentioning that, if you have a lot of hard data in front of you, it can let you look at things more objectively and analytically than you would if you had the game in front of you.  It tells you how things SHOULD be in an optimal environment, and can clue you in to where to improve.


In fighting games, people make a horrendous mistake when they claim that "it's not the character, it's the player."  It is, generally speaking, a combination of both.  If we're playing a version of Street Fighter where Ryu and Ken are completely identical, but everything Ken does has two extra frames of startup, two extra frames of wind-down, does 80% the damage of Ryu's moves, moves at 80% of Ryu's speed, has 80% of Ryu's health, and so on, and then has absolutely no special characteristics to compensate, he would be worse.  Period.  There would be no POINT to playing Ken other than to show a Ryu player how awful he is.  And if you just played Ryu instead, you would always perform better than if you were Ken anyhow.  Game designers generally try to avoid this situations, where characters have strengths and weaknesses that balance them and make them varied.  Game designers aren't perfect though.  Sometimes characters suck.  Sometimes though, they just LOOK like they suck, and are secretly good.

So to start, look at stats.  Look at moves with frame advantage you can exploit, moves with frame disadvantage to avoid.  Check the priority on moves relative to other moves.  Do you have a move that normally kind of stinks, but has this wonky hitbox you can abuse to win in unexpectated situations?  Can you COMBO off that goofy hitbox, or does it happen to shut down somebody's conventional approach option?  Do you have this move that is pretty bad, but has a weird timing that lets you sneak it in every now and then?  Does it work against safe players, risky players, players who think they know your frame data?  Examine how your character ACTUALLY works compared to the others, and play based on that.  You'd be surprised how many people don't break down the components of a character, and stubbornly play the "right" way when it's terrible.

The other thing to remember when it comes to stats is that it's not about having stats that are the *best,* it's about having ones that are *good enough to use.*  Particularly in conjunction with your character's other attributes.  I might have a move with worse priority, but my character has better maneuverability; even though you outrange me, the question could be "can I get in range to land this move first," or "can I bait you into whiffing then counterhit?"

Generally speaking, if you want to advance a character in a fighting game, you ask lots of questions.  You dose those questions with a lot of effort and intellectual honesty, and--surprise surprise--the answers you get will typically serve you well.

3) Replicate your accidents.  Accidents, surprises, and anomalies can often clue us in to how things work better than normal, common occurrences.  And being able to replicate something surprising (especially if it works to your advantage) can put you far ahead of the pack.  The more contrary your strategy runs to general experience, the harder it is for opponents to adapt.


In a fighting game you might find that a move randomly beats another one out.  In Melee you might accidentally dodge one attack by having a hitbox that bends your character around it... then you start replicating the accident and now people constantly find you dodging and counterattacking instantly.  In a MOBA, you might find that two abilities interact strangely and give you an advantage, so you reproduce the effect repeatedly in a matchup to win.  In an FPS, you might glitch a grenade through a wall... then test it, and find it lets you storm a position that's normally impossible to break against competent opponents.

When weird stuff happens, don't just shrug and say, "huh, weird."  Try and solve it, then try and integrate it should it appear useful.

4) Don't over-rely on gimmicks.  A gimmick is a strategy that DOES NOT and CANNOT work once the opponent has knowledge of how to beat it.  If you have a strategy that can be worked into your game with intelligent setups, and it doesn't get shut down just by knowing it *could* happen, then you aren't using a gimmick.  But if it can be checked for and defused every time, it's a gimmick.  It might be tough to spot and defeat, making it a very STRONG gimmick, but it remains a gimmick nonetheless.

Now, obviously if you've got a gimmick that your opponent doesn't know about and you figure it's got good odds of winning you the game, you should use it.  But centering your game-plan around a gimmick and hoping people won't know how to fight it is just putting an expiration date on your skill and success.  Heck, the guy sitting next to your opponent might know what to do and just tell him, and then you're screwed.


A rush in an RTS that can be scouted and shut down without deviating from normal build order counts as a gimmick, and relies on the opponent not noticing what you're doing (or if he does notice it, how to counter it).  If you have ways of forcing or tricking the opponent to build against a different kind of attack, then your gimmick falls more into the range of actual strategy and deception.  If you want a Chess comparison, the Scholar's Mate (a four move checkmate) would be a good example.  Not a perfect one though, because White has opening lines afterwards that are not actually terrible and it doesn't ruin your life if the other guy stops your plan.

5) Synchronize your strengths with your strategy's.  Do not try to play the game in a way that doesn't fit with your skillset.   There is obviously nothing wrong with expanding the things you are good at, but centering your style of play around a critical weakness will keep you from ever finding out if a strategy is sound or not.


In a fighting game, your character has a fast move and you have rather quick reflexes, and this allows you to interrupt moves most other people can't.  So rather than try and just guess or zone people the "normal" way for your character, you put yourself in situations where you can interrupt their moves based on reaction, and acheive success with your character that others have extreme difficulty duplicating.

6) Ask "is there more I could be doing in the time I have?"  Particularly in the RTS genre, being efficient with your time is everything.  Being elegant and taking care of more options with individual actions, or simply increasing the actions you can take in a given amount of time, can open up opportunities for you that others do not have.  In this vein, if you want to further your knowledge, you should actually check the math and stats on what you're doing to see if it's genuinely optimal.


In SSBM, there's a lot of room for people to add fast-falls or time moves closer to frame perfect that they don't explore because of 1) the difficulty and 2) because they generaly win even without it.  But then you encounter somebody who HAS spent that time optimizing and finding where they REALLY have an advantage, and you suddenly don't know what to do.  Be the second person.

7) Develop counters to your own strategies.  Don't just stop when you've found something neat.  Ask what could be beating it, and prepare counters to that.  This is what allows you to give the impression of being unbeatable and one step ahead of everybody.


In a MOBA, if you develop a toon/hero/champion/character/whatever to a higher level than most others, you may find that they will pick your character first so you don't get to use him.  So obviously, having a counter for yourself will give you the major advantage in the next game of a series.


I hope that helps you understand a bit more about the mechanisms behind becoming innovative and creative in your chosen game.  The final point is to never stop asking questions about your game.  What more is possible?  Where else can you optimize?  What makes something work in this situation, but not in others?  Use it to expand your knowledge of the game in total, not just the areas you specialize in, and you'll suddenly find yourself having realizations and ideas past the normal without even trying.

Thanks for reading.

1 comment:

  1. Remember when you trained Light against me for OC3? You're that guy who reveals gimmicks! Except it wasn't a gimmick, it was my whole play style. ;-(