If Numbers Are a Metric, I Give Zero Fucks

A recent thread in the Smite section of Reddit on the topic of player skill rating visibility prompted me to ponder the usefulness of being able to see the hidden personal performance numbers that HiRez, the studio responsible for Smite, associates with players in casual and ranked play. This led me to a broader consideration of the numbers associated with personal achievement in different game types.

Members of the Smite community have taken to referring to such officially invisible numbers as Elo, short for “Elo rating,” rather than MMR (matchmaking rating) or other alternatives. Typically, “Elo” by itself refers to a rating system for calculating the relative skill levels of players in competitive games, most notably chess. The system is named after its creator, American physics professor Arpad Elo. Smite uses the Elo system to track player performance, determine division placement in ranked play, and select suitable co-players during the matchmaking process.

Different games espouse different ideals and cultivate particular cultures when it comes to the quantification of personal achievement and ability. There is a tendency in competitive communities, for example, to channel one’s desire to win and perform at the highest levels into toxicity when other players are observed to have, in their estimation, underperformed. It may also occur when other players are not playing in the way that the player in question would have. Any material that can be used to further one’s argument or vitriol is sought after and wielded.

One of these items is a player’s personal numeric skill rating. Some competitive games make this information publicly available through official means. Other games, such as Smite, do not, and players are left to rely on third-party calculations in order to obtain this information. We see in other games that there is, perhaps, a certain art to being selective about what information to offer to players and what information to withhold. Enemy hit points are not displayed numerically in Guild Wars 2, for example, but via the length of a red bar. Classic EverQuest did not tell you a hostile creature’s level, only how deadly it appeared to be. Similarly, HiRez does not make official Elo ratings available to players. They must go to a site like SmiteGuru in order to view approximations thereof – that is, if the player has not chosen to hide their profile. Most players do not.

Smite Profile Screen.png
Normally these numbers are hidden from the world.

I am an exception. Like HiRez, I hide certain aspects of my information. I keep my profile hidden which means that other players cannot look up my information. When I am listed in official matches, only the god I played is listed – my account name is not. Additionally, hidden profiles do not feed data to SmiteGuru. This means that not only can other players not see my account-level kill-death-assist and win-loss statistics, they also cannot see my approximated skill rating. By the same token, neither can I. It may sound rather odd when you think about it. Shouldn’t I want to know how I’m doing? Why would I hide this information, especially from myself?

In the case of HiRez, there is a culture that is cultivated by not making certain information public. Hiding a player’s Elo rating implicitly tells the community that HiRez does not encourage the use of numbers as an exclusive means of determining the worth of other players. More specifically, it does not provide players material to be used for bad-mouthing / bad manners (“BM”) or ranking other players internally based on numbers. What is implicitly encouraged – that is to say, not structurally discouraged by official, public ratings information – is determining how good you think a player is by watching them play.

This is exactly what I do when evaluating my own performance in games. My metrics are generally based on subjective observations. Did we win? Did anyone die? Was I out of position? Did I utilize my abilities effectively? I rarely use healing or damage meters when they are available. My flirtations with them are always very brief. It doesn’t particularly matter to me exactly how much damage any of my abilities are doing or what my Heals Per Second output is. I’m a Feeler as opposed to a Thinker according to the Myers-Briggs Type Indicator. My preferred mode of interacting with game worlds is intuitive in nature rather than analytical. It’s not that I eschew numbers – I simply don’t worship them.

ScreenShot00003 (2)
These gods, on the other hand, have 50+ worshipers and are therefore mastered which makes them available for ranked play. It doesn’t mean I’ve actually “mastered” them, though.

Psychologically speaking, I think everyone likes to have a convenient metric that tells them roughly where they’re at. This is fine and dandy. We like to know what we’ve accomplished. How frustrating it must have been for numbers-driven players of classic EverQuest to approximate their experience gains with orange and blue bubbles. Numbers give us a tangible, definite quantity that represents, somehow, our investment, our dedication, our accomplishments, our time and effort. It’s reassuring, on some level, to have that definite, immediately recognizable reference point. It may even drive us to continue, as there are some of us who are motivated by such things. (I won’t stop playing tonight until I’ve achieved item level 170 on my Paladin. Ahem.)

Numbers tell us what we’ve done. They quantify our deeds, our ambitions, the totality of our play sessions. We have UI screens filled with numbers that tell us how powerful and experienced we are as a function of the adventures we’ve undertaken. There is a certain fascination to poring over the myriad statistics associated with our characters and our play sessions. We may wish to quantify them using tools such as spreadsheets, combat parsers, and simulations. This can be a game in and of itself. Numbers are great fun if that’s your sort of thing.

Meters and numbers are indeed quite useful as pointers: if you have mana issues and your meters tell you that you’re overhealing quite a bit on a class that does not feature overhealing as part of their ability set, it may lead you to investigate your ability usage patterns in order to solve the problem. The act of overhealing, however, does not make you a bad player – if other healers were to send you private messages telling you that you should delete your character and uninstall because you’re an overhealing noobmonster, how would that make you feel?

You’re not a bad player because the numbers say you’re suboptimal. And the numbers aren’t bad because they tell you things you may not like. Neither the player nor the meters are unwelcome. Why? Because the numbers aren’t even saying you’re suboptimal. Certain people looking at the numbers are saying you’re suboptimal. (It might even be you saying it!)

And so as useful as numbers are in describing damage done, achievements unlocked, and power-ups acquired, we must keep in perspective that they are only one part of the comprehensive picture that we must consider when determining our own progress in and contributions to any given game. That is, if we aspire to be something more than barbaric when it comes to sizing up our fellow players, teammates and opponents alike.

This, in a nutshell, is why I refuse to look at my own rating numbers or the numbers of others. This is a personal choice, not a philosophical approach that I demand everyone adhere to. When it comes to evaluating my personal progress and performance, I’d rather have fellow players use the same metrics that I would use for them. I can’t force them to, of course. I can’t even say that this is what they should do. They can do whatever the hell they want. However, I’m of the opinion that one’s Elo rating does not present the entire picture when it comes to a player’s effectiveness and contributions, nor do other associated numbers which may be pinned to mastery.

Smite Win Loss History.png
A somewhat more useful description of player performance.

Let’s consider an example from Smite. For those who are not familiar with the game, Smite is a MOBA (Multiplayer Online Battle Arena) in which players control gods from different historical pantheons and fight each other over strategic objectives. There’s nothing that tells you definitively whether a player is good at playing the god they’ve chosen. They may have a god mastery rank on the player card which is displayed to everyone prior to the beginning of the match. For example, I have Artemis at Rank I which means a small “I” will appear on my pre-match player card if I have chosen not to hide that information. What does this signify? It reflects the fact that I’ve played enough matches to acquire 50 “worshippers” on Artemis. In unranked play, players gain 0.2 worshippers per minute and an additional 4 worshippers for a win. A long win (30 minutes) will therefore grant 10 worshippers, while a short loss (15 minutes) will grant 3 worshippers. The “I” thus describes how much I’ve played that god without really telling anyone how many matches I’ve won. Other players are free to interpret this as they see fit, just as they may size up the “golden” skin which can be purchased for that god at mastery rank I using in-game currency.

At mastery rank X, which requires 1,000 worshippers, players may purchase a “diamond” skin for that god using in-game currency. Does this mean that an “X” indicates a good player? Well, there’s a correlation, perhaps one that we could only describe as “not strong” at best. But what if, for example, you were to encounter a diamond Geb, the Egyptian god of earth, who was using abilities seemingly randomly and not contributing to team fights? Well, perhaps their 1,000 worshippers and many, many hours of play wouldn’t be terribly useful as a barometer of skill. It could be that the person who put in those hours is part of a family that shares the account (so naughty!) or the account could belong to a player who is letting a friend try the game for the first time.

So it might not be that you’re using the numbers as a basis for making assumptions or statements that aren’t necessarily accurate. It could be that the numbers are lying to you because you don’t have the entire context available.

Let me put it this way: if the thermometer says it’s 20 degrees Fahrenheit outside, it’s probably pretty brisk, right? We know that objectively. But if someone who’s had a bottle and a half of whisky goes out in that weather, it’s 60 degrees Fahrenheit to them. We have to take into account the person’s inner circumstances in addition to any objective description of their environment that we may have at hand.

Smite Login Screen
Skadi, the Norse goddess of winter and Smite’s newest addition, is also unaffected by the cold.

In many cases, we simply don’t have access to this information, so our picture is incomplete. We can, of course, derive the most reasonable and plausible conclusions possible based on the existing information. We must know, however, that these hypotheses will be imperfect – something that every scientist worth their salt uses as the foundation of their endeavors. If the player is a streamer and we can see what they are doing in real time – both inside the game and outside the game (do they have a cold? are they tired?) – our available base of information is more comprehensive and we can come to a more informed, less imperfect conclusion.

So why in the world would we think that the use of a single metric or even a single source of information accurately represents the worth, value, and contributions of a player?

Because it’s convenient. It lets us be lazy. We can pin our losses, deaths, and situations with less than perfect outcomes on others. And we’re assholes, too, so why not?

Well, except those of us that aren’t assholes.

My level of experience when playing in matches against other players shows through in my ability to battle other players, make a difference in team fights, and in whether we win the game or not. I have no need to reference a mathemagical number called “Elo” to tell me whether I’m valuable or not. And I certainly don’t need to make it available to others so that they can use it as a marker of judgement with regard to my relative worth as a player.

Take your favorite MMO. Your level doesn’t tell people how many sights you’ve seen, air balloons you’ve ridden, horses you’ve tamed, or warehouses you’ve constructed. It simply doesn’t reflect all of those things. Certainly there are numbers that do reflect those things, and those numbers must be incorporated into a comprehensive overview of a person-player if we want to understand their contributions to the world they inhabit, even if those contributions are mostly solo endeavors.

It’s just that I have little use for those numbers. Yesterday evening, I played three matches in which I underperformed. My team still won every time. Does my Elo rating reflect how I did in those matches? Nope. It reflects the superiority of my team’s mechanical ability to play their chosen gods and score kills. My part in this whole shebang was minimal at best. I felt bad about it because I’ve played much better. Perhaps I was just having an off night?

And there it is. You wouldn’t know unless you talked to me or saw inside me. You can analyze replays of my matches and point out weaknesses and flaws and from this gain insight into my level of ability and my contributions to our wins. My Elo rating tells you about my win-loss history and how I’m expected to perform. From this, you can’t reasonably extrapolate anything about my character, personality, play style, or any of that. What you have is a number whose usefulness does not extend to all domains.

Which is why I agree with HiRez’s philosophical stance on this matter. Not making those scores available is a clear signal to players that they will not provide ammunition for toxic behavior. Those who are determined to be asshats to their fellow players will, of course, make use of whatever they can find in doing so. The fact that they will do so whether Elo ratings are publicly available or not does not mean that HiRez should go ahead and make them visible. It’s about the principle of the thing. By cultivating a set of principles – an ethos – via what they do and do not do, they send a message to the community of people who play their game: evaluate others as players, not numbers.


9 thoughts on “If Numbers Are a Metric, I Give Zero Fucks

  1. As someone who switches off all in-game metrics at the earliest opportunity (albeit mainly for aesthetic reasons rather than philosophical ones) I would prefer my games not to have any option to display them whatsoever. I would go further and not show any kind of stats on gear or characters either. Nor would I display descriptive equivalents like color-coded quality or “Weak”, “Strong”, “Mighty”.

    What I would do is have the apparent qualities of the items reflect their effectiveness at what they do, so a steel hammer would hit harder than a wooden one. How much harder, though? That you would have to decide for yourself by hitting things and seeing what fell down.

    Liked by 1 person

    1. That sounds absolutely fascinating. Reminds me of a BBS door I used to play in which everything was pretty much based on your interpretation of things. I forget what the name of it was or what you actually did in it; I never figured out how to accomplish anything because it mostly confused me. But today I think such a game would be rather enthralling to play.

      I don’t know that I’d go so far as to disable telegraphs in WildStar as you did on your warrior in a game that is designed with such things in mind. Something like Minecraft in which an iron pickaxe breaks blocks faster than a wooden pickaxe, however, without revealing the numbers behind anything (although there are some mods that show you the numbers) is a nice counterpoint to our collective obsession with (or forced homage to) statistics.


  2. I’m a believer in numbers available for personal use. It’s the interpretation (often fairly loose, and sometimes in no way resembling reality) by others that makes me want to pull out my hair.

    Preventing others from easy access to those makes the convenient blame game much harder to resort to.

    I wonder though if it would encourage self-reflection and learning if statistics showed you your own personal stats in comparison to a generalized average (taking into account your matchmaking, class roles or what not) and added a few positive coaching comments. Something like “Your healing this match was 13% better than the X average, keep it up!” or ” the average assists for X class was 8, and you achieved 4 this match, try to help your fellow players more next match!”

    Liked by 2 people

    1. I wouldn’t have any major objections to making personal Elo ratings available to ranked players in Smite as part of their private profile. You would run the risk of having other aggressive players trying to extract their Elo rating from them in chat, however. (Not to mention that this information is already available via third parties in the form of unofficial estimates.) I don’t know that I would be on board with making ratings available to unranked players, though – seems a bit counterproductive for those who are playing “casually” (i.e. in non-ranked games).

      I think some sort of coaching function that relied on statistics without necessarily referencing them directly wouldn’t be a bad idea, perhaps as a checkbox (“Coach Me!”). Comparing your statistics to a global average based on your rating and making that information visible to you…I, too, would be curious to see whether that would be helpful. Interesting idea.


  3. I love pouring over the post-match data in League of Legends, and I can see where the same sorts of systems already exist within Smite. I have been dabbling in the latter game as of late, so I can see where I would enjoy the numbers in it just as much as I do in LoL.

    You’re right though, the numbers allow us to be lazy. Rather than pouring over a replay of your game and making notations of the things you (or your team) did wrong (or right), we look at numbers that say “I had more kills than you” or “I have more CS than the rest of the team” or “I did more damage to champions than anyone else.” I highly recommend recording your own gameplay and watching it later to improve. Watching pro players is also a great way to learn tricks of champions you like to play (or Gods in the case of Smite).

    I try to take into account the mistakes made each round and correct those, as we can only really control ourselves. We can only improve our own gameplay. People are assholes (myself included) but most people at higher levels of play tend to be more rational. Teams tend to flame less when you (and your team) are doing everything in their power to win but just give up that first blood that snowballs into a win for the other team. It happens to the best of players.

    I think stats when applied to learning are great. Just using them as a way to add inches to your ePeen doesn’t really do much for anyone.

    Liked by 1 person

    1. I’ll look over the post-match stats briefly to see whether there was an area that objectively didn’t match up with my subjective feel for things. I’ll be playing unranked (casual) for quite some time so I’m not yet at the point where I’m interested in watching replays of my own matches.

      Instead I have watched lots and lots of pro players for every role. I’m still trying to decide which role I prefer. And I’m still working on improving my game both mechanically and strategically by mastering every god and getting in 500+ unranked games before ever setting foot in ranked. So I have a while before I’ll be playing at a higher level.

      I agree that stats are great as long as they’re used mindfully.


      1. You have the right approach, but I’d advise against mastering all Gods. Play them all, sure. But only master those that fit the role/lane you are best at. I have a dozen or so champs I keep practicing that shift with the meta. At least this is legit advice for LoL, where there are 130+ characters to learn.


      2. I’ll keep that in mind. There are 74 gods in Smite right now and I currently have 16 of them mastered with several almost mastered. This weekend’s double everything event is speeding the process along nicely. If it gets to be too much of a grind I’ll consider mastering the top 3-5 for every role (since you’re expected to be able to play pretty much any role competently in Smite conquest).


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s