Does mastering a game really mean committing 10,000 hours to it, or is there something else?
How many times have you heard, thought or said: “I’m going to get really good at (insert game name here). I’ll just put in 10,000 hours of practice and bam! I’ll be rolling in esports mega-bucks with a big sponsorship (hi, SteelSeries!) and professional team contract.”
There is a lot in those couple of sentences to be criticized, debated and argued. But let’s talk about the number: 10,000 小时. It turns out there is some scientific “controversy” about the 10,000 hours idea, but I’ve put that word in scare quotes because it’s a classic scientific argument where there’s a lot more everybody agrees about than disagrees.
Better yet, understanding the controversy might even give you some ideas for becoming a better gamer yourself, whether your intentions are to go pro or just get better.
Credit: Anton Porsche
The Power Law of Learning
The idea that you need to practice to get better is not controversial. Nor is the idea that you need to practice a whole lot to really get to a level of top-level competitive ability. A key aspect about why you need so much practice is captured in the theory of The Power Law of Learning, which sounds cooler than it actually is. The name does not actually refer to powering up with practice. Rather, it is based on a mathematical function (power function) that is one of a class of negatively-accelerated curves that are effective for describing learning rates.
That sounds fancy, but in fact, it is something every gamer already knows: the better you are, the more you have to practice to improve further. When you first start out, you learn really fast – the learning curve is technically steep, but here that means easier. But as you get better and better, you have to work a lot harder to get additional gains.
Credit: Element5 Digital
To get an idea of what this intuitive idea means in practice, if it did take you 10,000 hours to become a competitively elite player, with typical values of the function, you should get around 10x better in the first 1,000 hours that you play – literally 1,000% better. Then it starts slowing down. A lot. From 1,000-2,000 hours, you only get about 25% better than you were at 1,000 hours. And it keeps getting worse. Over the last 1,000 hours of practice, you’re only improving by around 3-4% more, which is a lot of work for pretty small gains.
This makes sense of a lot of things hardcore gamers already know. For example, the skill gap among the best players in the world at the game is generally very small. They are all way, way better than us filthy casuals, but the gaps among the top-ranked aren’t large. Somebody who is playing for the world championship and is executing consistently 2% better than their competitor is probably going to win – as long as there isn’t a lot of randomness in the game (which is why pro players hate RNG games in general).
OG didn't win back-to-back Internationals without a whole lot of practice.
As another example, learning a new game is generally pretty fun, but pushing to get really good often feels like a grind. We all like the feeling that we’re getting better when we are playing. That’s going to make the beginning of the learning curve where we’re rocketing up in ability (and MMR, ELO, etc.) feel very rewarding. That’s one reason you end up with a whole lot of games in your library. But there’s something special about committing to pushing yourself to keep improving over thousands of hours of practice where the gains are slimmer and slimmer.
As a reference point, if you worked at a normal job 40-hours/week (9-5 M-F), you’ll work 2000 hours in a standard work year (52 weeks with 2 weeks vacation, 50 x 40 = 2000), so we’re talking 5 years of full-time effort to get to 10,000 hours. Gamers probably train harder than that, but we’re still talking years of concentrated effort to get near the top.
Nothing about the basics of power-law learning is really controversial (ok, technically, there’s some controversy about whether the power function is the best description or if it’s some other similarly shaped curve like an exponential function and whether the curve should be considered smooth or have some step-like elements, but none of those arguments change the basic overall shape or idea that you have to practice a lot and honestly, scientists fight about everything, it’s kind of intrinsic to how scientific progress works). The “controversy” about 10,000 hours has to do with how it’s gotten out into popular understanding from the scientific literature and what it actually implies about how you should approach your practice.
The idea of “10,000 hours” was made popular by Malcolm Gladwell in his book Outliers. Gladwell is a wonderful writer, but he’s a science journalist not a scientist. The research he drew from to make this point was actually done by Dr. K. Anders Ericsson, a professor of Psychology at Florida State University. Curiously, Gladwell’s popularization of the concept seems to have annoyed Dr. Ericsson considerably. Although researchers like myself in the general field felt that Ericsson should get the credit for the core idea, on his website, he posted a fairly long rant essentially denying the idea was his and insisting he not be cited as the source of the idea.
Practice vs. Deliberate Practice
The problem with the popularization of the “10,000 hours” idea is that the research Dr. Ericsson was doing was focused on differences in achievement after 10,000 hours of practice. He was studying expert musicians who had all put in 10,000 hours of practice but some seemed to be performing at a clearly higher level than others. The core theoretical idea he was advancing was the importance of “deliberate practice” which reflects what you actually did during those 10,000 hours. He was specifically arguing against the idea that simply hammering away for 10,000 hours would make you best in the world.
It's been 10,000 hours, am I an expert yet? Credit: Marius Mann
Instead, you had to practice in a focused, structured, way, perhaps with help from high-level coaches, trainers and instructors. The popularization of the idea generally doesn’t discuss that idea, so I suppose you can see why Dr. Ericsson was unhappy with Gladwell’s description (note to Gladwell, Michael Lewis or any other really talented science writer – if you would like to make any of my scientific work really famous, even with some slight mis-description or unexpected emphasis, please go right ahead. Or call me, really, anytime).
Turns out it takes more than time to get good at Dota 2.
Again, for gamers, the idea of deliberate practice is very intuitive. I have a lot of hours in Dota 2 - in fact my Steam interface report currently has me at over 10k hours playtime, and I’m completely terrible at the game. Maybe it’s partly because I’m old, but I’m pretty sure I haven’t approached learning the game properly either.
The interesting question for both gamers and scientists is figuring out what the best kind of deliberate practice is for the game you want to get good at. The real limitation of Ericsson’s deliberate practice theory is that it doesn’t tell us what it is. For Dota 2, better players always tell me to get better at last-hitting creeps. I remember watching talented players streaming last-hitting practice with really low damage heroes (or even creeps) and I can’t do that. However, I’m also pretty sure I need more than that, but I’m not sure what it would be.
If you are trying to figure out what deliberate practice is for the game you are playing, remember that it is generally not just spamming public matches. You can see examples in other areas that people become expert. In any athletic (physical) sport, there are always specialized drills to hone specific skills. Musicians practice scales and training pieces. Back when I played chess, the good players told me to study endgames more. Dr. Ericsson’s group just published an interesting analysis of Scrabble players focused on the deliberate practice approach of memorizing all the short allowable words in the dictionary (e.g., up to 4 or even 5 letters). In each case, a specific subpart of the game is selected out and trained intensively.
Figuring out what the right deliberate practice is for your video game of choice will probably help a lot. My son who plays Overwatch well spends a fair amount of time in ‘headshot only’ custom maps to hone his aim. A graduate student in our department who is expert in Super Smash Bros talks about ‘frame analysis.’Back in the day, I remember hearing about StarCraft/SC2 players who trained on the game Osu to get their APM up to speed.
Above: Anssi “mxey“ Pekkonen – FaZe Clan PUBG
Whatever you are playing and trying to get good at, there’s something outside the main game but related you should probably be training on. To get better, you can probably learn from players who are better than you not just in game, but by finding out what they are doing for deliberate practice that isn’t just playing. The problem, of course, is that training is less fun than playing, so if you are like me, you might just end up just playing Dota 2 for fun, missing those last hits and generally playing like some strangely highly experienced noob.
So let that be a lesson to all you ambitious youngsters out there: it’s not just 10,000 hours but what you do with them.
Spend them poorly and you’ll end up a 10,000-hour-bad like me. I should really go launch Demo Mode and hit some creeps right now – but you know I’m going to just go pound the Play Dota button and just embarrass myself as usual.