Score in Videogames: Why No One Cares, Why They Should, And How We Can Make Them

Since I wrote my related post, “On Score“, I’ve come to a lot of new realizations about score and its strange, rocky relationship with digital games.  Our history has been so rife with problems that there are many of us who believe that score is necessarily, inherently irrelevant.  I know people who say that they “don’t care about score in videogames”, without even talking about a specific game.  While I’m certain this is a faulty idea, I feel that I do understand why they feel that way.

Historically, videogames have done a completely awful job of implementing score systems.  I’ll get into exactly why this is a bit later, but for now it should be no surprise that anyone is tempted to dismiss score outright, if they were raised in this time period with scoring systems like that in most Roguelike games or even Tetris.  I can’t think of a single digital game that has a really great scoring system;  probably the best ones are very early games like Galaga, but even they have issues.

It’s worthwhile to recognize that score is an abstract gameplay concept.  It is not a thematic element, and it is (or at least, should be) implemented a bit differently in each game you play.  To “not care about score” is similar to “not caring about position” or “not caring about resources“.  These are abstract elements and whether or not you care about them will be based on what their endogenous meaning inside the system is.

Consider a sport like baseball, or football.  Obviously, if you’re going to play one of these games, you’re going to care about score.  Very, very few people watch a football game with no interest in knowing what the score is.  If you’re going to watch a game, and not care about score, then what keeps you watching?

I’ve heard the argument that this is because those games are multiplayer.  I don’t think that this is the reason, though, because actually, all score-based single-player games are technically multiplayer.  You’re competing to get the new high score – a high score that has been set by another agent.  That other agent may be another human, yourself, or one of those preset ones that NES games came with.  You don’t play against each other at the same moment, but does that mean Golf isn’t multiplayer?

Or boardgames:  surely, people playing The Settlers of Catan, Carcassonne or Ticket to Ride are all caring about score, right?  I mean, if you don’t care about score, these games really end up making very little sense.

So this begs the question:  why?  Why do people generally care about score in every kind of game except digital games?

 

The Smaller Cause:  “Completion”

I’ve talked about this idea that videogames, at a certain point, started to become “about completion”, not about what games have always been about, which is “pursuing mastery”.  This point was, unsurprisingly, the exact same point at which games started tying their play to a linear story.  Of course, stories fundamentally have beginnings, middles, and ends — they are a sequence of events, by definition — and so it seemed to make sense for games to start having an “end point”, since we were now tying these to stories.

This was a terrible idea that caused all kinds of problems that we’re still struggling with today, but one of the subtler effects was that we started having a lot of games that had both a completion point and a score system.  Games like Sonic the Hedgehog and Super Mario Brothers.  Every single person I knew was utterly confused by the inclusion of “score systems” in these games.  Who cares about points in these games?  The objective is obviously to get to the end and “beat” the game!

In Super Mario Brothers, the score really was irrelevant.  The game had a completion point, and it was pretty clearly all about reaching that completion point (rescue the princess!).  The game also was not randomized, so it didn’t really make sense to play it over hundreds of times and compete for score.  This series, and hundreds of other early popular games, set the tone for how people would think about score in the future.

 

The Larger Cause:  Videogames Have Been Doing It Wrong

You don’t care about videogame scores, because videogame scores are simply not well-implemented.  Even without the “completion” element, scores are not built into games in a way that would make us care about them.  The first problem is in the nature of the kinds of numbers videogame scores produce.

Look at a team’s score in baseball:  “4″.  What does that score mean?  Well, it means the team got four runs.  How about a score of “21″ in American football?  Most likely, it means the team got three touchdowns.  Same can be said for basketball, ice hockey, and almost every other sport you can name.

Boardgames are a good bit more complex in general, and they therefore usually have more complex scores, but I know of few boardgames where the scores routinely go well over 100.  A score of over 100 in Dominion is very, very good.  It’s practically unheard of in Puerto Rico.  Ticket to Ride’s score counter on the edge of the board only goes up to 100, and you usually only go a little bit above that in a game, unless one player just dominates.

Now let’s look at some video game scores.  You play a game of Galaga and then it tells you your score was 744,315.  Okay.  You play a game of Pac-Man and get 253,150.  Cool.  You play a game of Civilization and get 1,351,881.  I think my best Dungeon Crawl score was around 33,000.  Well, alright.

These kinds of numbers are extremely hard for people to derive any real sense of meaning from.  I can’t find a link to support this, but I swear I remember reading that the reason Blizzard split up their Starcraft 2 leagues into “Divisions” of 100 people was because after the number 100, humans have a difficult time attributing any meaning.  People don’t have a good innate sense for how much larger 551,365 is from 89,051, but they do have a very good sense of how much larger 70 is from 20 (this may be – and I’m purely speculating here, I haven’t looked this up – because early evolutionary humans traveled in groups of sizes roughly between 20 and 100?  Someone feel free to correct me if not).

Why 10, and not 1?

Careful with the Equations!

Equations are the other reason videogame scores suck.  I often think about my favorite boardgames and how each time I score a point, I take a little victory point chip and put it in a little pile in front of me.  In Through the Desert, I earn VP chips for connecting to oases, for picking up watering holes, and for getting the longest caravan.  I do those things, and BAM, I get a victory point chip.  There are also points you get for territory you control, but that also has a very direct look, and feel to it.  I see that space that I personally carved out with my own actions and decisions, and I count up its size.  Bam, that’s worth 12 points!  I fully understand this feedback that I’m getting here.

Now compare that to a videogame’s scoring system.  Very frequently, there is some weird equation that goes on, adding up tons of arbitrary numbers, and then dividing it by some numbers, and applying a multiplier for this, and so on.  Eventually, a number comes out the other end:  1,135,640.3.  Oh, okay.  I don’t have any feeling of connection to this number.  I can’t really see a direct relationship between my agency and the score I got.  If the score was 500,000 off (that’s 50% larger almost), would I really be able to call it out?  I’d probably go “Oh, okay” either way.

We have to be really careful with equations that have multipliers and division in them.  These serve to disconnect players actions from the feedback they get in the form of score.

 

Focus the Player On Your Game’s Actual Goal

Too many games are confusing about what their actual goal is.  Team Fortress 2 is one of my favorite digital games, but it’s guilty of this.  Why, when I hit TAB, am I even allowed to see my personal, individual score?  Does this have any bearing on the outcome of the game?  It’s a team game, and only a team can win or lose, yet — and feel free to call me out on this — many if not most TF2 players are probably a bit more concerned about their personal score.  This is because the developer showed information that wasn’t relevant.  The same thing would happen if you had a visible counter that ranked people by “most time they melee’d the floor” — you’d have people meleeing the floor all the time instead of trying to win the match.

In roguelikes or other similar games, the “fake goal” is “getting to a higher (deeper) dungeon level”, which is a vestigial thing from our unfortunate history of tying games and stories together.  The real goal is getting a high score, but I’ve had arguments with many who say that they don’t care, that they really care about getting to a higher dungeon level.

For AURO, our solution is to not even show what dungeon level number you’re at.  Neither during the game, nor at the end of the game, would you see this information.  Why not?  Because it’s irrelevant!  You should just as soon ask that the game display how many times you “canceled a spell” or “walked north”.  It’s not relevant to the goal of the game, and displaying it only confuses players.

 

Videogames Can (And Must) Do It!

I want to be clear:  it’s not at all due to any inherent property of digital games that they are so bad at score.  Videogames can do it just as well as any other kind of game;  it just requires a little more discipline on the part of a game designer since there are so fewer physical and practical limitations.

I’m not saying that you need to see some digital version of a VP chip fly into some digital bag every time you do something in a game.  Videogames aren’t boardgames, and they aren’t sports.  They are their own animal, and they can afford to be a little bit more complex than either of these things.  But you really have to have serious discipline about it — we’ve been going off the rails with this since the beginning.

Videogames must do this, because single player games require some system of score in order to be endlessly replayable.  And learning to master a game, just like learning a musical instrument or learning to paint or learning any other skill, should be an art that you can explore for the rest of your life.

Posted on 05/09/2012 by keithburgun. This entry was posted in Game Design, Score. Bookmark the permalink.
  • Matt Kloth

    I agree with all of your points. I think being able to actually understand and feel attachment to a score makes a difference. I think that idea goes further into game design also. It’s very common for video games (especially rpgs) to have combat systems so complex nobody can intuitively wrap there head around it easily. In the tabletop world humans have to be able to figure out how much damage something does, but videogames often create byzantine formulas that only the most dedicated bother to learn. That creates a huge problem for players who want to improve their playstyle. They have to crack open a spreadsheet just to get 5 more dps or figure out what talent to pick.

    The 100 person limit made me think of the monkeysphere article over on cracked: http://www.cracked.com/article_14990_what-monkeysphere.html

    • http://www.dinofarmgames.com keithburgun

      Ya. I think that the “byzantine formulas” is just classic weak, bad game design. It’s like creating a smokescreen of noise so that people don’t see how boring your actual mechanisms are.

  • http://blown-to-bits.blogspot.com/ Kdansky

    When you made the point with the victory chips, I realized something; Video games have been implementing score-like systems more than ever in the recent years. You do something awesome, and then you get a huge “DING” event, and your score goes up by one. They call it “level”, and I have yet to see more than one game where your level doesn’t go from 1 to just below 100. We care immensely about those numbers going up. I wonder if you could put that to use in Auro? Instead of giving out one ability per floor, give one ability per 100 score achieved, but since you tie score-payout to reaching of the stairs, this would do exactly what you want anyway, except trick people into their WoW-mindset for the greater good?

    How not to do it: BlazBlue has a scoring system. I have to look up what the numbers are called, because they are around twenty to thirty digits. That’s not how you do it! And they also have zero relevance, apart from Score Attack mode, which nobody plays, and nobody cares about.

    • http://www.dinofarmgames.com keithburgun

      The problem with “one ability per 100 score achieved” or something similar, is that the game actually gets *easier* for players who are better at the game, and harder for players who are worse.

      Like, obviously if you’re better at, say Tennis, the game is effectively easier for you, naturally. But, imagine if Tennis had some rule where it was like “if you gain a point, then you get a giant racket for the next round”.

      Players who are good don’t need to be showered with unnatural rewards, and players who are not good don’t need to miss out on basic gameplay stuff like getting more abilities. It creates a snowball effect or a brick-wall effect, neither of which are desirable.

      I do agree with your general Ding! scoring type stuff but usually in modern games, it’s not like you’re competing with a high score. Instead, you’re just collecting.

      • http://blown-to-bits.blogspot.com/ Kdansky

        You are right, it leads to annoying slippery slope difficulty. I really despise CoD or League of Legends for that enough to not play them.

        I have a neat idea though. What about getting stuff for scoring high, but don’t make them player advantages? Or even make them unlockables? Background tile sets, or a new staff for auro to carry, or green demonic flames (work the same way, but are green). Or even go the route of Desktop Dungeon: High score results in unlocking more enemies, which results in a harder game for the better players. Which is absolutely okay in my book. And I think they should be “badly” paced: 90% of all stuff should be totally up-front, such as “100 points” -> unlock basic tile set. That way, you train the player to go for a high score, but he will have unlocked the actual content of the game long before he’s close to getting good. I mean, it really doesn’t matter if your first three rounds only have a single tile-set. You would not even realize that there are more tile-sets when you can’t get to the second floor anyway.

  • cuc

    Hardcore shmups are all about score, and I imagine a devoted player of a given shmup would know how to parse its score, even though these score systems are, as you said, highly inaccessible to outsiders.

    But I think you already knew that :-)

  • http://www.gruesomegames.com/ Darren Grey

    I think a game must have a score or an end goal, but not both. An altruistic statement that likely fails in many individual cases, but it’s a good general rule of thumb for me. If the game is about getting to the end by whatever means you find most enjoyable then a score will only interfere.

    The only game in which I’ve ever truly cared about the score is Geometry Wars, an arcadey shoot em up where you get a gradually increasing score. The mechanics of the game tie in with it very clearly – you get points for killing enemies without bombs, and your multiplier goes up with number of kills as long as you don’t lose a life (thus making the use of a bomb or the risk of a life both very important decisions). It also, in spite of its minimal interface, has the current score and top score displayed in a large font at all times. The makers obviously had the score in mind at the centre of all design decisions, and this is communicated in a very clear way to the player.

    In too many games the score is a tacked on thing, often because it’s felt “It’s a game, players want a score”. I think this bad approach shows in all the poor score systems that have no real link to the gameplay – or worse, detract from it. Score, as you say, must be a designed thing and encourage gameplay in the right way.

    I think, if done right, achievements/trophies can be a good replacement for a score system. They can be used to encourage the player to try out different gameplay tricks, explore new areas/features of the game and get better at the game in certain ways. They also are awarded for discrete things rather than scaling up to high numbers. Unfortunately achievement systems are just as badly misused as score systems…

    • http://www.dinofarmgames.com keithburgun

      >I think a game must have a score or an end goal, but not both.

      Yeah, I agree. Having both is confusing for players. In fact, Auro has a story mode (that functions also as a tutorial) with an ending and that’s substantially easier than the Normal Game. In Story mode, we’re considering disabling the score system, since it’s kind of irrelevant to that mode (then again, it’s the mode that teaches players to play the real game, so… not sure).

  • http://PlayerDuel.com Yaniv

    I think that LB is a must in any game but you cant change the way all of us as a player fill about it.
    We checked it in the past year and found that most of the people should like to win a tournament instead of getting a score that will be ranked as number 256,734 in a game.

    So we changed the way players think by helping them to compete in each game that support us.
    The score doesn’t matter any more….only winning!!!

    I wrote this cause I really believe in it. I have to say that I’m the CEO of PlayerDuel as well.

  • http://roguetemple.com/z/ zenorogue

    What’s wrong with the scoring system in Tetris?

    I don’t agree that end goals and scores don’t mix…

    As far as I see it, the purpose of score in the traditional roguelikes is to have an automatically organized list of past characters who are closer and closer to the goal. This is very nice. I don’t care how it is calculated, I don’t understand the numbers shown, but it seems to work reasonably, characters who achieved more seem to be higher on the list. And I love the roguelike traditions here (fully described cause of death, 100 entries). I don’t know what is confusing there.

    On the other hand, in most roguelikes I find that score is quite meaningless once you have characters who have already won, and probably it would be not very reasonable to compete for high scores with someone else. But it is possible to have a meaningful system, where getting a high score is a nice challenge for the best players. Maybe you are right that it should be shown only when requested, or as a unimportant stat instead of a big thing, to avoid confusing new players. And that it should be something clear, instead of a “byzantine formula”.

    There is a “Dunbar’s Number” theory which say that a human can maintain social relationship with about 150 people… not sure whether it is related, but you might want to read about it.

    • http://www.dinofarmgames.com keithburgun

      >As far as I see it, the purpose of score in the traditional roguelikes is to have an automatically organized list of past characters who are closer and closer to the goal.

      Well, that could be the case, but I think it’s less interesting that way. If that’s the case, you lose 100% of your games, and then finally you win.

      It being score based, rather than completion based, is better, because it’s renewable perpetually. You win when you beat your high score. You lose when you don’t.

      >in most roguelikes I find that score is quite meaningless once you have characters who have already won,

      Roguelikes should not have a “game completion state”, this is a vestigial design flaw of almost all of them. My own game 100 Rogues even has this flaw.

      Dunbar’s Number sounds interesting, I shall look into this.

      • valrus

        Dungeon Crawl Stone Soup, which I will defend to the death, has both score and a win condition, but (some) people will still try to beat scores for a given build even if the current high score was achieved by someone who won the game. There are kind of two tiers to the “win” idea, then: less-good players can just aim to win, which is difficult but achievable and feels like an accomplishment in a way I’m not sure “get X points” ever could; power players can try to beat each other’s high scores by doing speedruns, getting more than the required number of runes, etc. By the time a player has won with enough different builds that that variety is no longer interesting in itself, which will take a long time, he or she will probably be ready to start trying to beat others’ scores, or their own.

        I love your blog, by the way. It seems to speak to why I find so few games compelling these days.

        • valrus

          Didn’t notice someone already commented on Crawl below.

  • Lavaflyer

    Just wanted to point out a few things. Earlier you talked about people playing carcasoone careing about score and I agree but i found it funny that you didn’t talk about how high scores get in carcasoone (anywhere from 100-400 with alot of expansions-atleast). Obviously you did this because you were basing it off of 100. And also didn’t know if you were a where of these next two things or not. First of all, your game 100 rogues (obviously has both a score system and ‘completion system’)but its scores just like the arcades frequently go too high. For example my highest ever score was 296,140- much higher than 100 but it does sound like you’re fixing this with auro. I similarly like how you’re doing the hp system because what is the point of having enemies doing 10dmg and you having 100hp when you could have 10hp and take 1dmg per hit– I’ve always had this kind of problem with hp systems -lol. And finnally, 100 rogues also exhibits weird multiplier equations in its combo mode which also (as you described make it hard to distinquish score).

    • http://www.dinofarmgames.com keithburgun

      Yeah, expansions tend to make things go a bit nutty, and I was just thinking of the base game. However, even 400 is nothing compared to most video game scores.

      I think 100 Rogues has this same flaw in its score system, yes. I also didn’t design Combo Mode so I don’t really know/remember what its equations are, but yeah, I’m sure they’re a bit on the weird side.

      Yeah, for Auro, everything is boiled down.

  • http://outsideyourheaven.blogspot.com/ Matthew “Sajon” Weise

    Maybe score needs to go, so story can flourish. :)

    • http://www.dinofarmgames.com keithburgun

      Score has never gotten in the way of story. Interactivity, however, does. And I do think we should do away with interactivity so that story can flourish. Novels > Choose Your Own Adventure Novels.

      • http://outsideyourheaven.blogspot.com/ Matthew “Sajon” Weise

        Sassy.

  • http://outsideyourheaven.blogspot.com/ Matthew “Sajon” Weise

    Seriously, I think it just depends on what the player wants. I agree score should be meaningful, but it (and interactivity) does not clash with story in some fundamental, cosmic way. It depends on how it’s implemented. Not all games should have stories, of course. And not all games should have scores. Isn’t the world big enough for both kinds of games?

    Also, the interactive fiction movement has demonstrated for years that you can do a lot more with story and interaction than Choose Your Own Adventure. Are you aware of IF at all?

    • http://www.dinofarmgames.com keithburgun

      I do think that they clash in a fundamental (dunno about “cosmic”) way. I don’t think it depends on how it’s implemented, except that “the less it is implemented, the less the game will be damaged”. I wrote about this pretty extensively here: http://www.dinofarmgames.com/games-hurt-stories-stories-hurt-games/

      The world is big enough for whatever. I’m not saying people can’t make these things, I’m saying they shouldn’t. Because sure, the world is big, but our lifetimes are finite and I want us to spend as much time as possible playing the best things we possibly can.

      >Also, the interactive fiction movement has demonstrated for years that you can do a lot more with story and interaction than Choose Your Own Adventure. Are you aware of IF at all?

      Totally aware of IF, and I am not at all convinced that it’s fundamentally more than CYOA. I think it’s just a more complex CYOA, but it still comes down to being a trial and error advent calendar.

      • http://outsideyourheaven.blogspot.com/ Matthew “Sajon” Weise

        “The world is big enough for whatever. I’m not saying people can’t make these things, I’m saying they shouldn’t. Because sure, the world is big, but our lifetimes are finite and I want us to spend as much time as possible playing the best things we possibly can.”

        Your best game might not be someone else’s.

        • http://www.dinofarmgames.com keithburgun

          Of course. We’ll never reach a place where everyone agrees on what’s good. However, we *can* establish principles, guidelines — a “music theory” of game design, and that’s what I’m after.

  • Pingback: Superlicious | Superlevel

  • Hypocee

    Heavy modern score involvement:
    Spelunky – http://www.rockpapershotgun.com/2009/03/30/snake-to-death-the-majesty-of-spelunky/
    Grid Wars 2 – http://wosblog.podgamer.com/2011/08/27/this-frequencys-my-universe/

    Not? coincidentally, two of the greatest games in history, and two of only four or five I’ve discovered that feel like they actually improve their players’ quality of life outside play. Spelunky has not one but two tiers of ending, which are typically reached by sufficiently skilled players in under 15 minutes.

  • http://www.smestorp.com/ Michael Brough

    This is a good article, I agree with most of it. Especially the “big number” creep.

    > The real goal is getting a high score, but I’ve had arguments with many who say that they don’t care, that they really care about getting to a higher dungeon level.

    This is why in Zaga-33 I made the score simply be what dungeon level you got to (plus a bonus if you complete it from how many items you have). Also keeps the numbers nice and small.

    Score and endpoints mix best in short games. (I’ll argue sometime that short games are better in general, but that’s a different post.) Every Extend was a bounded game I spent a long time on trying to improve my score; you can get to the end and kill the boss in a few minutes, but the real point of the game is trying to give the best performance you can in those few minutes.
    In fact, if there isn’t a bound on game length, then “survive as long/far as possible” will always become an implicit goal, an alternate score system (unless this is precisely what the actual score measures).

    • http://www.dinofarmgames.com keithburgun

      Hey, it’s the Zaga-33 guy! Good to hear from you. I really enjoyed your game.

      I think your score system is totally coherent, and clear! So I give it a lot of props for that.

      Actually though, the thing that bothers me, and it’s also the thing that got me to stop playing Zaga-33, is that your score system is actually TOO tight. Like, the problem is not that the numbers are small, it’s that they are inelastic.

      I think it’s important for a single-player game score system to have some elasticity. I think videogames generally go too far with it and have these outrageous multipliers, but I really want a way to have the score express something about my really awesome play.

      I should have mentioned Zaga-33, really, in that it’s a very interesting case. I think that unlike most videogames it is a coherent scoring system, but I personally think it goes a little bit too far and it would be better to move up a small amount in terms of complexity from where it is.

      • http://www.smestorp.com/ Michael Brough

        Thanks!

        Sure, I definitely wouldn’t want every game to have the kind of scoring Zaga-33 has, but I think how it is is right for the game. It is very coarse, but that’s kind of the point – it kind of says “I don’t care how you approach each level, just get through it”. The design’s too focused to have much concept of a “really awesome play” other than this – if it gets you ahead it’s good, if it gets you there with more resources left over it’s better; not because the resources are counted, but because you can use them to keep going further.

        It also smooths over the randomness a bit – e.g. if you get points for kills, then if you happened to draw good offensive equipment you’d score more than if you got stealthy or defensive equipment. This doesn’t reflect at all on how well you played! A more deterministic game can obviously get away with a more fine-grained score.

        • http://www.dinofarmgames.com keithburgun

          Yeah. You are a good game designer and I really appreciate Zaga-33. Have you made any other games? Link me to your site or somethin’!

  • Pingback: The Monday Papers | Rock, Paper, Shotgun

  • k

    I don’t think having an end goal is necessarily detrimental to a score system.

    Dungeon Crawl has a pretty good scoring system, other than the numbers being too big to be meaningful in and of themselves.

    For a winning game the main scoring factors are
    a) runes collected;
    b) turns taken.

    Runes being the main factor, with world class high scores requiring all 15, though a quick 14 rune game will beat a slow 15 rune one.

    The only confusing part is that a small amount of score is awarded for exp and gold gained. (Which is there so that new players can measure their progress, your 33,000 is effectively 0 in the “real” scoring system.)

    Maybe it would be better if, instead of a single incomprehensibly large number, it just sorted games by 5 parameters in descending order of importance:
    Winning, rune count, turn count, exp, gold.

    And it makes very little sense to generalize the dungeon level display argument to the whole genre of roguelikes, when in a game with a more complex dungeon layout (Crawl) it is actually gameplay-relevant information.

    • k

      With turn count only being taken into account for winning games, obviously.

    • http://www.dinofarmgames.com keithburgun

      The problem with “completion” in a score based game is that it’s confusing for players. They don’t know if the goal is to get to the end, or if it’s to get a high score (in fact they almost CERTAINLY think it’s the former, having been trained by modern storybased videogames).

      For instance, Mystery Dungeon: Shiren the Wanderer, probably the *BEST* roguelike ever created, got kind of bad reviews because it was “too hard”. Of course, it was only “too hard” if you are expecting to “complete” it, and that that’s the only way to win. When you realize that it’s actually score based, and you win roughly half the time, it’s suddenly not too hard anymore.

      • k

        But is such a reception really the game’s problem, when it is not a flaw in the game mechanics? Though I suppose it matters more for a commercial game.

        • http://www.dinofarmgames.com keithburgun

          It is the game’s problem, I think. The game doesn’t state what its goal is clearly. I think people are actually somewhat reasonable in expecting that it’s about completion, being that it has a total JRPG presentation.

  • Fhnuzoag

    Hypocee:

    People play Spelunky to try to pass all 15 levels. A few more play Spelunky to try to pass all 15 levels in the minimum possible time. Only a tiny number of people play to maximise score – because maximising score is often exceedingly tedious, being dependent on the player taking as much time as he can to comb every part of every level for money. Which involves little skill, and so has little interest.

    Really, this points to the problem with score mechanics. Achieving a higher score – in contrast to diving to a lower level, does not usually necessitate a corresponding improvement in skill level. More often than not, a high score is tied to an abundance of patience than anything else. This is the root cause behind people wanting to measure their skill level by number of levels – because lots of levels means they saw lots of new and more dangerous enemies, which points to their greater mastery of the game. Lots of score just means they had enough time to farm easy levels for points. In games with experience points and level up that reward score for killing enemies, the greatest skill level would be those who can dive to deep levels, and *minimise* their score.

    Removing display of the dungeon level misses the point entirely, and telling player what they should want is just a bad attitude.

    • http://www.dinofarmgames.com keithburgun

      >Only a tiny number of people play to maximise score – because maximising score is often exceedingly tedious, being dependent on the player taking as much time as he can to comb every part of every level for money. Which involves little skill, and so has little interest.

      If the game isn’t about score, it shouldn’t have a score system. That goes without saying.

      I’ve always thought Spelunky is actually a really haphazard, poorly-designed game, and your claim about its scoring system backs that up for me.

      >Achieving a higher score – in contrast to diving to a lower level, does not usually necessitate a corresponding improvement in skill level.

      Hey now, don’t go from “In Spelunky it’s true” to “that means its true everywhere”. In Tetris, getting consistent Tetrises is more valuable than getting to higher levels. In fact, you want to get as many points as you can without getting to higher levels, because doing so is a threat to you.

      Telling the player what they should want is a bad attitude? It’s called game design, dude. The game designer chooses the goal of a game.

  • Fhnuzoag

    >If the game isn’t about score, it shouldn’t have a score system. That goes without saying.
    >I’ve always thought Spelunky is actually a really haphazard, poorly-designed game, and your claim about its scoring system backs that up for me.

    Hohoho. Them’s fighting words. I will be interested in seeing how good your game is, then. Right now, I’ve replayed Spelunky 526 times. This is your score to beat.

    To be fair in Spelunky there is a mechanic whereupon the highest scores still require skill: the maximal money (because that’s what score is in Spelunky) is attained by not purchasing anything from the game’s stores, and by the exceedingly dangerous trick of ‘ghostrunning’. Playing in this mode is unattainable to most players, though.

    >Hey now, don’t go from “In Spelunky it’s true” to “that means its true everywhere”. In Tetris, getting consistent Tetrises is more valuable than getting to higher levels. In fact, you want to get as many points as you can without getting to higher levels, because doing so is a threat to you.

    But then Tetris lets you pick *the initial speed you start at*! If the game was solely about score, choosing any speed other than the slowest speed would be non-optimal. Hence, the higher scores go to those players with the patience enough to go through multiple minutes at low speeds, rather than those looking for a challenge and going in at speed level 5 or whatever.

    >Telling the player what they should want is a bad attitude? It’s called game design, dude. The game designer chooses the goal of a game.

    The game designer chooses the *intended goal* of the game, perhaps, or more generally provides a range of ways to play a game, but it’s not the role of the game designer to restrict the methods in which the player might enjoy it, or might discover goals of their own. You can’t force the players to enjoy one aspect of the game by removing aspects that they prefer. If the players are uninterested in your score, then you’d better find a way to interest them in it, or rethink fundamentally your design.

    You are basically falling into the exact same trap as the Call of Duty games by providing a score-lined breadcrumb trail for the players to follow, and seeing players being uninterested in your trail as a problem to be solved by removing reasons to deviate from it (or making such deviation impossible). It’s the same sort of attitude.

    • http://www.dinofarmgames.com keithburgun

      My game doesn’t have to be good for me to say that Spelunky is a poorly designed game. I don’t need to have made a game at all to make this claim. What you’re basically asking for is an argument from authority – a logical fallacy – instead of asking for me to clarify my reasons. Even if I had designed the best games ever made, I could be totally wrong, and if I had designed no games at all, I could be totally right.

      >But then Tetris lets you pick *the initial speed you start at*

      Yes, and it is ALWAYS best to start at level 9; it’s totally optimal, which I think is kind of bad design. They shouldn’t have let you choose what speed to start at.

      Your thing about…

      >You can’t force the players to enjoy one aspect of the game by removing aspects that they prefer.

      …Makes no sense. How can people “prefer” X over Y in a game that is truly new (which Auro is) and that they have never played before? Totally ridiculous. And I’m not “forcing people to enjoy X”, I am saying “X is the goal of the game. If Y is NOT the goal of the game, if it is not relevant, then it shouldn’t be shown”.

      To illustrate, hey, maybe some players would like to see “total tiles walked” at the end of the game! Of course, this has zero to do with the gameplay, but hey, “maybe some would enjoy it”.

      I am not a “maybe some would enjoy all kinds of things, so throw all kinds of things in” kind of designer. In fact, I don’t think that that even qualifies as design. To me, design is careful, intentional construction of a coherent and focused system.

      To clarify how Auro is nothing like Call of Duty, Auro is all about one thing: interesting, meaningful and ambiguous decision-making. You use skills, tactical skills against monsters who also have their own tactical skills. The level layouts are random and they are not puzzles (as CoD single player almost certainly is, an execution/memorization puzzle). Execution and memorization helps you zero in Auro. What helps you is gaining skill at the game, and the game is about making ambiguous decision-making. The system will encourage expression and creativity in play, through combining skills in synergy with each other.

      • Fhnuzoag

        >My game doesn’t have to be good for me to say that Spelunky is a poorly designed game. I don’t need to have made a game at all to make this claim. What you’re basically asking for is an argument from authority – a logical fallacy – instead of asking for me to clarify my reasons. Even if I had designed the best games ever made, I could be totally wrong, and if I had designed no games at all, I could be totally right.

        I’m not making an argument from authority. I’m saying that your idea of a good game design, and of mechanics that encourage replay specifically, do not seem to relate to past experience in terms of what people (and I, in particular) enjoy and replay. If I was saying ‘But Derek Yu says different!’ then I’d be arguing from authority. Here, I’m arguing from what has worked in the past, from evidence. You had best present your own evidence to counter this.

        >…Makes no sense. How can people “prefer” X over Y in a game that is truly new (which Auro is) and that they have never played before? Totally ridiculous.

        The purpose of this discussion is to have people play Auro more than once, right? So after the first playthrough, they will have played your game before, and they may well be sick and tired of score maximisation. The funny thing about being a game designer is that in reality, you might quickly find that your players have played your game more than you have, and are better at it than your imagine. So the prescriptive attitude of ‘I’m the designer, I know better!’ quickly becomes silly.

        >To illustrate, hey, maybe some players would like to see “total tiles walked” at the end of the game! Of course, this has zero to do with the gameplay, but hey, “maybe some would enjoy it”.
        >I am not a “maybe some would enjoy all kinds of things, so throw all kinds of things in” kind of designer. In fact, I don’t think that that even qualifies as design. To me, design is careful, intentional construction of a coherent and focused system.

        Including different measures of success can be done haphazardly, for sure. But it can also be part of a focused design – just one that recognises that different players might have different values to which they attach notions of success, and that varied victory criteria encourage players to play differently and find different aspects to the game.

        Indeed, I’d argue that score is *fundamentally broken* in this respect of producing creativity. No matter how you weight things, the fact that there is a single summary statistics means, mathematically, there is a single method of maximising this value. You might allow both magic attacks and melee attacks, but once people figure out magic attacks have a 1% score bonus over melee attacks, for score maximisers melee attacks basically might as well not exist. Once the optimal strategy is determined – and soon it will, and the proliferation of GameFAQs and so one means soon everyone will know it, all increases in score arise from further perfection in execution of this strategy and from random luck. And then your game is dead.

        • http://www.dinofarmgames.com keithburgun

          >I’m saying that your idea of a good game design, and of mechanics that encourage replay specifically, do not seem to relate to past experience in terms of what people (and I, in particular) enjoy and replay.

          Point out specifically what about my idea doesn’t work.

          >So after the first playthrough, they will have played your game before, and they may well be sick and tired of score maximisation.

          After ONE GAME they’ll be tired of trying to get a high score? Wouldn’t that also be the case for Soccer, and Go?

          >>The funny thing about being a game designer is that in reality, you might quickly find that your players have played your game more than you have, and are better at it than your imagine. So the prescriptive attitude of ‘I’m the designer, I know better!’ quickly becomes silly.

          Being good AT a game is a VERY different skill than being good at game design, I’m sure you made a silly mistake when you typed this.

          If score is fundamentally broken in respect to producing creativity, then I guess Go and Soccer are fundamentally broken and don’t produce creative play.

          • FhnuZoag

            >Point out specifically what about my idea doesn’t work.

            Until your game is released, that’s going to be hard isn’t it? But as it is, I can’t find myself very interested in it. I address more your argument that Spelunky is a poor design. But people enjoy it! So is your discussion of what is or is not a poor design actually tied to what people do or do not enjoy, or is it purely an abstract statement with no relation to how people actually play games?

            As to why score systems don’t work, well, I’m talking about it, aren’t I?

            >After ONE GAME they’ll be tired of trying to get a high score? Wouldn’t that also be the case for Soccer, and Go?

            Yes? Now a small subset of people don’t get tired of it, but its inescapable that Soccer and Go are kinda *unpopular* as games people actually play and keep playing competitively. Indeed, probably most play of those games are in spite of, instead of because of the scoring system. Playing football and Go allows people to socialise, to exercise their mind or their body, to work out aggression (and because their teachers force them to) and so on, and that’s why people play on even though they have no expectation of consistently improving their performance. Individual players keep track of their statistics like ball possession, number of assists, and so on, again when they aren’t factors of the final match score. For beginner players, who have no expectation of winning, it’s these factors that keep them playing, and well, if you are okay with simply ignoring them, then I think you are making a big mistake.

            Of course, these games are kinda a false comparison because Go and Football are multiplayer games. Being multiplayer increases the strategic space, because many multiplayer games have a rock-paper-scissors effect whereupon a strategy is good only versus certain other ones. This reduces the likelihood of there being a single dominant way to play, and this is what produces the creativity in those games – arguing that Go because it has a scoring system leads to more creativity than Chess which has none is silly. Finding creative single player score maximisation games would be a lot more difficult. And that’s ignoring the fact that the vast majority of low level Go and Football players play games entirely uncreatively.

            My argument is that score doesn’t produce creativity – and often discourages it. I do not claim to argue that a score system makes creativity impossible. But merely consider this: on a replay, or a new phase of the game, the player has essentially two options. The player may continue with what he did previously, and attempt to perfect it. Or he may try a radically new and unfamiliar strategy. With the former, he will usually incrementally improve his high score. With the latter, because he’s unfamiliar with the new strategy, he will almost certainly in the short term perform worse than his original strategy. Which means that if he is, as intended, playing to improve on his high score, then he is directly penalised for trying something new. More dangerously, it might not be obvious to him that with more work, his new strategy will perform better than the old one. So he returns to his old strategy. This is the mechanism by which score maximisation discourages creativity, and encourages people to specialise until they get bored with your game.

            Using a completion mechanic flattens the cost of the new strategy. Okay, you are performing for the time being worse than before. But hey, you still got through the chapter! Maybe you can get through the next? So you keep at the new strategy, at least long enough to evaluate it properly.

            Offering different success metrics also removes the cost of the new strategy. Oh, you have a lower score X now. But hey, check it out, your score Y is now higher! Maybe you can figure out a way to have score X and score Y both higher, or maybe you decide that you like the game style encouraged by score Y more? Again, you are encouraged to explore.

            • http://www.dinofarmgames.com keithburgun

              People enjoy Spelunky more because it does a few things right (i.e. random elements, which are REQUIRED for a single player game to be replayable at all) in a world where almost all videogames do almost everything completely wrong.

              Good design isn’t *directly* related to “what people enjoy”, no. There are plenty of horribly designed things that people enjoy.

              >But merely consider this: on a replay, or a new phase of the game, the player has essentially two options. The player may continue with what he did previously, and attempt to perfect it.

              For a non-random single player game. This is why all single player games need random elements, so that the player’s skill is being tested, not his memorization.

              >>inescapable that Soccer and Go are kinda *unpopular* as games people actually play and keep playing competitively.

              Are you *insane*? What world are you living in? Look, if my games become as “unpopular” as Soccer and Go, I think I’ll be pretty happy with that.

              >no expectation of consistently improving their performance

              I seriously think you must be trolling at this point. Soccer and Go players aren’t trying to improve their performance?

              You are either trolling or your brain is hopelessly twisted by the industry into an illogical shape.

            • Hypocee

              OK, so this guy’s that kind. Never mind that, then. I do mildly disagree with your statement about the theoretical mathematical optimum ruining score games, and I suspect and hope you might not really hold to it since it’s a depressing way to feel. The situation you describe is certainly the norm, since few are lucky enough to implement a classic game design. I mentally map the phenomenon to danmaku shooters more than any other genre, and Ikaruga in particular. I loved Ikaruga when I was just playing for survival and efficiency, but my opinion of it took a big hit when I discovered/noticed the 3-chain combo scoring system and the patterns where you’re supposed to plink here, weave here, etc. I could see that there was a singular perfect path constructed through the game, that it had been carefully built within human capability so one was always playing either perfectly or incorrectly, and that I had no interest whatsoever in pursuing that brittle system. It turned the score readout into a constant reminder of the part of the game that was antithetical to me.

              The escape from the watershed trap is that most excellent score games have A. multiple optima which B. exceed human possibility C. along a reasonably smooth gradient. When those criteria are met, it ceases to matter which one is theoretically “highest” above the ceiling – the game becomes about exploring the fractal spaces bordering the impossible, and allows self-expression both in choosing at any moment which optimum best suits your mental state at a given day or moment, and in many cases which best suits the environment the RNG has thrown up and which of those two paths is more important to you. High-scoring players still post fundamental disagreements about, say, the pitcher’s mitt in Spelunky; some claim it’s useless, others that it’s a good value. Stuart Campbell claims in his piece that black hole farming is essential to high scores in Grid Wars 2. Happily he’s wrong; my twitch-god little brother reached similar scores with a policy of destroying all black holes ASAP. In both cases, the design is expansive and robust enough to accommodate the development of personal styles.

              Of course, SpaceChem is the epitome of your proposed approach – it rates various criteria of solutions on a curve against all players, and it is literally impossible to perform well on many at once. The optimizing player is forced to choose an optimum to pursue at any time, and always has the other choices in front of them encouraging exploration.

  • Burzmali

    The problem I have was scoring systems is that the skills required to maximize score are largely independent of the skills needed to progress in a game. Take Tetris for instance, if you want to maximize your score, you need to drop piece from as high up as possible, however survival is more related to your ability to carefully position pieces. It’s similar to speed-runs where the bizarre risks and highly precise movements are more important than tactics.

    For a game like Auro, maybe you have a bonus for killing multiple enemies with the same attack or a combo multiplier for killing enemies in sequential attacks. Either way, you are rewarding the player for taking risk that increase score without aiding in survival, which is pretty much the bog standard mechanism of the puzzle game genre. If you just reward killing monsters and finding loot, you are doing the opposite and rewarding the player for being extra cautious, which is effectively ensuring that the player will grind. Rewarding “excellent play” in a procedural game is usually synonymous with reward excellent random placement of obstacles.

    • http://www.dinofarmgames.com keithburgun

      “The problem you have” is like a fundamental quality of games. It’s like saying “what’s annoying about boxing is that you have to attack, but you also have to defend”. Basically every single decent game in existence has this two-axis “balancing act” that has to be maintained.

      In Tetris, the balancing act is between “survival” and “doing things to push for more points”, such as the hard drop and saving up for tetrises.

      This two-axis balancing act is what makes games interesting! If it weren’t for that, then they would probably be contests. Like a pie-eating contest – in that, all you have to do is EAT MORE PIES THAN THE OTHER GUY AND NOTHING ELSE.

      >Rewarding “excellent play” in a procedural game is usually synonymous with reward excellent random placement of obstacles.

      What?

  • Burzmali

    >“The problem you have” is like a fundamental quality of games. It’s like saying “what’s annoying about boxing is that you have to attack, but you also have to defend”. Basically every single decent game in existence has this two-axis “balancing act” that has to be maintained.

    Many games have you balancing survival and progress, by granting score equal status to those, you add another axis to fall off of. If you can manage it, cool. Typically, the reward for survival is continuing to play, the rewards for progress are new skills that let you continue to survive, but the reward for score is usually just bragging rights, unless you plan to add scoring gates.

    >>Rewarding “excellent play” in a procedural game is usually synonymous with reward excellent random placement of obstacles.
    >What?
    If you get points for completing a level quickly, location and type of monsters is at least as important as skill for determining how quickly a player will reach an exit. Even more so for reaching an exit with the least damage. Most metrics that give points in non-random games tend to be heavily influenced by the RNG in a procedural one. See wizards in Nethack.

    • http://www.dinofarmgames.com keithburgun

      In no world other than the world of videogames is it inherently considered always a “good thing” that the game is going on longer. It’s obvious to everyone BUT video-gamers that “game end condition” isn’t some inherently bad thing, but we’ve conflated it with death, completely.

      >If you get points for completing a level quickly, location and type of monsters is at least as important as skill for determining how quickly a player will reach an exit.

      If that’s true (I don’t think it is) then it’s also true for Tetris. If my game is as bad as Tetris I think I can survive.

      It’s not just quickly, by the way. It’s moving quickly and killing monsters in combos that both give points.

  • Bucky

    I think by your standards, completion time itself is a good scoring metric. It’s pretty obvious what “0 hours, 32 minutes” means on the game over screen, the goal of “do it fast” is easy to communicate, it is clearly related to skill, and it even sidesteps the “completion point” issue.

    • http://www.dinofarmgames.com keithburgun

      Right, well we’re sort of doing something like that. Of course, we’re not using “seconds” (as our game is turn based and therefore uses discrete time(turns), not continuous time). We also want to reward killing monsters efficiently. The game is about efficiency, in general.

  • rubybliels

    “For AURO, our solution is to not even show what dungeon level number you’re at.”
    Does the level generation change for different level numbers (ie. certain enemies show up on floors 4-8)? If so there would be good reason for players to know the level number. And while showing a bunch of different stats can be confusing, I don’t think it is always bad. Look at baseball and sabermetrics for example. The goal of baseball is still to score more runs than your opponent. But stats can illuminate “how” to win the game.

    • http://www.dinofarmgames.com keithburgun

      >Does the level generation change for different level numbers (ie. certain enemies show up on floors 4-8)?

      No, not really. I mean, loosely. There are some monsters who will only appear after level 4, for instance, or before level 3. But it’s not like 100 Rogues where “on level 3 you fight Black Bandits” and such. Almost every monster can appear at almost every level.

      I think people are obsessed with numbers in Baseball because the game part of Baseball is… not that interesting. To quote the Simpsons “it has a level of excitement and passion that is only exceeded by every other sport”.

  • http://roguetemple.com/z/ zenorogue

    Good luck with creating a game with score and no “game completion”, then.

    My HyperRogue II had everything available from the beginning, and no game completion. I have seen players complain about this: “Shame it has no real end other than dying (or does it?)”, “I feel I have explored the game in its entirety” (the second comment actually made by a player who has not yet seen some of the cool areas). I agree with the feeling in the first comment, it is much more attractive for the players to aim for something like “wow, you have seen mostly everything, and play quite well… you are a winner then, congratulations!” than “your score: 138″.

    Thus HyperRogue III changes this (there is a clear quest that can be completed only by skilled players who have seen almost everything, and some areas are available only after proving yourself in earlier ones). One thing that this post has made me realize that I have designed incorrectly is that I gave too much points for completing the quest (which means that you do need to care about the main quest when going for a high score, which is bad IMO).

    • http://www.dinofarmgames.com keithburgun

      I mean, if these are people who would complain that Tetris has no “completion point” then I am comfortable with that.

      I understand we are in a culture of “seeing all the content”, but I think this is really terrible and I want to do what I can to combat that. Auro is most certainly *not* a game about “seein’ all da stuff”.

      Actually, we will have win/loss conditions very tightly tied to high score.

  • Pingback: URL

  • jrodman

    In roguelikes, the real goal is indeed to get farther, and in fact complete the game. That’s the real goal.

    The score goal is the fake goal. Very few players are interested in this.

    Some develop alternative scores, such as completing the game in fewer turns, and compete on this basis. But claiming that the score is the real goal is just parading your ignorance of real player motivation. Again.

    • http://www.dinofarmgames.com keithburgun

      >In roguelikes, the real goal is indeed to get farther, and in fact complete the game.

      If that’s the case, then I think they are not well-balanced games, and they are generally not well designed. The player is basically guaranteed to lose his or her first 50-100 games (actually, I’ve played hundreds of DCSS games and same with Shiren the Wanderer; I’ve still never “won” by that definition). Then, when he DOES win, he’s increasingly guaranteed to win from that point out.

      Even if you think the current “real goal” of roguelikes is “completion”, a dynamic “can you beat your last high score” goal is much better.

      >”the fake goal”

      Games have fake goals? What is the purpose of a fake goal? To confuse the player?

      >Very few players are interested in this.

      This is true in videogames because of bad implementation and unclear / conflicting goals. Look to sports and boardgames to see that people do indeed actually care about points when it’s implemented correctly.

  • jrodman

    Increasing chances to win in all games is typical, and interacting with avid players of roguelikes over long periods, many are satisfied to keep playing without ever expecting to win. I think that’s really part of the draw. It’s out there somewhere and eventually you’ll find it.

    Players will just find harder goals if the stated goals become too easy, in the context of a complex game like a roguelike. It’s not really that big a problem. Meanwhile, not every game needs to be ‘a lifetime to master’. There are way too many games for that to even be a worthwhile goal for most.

    You claim one goal is real, which means the other goals are not. Fake goal is just a logical consequence.

    Most players of sports and boardgames are not that interested in the points either. Many people play sports and boardgames socially. This is *much* more common than playing competitively in sports beyond age 25 or so. Boardgames it varies by group. Dyed in the wool gamers are much more likely to be very interested in the score, while people who play now and then are much less likely to be. In board games, typically nearly everyone accepts that the score is part of the game, and by doing so avoids wrecking the game (to do otherwise in most multiplayer games would be very unsocial), but don’t really care too much what the result is.

    Altogether too much projection.

    • http://www.dinofarmgames.com keithburgun

      >Players will just find harder goals

      Right, but this is sort of the player designing the game. Which isn’t a problem in and of itself, but it is a mark of a less-well designed game if people have to house-rule it and such.

      >Many people play sports and boardgames socially.

      Even when they do play socially, they are still totally interested in points, since they are the object of the game.

  • Pingback: Adventures of a Newbie Roguelike Dev » Don’t Ask What, Ask Why