You Are Not Your Rating

In chess, we frequently say, "I am a 1500 player," or whatever your rating is.  When meeting another player, it's one of the first questions we ask.  When we ask questions online, we invariable include our rating.  We celebrate when it goes up and mourn when it goes down.  We place great weight on this number, to the point it often becomes the defining feature of us as a chess player.  "I am a 1500 player."  The very words we use show how we identify with our rating... though everyone also thinks they are underrated.  I may be 1500, but I deserve to be 1800, that sort of thing.

I want to push against this.  Telling me your rating tells me almost nothing about you, to the point that "I am a 1500 player" becomes empty and almost meaningless.

Yes, this is a hot take, and yes, I am exaggerating for effect ... for only slightly exaggerating.  Our ratings certainly tell us far less than we might think, though, and I want to give a gentle reminder that the number next to our name is not the be-all and end-all.  I've written on this topic before, but it deserves a deeper dive.  Let's go.

What is a Rating?

In any sort of competitive endeavour, no one enjoys a mismatch.  We want games to be close, with either side having a chance to win.  In a perfect world, games are 50/50. Either side can win any given game, but over time, the results stay even.  If we achieve this, we get competitive balance.  The players are evenly matched, and in chess terms, they would have the same rating.

This is the main takeaway: ratings are a predictive tool.  If two players have the same rating, they should have the same results.  If not, then the ratings need to be adjusted, and it will keep adjusting until it gives accurate results. Once it does, then it is stabilized.  You have your rating.

Most of us think of our rating in terms of progress, with a rating going up meaning we are getting better.  Certainly, a higher rating means you are predicted to do better against lower rated foes, and so it serves a useful proxy for improvement, but it has several limitations:

1. It's not just you.  Your opponents might be getting worse, or they might have a bad day.  Your rating can go up based on the actions of others.
   
2. It's based on past data.  It looks at all you have done and then predicts what might come next. Effectively, it is a number based on what you have done, not who you are now.
   
3. It's based on long-term trends.  Individual games mean little.  You can have a good day / week / month and gain rating ... and then "lose" it over the coming days / weeks / months.  We didn't get better and now are getting worse; we are simply smoothing out the short-term variability.
4. Relatedly, a rating is a static number.  Improvement cannot be measured in isolated incidents but only over time.  It's not enough to reach a new rating point, but rather you have to stay there.  This dynamic aspect is ill-suited for a static number.

 Ratings can help us show progress, but it's main purpose is predictive.  The two are not the same.

Predictions, Not Descriptions

When we say, "I am a 1500 player," we often think we are saying something about ourselves, as if it were equivalent to "I have an IQ of 125," or "I got a 175 on the LSAT."  It isn't.  Your rating tells you almost nothing about you as a chess player.

Case in point, take Alekhine and Capablanca.  These former greats had no official rating, but as World Champions, convention has them both rated at approximately 2700.  Based solely on their rating, we might think they played similar chess, but nothing could be further from the truth.  No one would confuse the mercurial and magical Alekhine with the steady, machine-like precision of Capa.  The same rating, but very different personalities.

The same applies at the mortal level.  "I am a 1500 player" says nothing about you.  It is not a statement of your tactics or strategy or opening preferences.  It tells us that if you play against a 1200 you will likely win, but it doesn't tell us why and it doesn't tell us what you need to do to improve.  It is a prediction, not a description of who you are.

Predictions, Not Skill Sets

I said earlier that two identically-rated players are equally matched.  Notice my words: equally matched, not equally skilled. I've written before about chess skills, about the smaller base components that make up your overall chess ability.  Let's take two chess stereotypes as examples: the firebrand club player that loves to attack and has never studied the endgames vs the retired master whose brain is slowing down from age.

These players might have the same rating, but they are very different chess players.  We can see that by looking into their sub-skills, that is, the smaller chess skills that make up their overall chess skill: tactics, calculation, positional play, etc etc.  We might get something like this:

Both players are rated 2,000, but they get there in very different ways.  The Firebrand is a tactical monster with an aggressive opening repertoire ... but if he can't attack or Queens get traded he plays much worse than his rating.  Conversely, the Elder has excellent positional understanding, but opening knowledge is dated and his calculation capacity has withered with age.  Both are rated 2,000, but their individual situations and skillsets are worlds apart.  Indeed, their respective ratings are the least interesting part of their make-up.

Said another way, imagine both of them ask, "How should I improve?  I am rated 2,000."  Would you give them the same advice?  Of course not.  They are in completely different situations.

Trends, Not Instances

Do you know how much luck is involved with chess?  I used to think chess had no luck, and while it clearly is not a dominant factor, it still exists.  For example, let's say you I play one game.  According to lichess, I score 54% with White and 48% as Black.  That's a 6% swing before the game even starts.

I get White and play 1.e4.  How does my opponent respond?  If my opponent responds with 1...e5 or 1...e6, then my scores jumps over 60%, but if he plays the Sicilian, Black's score jumps way up.

In other words, my opponent's choice of repertoire, something I have no control over, plays a large role in how the game might play out.  This also works in reverse.  Historically, as Black, I score near evenly against 1.e4 and 1.d4, but in the last few months, I've been doing terrible against 1.d4:

My opponent's chance to win basically doubles if he is a 1.d4 player.  Again, I have no control over this.  Before the game even begins, the choice of White or Black and my opponent's repertoire, drastically influences the result.

I might play several games when I get nothing but 1...e5 and 1...e6 from my opponents and thus win a lot.  My rating goes up!  Yes!  But then the next bunch of games are against Sicilians, and my opponent's play more 1.d4 as White.  I lose more.  My rating does down.  Am I getting worse?

No, of course not, because improvement is not based on individual games, or even a small batch of games.  These are just small, isolated instances, and what matters is the long-term trend.  Indeed, my 1.d4 results as Black are basically meaninglessness; it's only 14 games!  If the score still looks like that after 140 games, then that's bad.  Right now, though, it's just a spike of variability.

Especially in online games, this luck factor can spike our rating in either direction.  Our rating is changing, but it has nothing to do with us.  The more we focus on our rating, the more we are at the mercy of these luck-based rating spreads.  Our ratings tells us nothing about one game, not five games, not even 10 games.  It's only the long-term trends where our rating can tell us something.

The Take-away

At the time of writing, my blitz rating is 2203.  What does that tell us?  Well, if I play a game against another 2200 player, I should win 50% of the time.  That's the only thing it tells us.  It doesn't tell you what I am good at.  It doesn't tell you my weaknesses.  It doesn't say what I should do to improve, nor say that a given book is too easy or too hard for my ability.  It only tells us who I'm predicted to beat.

I also haven't played a game since July, so it's not a particularly useful rating, either.  I was 2200 in July; who knows if I still am today.  My rating is only as useful as the most recent games I've played.

You are not your rating.  Your rating is merely a predictor of what might happen.  I've always found this reassuring, even positive.  My rating doesn't define me.  I'd like to achieve rating goals, but I don't need to chase after them.  If I focus on me and my chess journey, my rating will take care of itself.  That number, by itself, doesn't say anything about me.

When I went from 1400 to 1800, I celebrated the milestones but otherwise rarely thought about my rating.  When I eventually plateaued at 1800, I thought about my rating a lot, and that was the least happy time in my chess career.  Plateauing is never fun, but focusing on it makes it worse, and it's no surprise I quit chess for several years.  When I eventually came back, I focused on studying and enjoying the game, ignored the numbers, and within two years I broke 2000. 

Ratings are like cats.  The more you focus on them, the more they run away.  But if you ignore them?  Then they come to purr in your lap.

[Insert picture of Smithy with cat on lap]

Now, of course, a higher rating is nice to have, and certainly there are more 1400s than there are 2400s.  And certainly I have cherry-picked some extremes, as many 1500s are fairly similar ... but none are identical, and I'd say the majority are distinct in several ways.  I can talk about trends, stating that most 1500s have certain strengths and weaknesses, but that's only a trend.

And I think that's great.  You are not your rating.  Your rating is merely a reflection of your past results and a predictor of future ones.  And honestly, it doesn't do it that well, because they sure do swing wildly.

So celebrate your rating wins but don't feel defined by it.  You are not your rating.