How easy is it to become a GM?

Massively, 

When I started this thread my rating was below 1000, I have reached a high  of 1357 earlier this year.  My rating has slipped since, but that's only because I haven't been watching as many Kingscrusher videos. 

@OP

I think something that is key to understanding the difficulty of becoming a GM (or even IM, FM, NM) is the non-linear nature of rating improvement.

At 14 years old, I reached my peak rating at 1985 while playing casually, and without the benefit of a coach, chess club, or computer analysis (my family had just gotten dial-up internet that year, and not the blazing fast 56k either!). The next year I started College and no longer had time to play. It has been 20 years since the last time I played a game of chess. I remembered little beyond the rules and basics. 

With only an hour of study and practice each day, I expect my rating will rise to 1200-1300 shortly. If I managed to attain a 1450 rating in 1 year of play, does that mean that I am on track to reach 2900, a GM norm, and rival Magnus in 2 years? Of course not. But why? 

Elo ratings are not linear, and take the difference in rating, and expected performance against an opponent into account when determining movement in either direction. As such, several factors come to play a more significant role as your rating increases:

1.) Elo is more reflective of an exponential scale than a linear one: A 2900 rated player is far more than twice as good as a 1500 rated player. A 2900 could likely give a (possibly blindfolded) simultaneous exhibition against 4 1500 rated players, with half the clock time, playing black, and winning all 4 matches without significant difficulty. As your rating goes up, so does the quality of your competition, and by a more significant degree than a linear Elo comparison would suggest.

2.) Early Increases are Unsustainable: If your Elo is 800 by virtue of games played, but your actual skill level is closer to 1500, and you are playing opponents with higher ratings (1220, 1065, 914, 920) and performing well above expectations for someone who's actual skill level is 800, you will see your rating climb rapidly. Conversely, if your rating is 1500, and you are evenly matched with 1500 rated players, you will see your rating increase at a much slower rate. Don't forget, your opponents are learning and improving just as you are!

3.) Pickings get slimmer: When playing low level competition, there are plenty of novices who are easily beaten. Also, players rated 1200, 1400, 1600 have a much greater instance of committing major blunders, hanging pieces, or sacrificing for a mating combination that was inaccurate. This often results in games you were loosing, or may have squeezed out a draw in, turning into wins for you. Winning is the best way to increase your Elo. However, players rated 2400+ will almost never blunder, or hang a piece unintentionally, or miscalculate a major sacrifice. Players rated 2600+ are so good, you rarely see them do something as simple as hang a pawn without positional compensation, since a well located pawn advantage between 2600's is similar to being up a rook between 1500's. 

The result of this is that wins become less common, and draws become more common, as you ascend to higher levels of competition, thus slowing your rate of Elo progression drastically the higher you go. No matter how good you are, forcing a win against a GM is no easy feat, so even playing objectively "better" against a highly rated opponent, you are less likely to end up with a win, just because they are so much better at not loosing. 

When playing in a tournament with top level competition, you quickly come to realize that those you are playing against are approaching the pinnacle of human ability in playing this game. The closer you get to the ceiling, the more difficult those small increases in rating become. Going from Elo 2450 to 2500 is far more difficult than going from 1500 to 1600.

4.) Age comes for us all: The sad reality is that our cognitive abilities decline with age. Some GM's who have been in the top levels of competition for many decades devote 12-14 hours a day, 6-7 days a week, to studying and playing chess and still decline in rating as they struggle to calculate with the efficiency and depth of younger players. There is a reason it is unusual for people to become GM's late in life, and most of those have been playing since childhood.

An analogy: 

Chess is like any other elite sport today, and should be looked at in a similar way when setting goals. There are several realities you should confront with perspective and understanding:

A. Your potential:

Is not unlimited, and you may simply lack the mental prowess to attain the rating you aspire to. (Example: if you are 4'6" and 95 lbs, no amount of practice will enable you to dominate the NBA like LeBron James. If you have a naturally slow running gait due to your bone structure, you will never compete with Usain Bolt in a race, even if you run 20 hours a day for 20 years. If you were born with Osteogenesis Imperfecta, no amount of weight training will allow you to even compete in Heavyweight Olympic Clean-and-Jerk. You have to have the potential, and some people's minds don't work in a way that is conducive to elite chess play. That's not an indictment, it's simply a fact of life, and part of the imperfections that make us all different. (Some brilliant chess players struggle with social interaction because their minds don't work as well in that way, see Fischer.)

B. Diminishing Returns:

As any athlete approaches the pinnacle of their personal limits, and pushes the boundaries of performance in their sport, harder and harder work, more and more exhaustive practice and study, and greater amounts of effort are rewarded by smaller and smaller gains in performance. The same is the case in chess. Your 5 hours a day, 5 days a week study may net you 100 points of improvement every year until you hit 1900, then suddenly you find that 5 hours a day for 5 days a week is barely enough to add 25 points in a year. At 2000, you find that it's barely enough to help you hold on to your 2000 rating, and you're not really improving at all relative to your competition.

C. Your opponents are improving as well:

Nothing is static. You are one-dimensionally equating practice to result, but an Elo rating is based on your performance against other players. Putting in 5 hours a day will not necessarily mean you getting to 2500 and a GM norm. The fact is that those players at 2000+ are practicing, studying, and improving as well. Trying to become a GM with 5 hours a day of practice is presuming the GM's you'd have to compete against and equal or beat consistently, don't practice and play more than that themselves. They are constantly improving their game with greater practice and stiffer competition than you are, and in order to become a GM, you need to perform well AGAINST them. 

In summary, you may find yourself a GM in 10-15 years if you follow your plan and have a natural talent for chess that you have never fully explored. However, it is far more likely that you will improve considerably and be a very good player, but find yourself reaching the peak of your ability somewhere well below the 2500 rating. There is nothing wrong with striving to be a Grandmaster, and nothing wrong with achieving an International Master, FIDE Master, National Master, or any other honorific in the chess community. However, the odds of anyone becoming a GM without either being a savant at visual spacial calculation, or making it a full time career that you live and breathe every day, is exceedingly low. 

Good luck in your attempt, but I advise you to temper your expectations so you are not so easily frustrated when you spend a year studying 5-6 hours a day, 5-6 days a week, and can't seem to get out of the 1900's no matter how hard you study. You will experience it, and then you'll understand for yourself how much respect is truly owed to those who have become Grandmasters. 

It's very easy. Here it is:

The Elo system is designed statistically. Every 200 points represents a standard deviation, an entire class difference in skill.

A1600 is twice as good as a 1400.

The Elo system is designed statistically. Every 200 points represents a standard deviation, an entire class difference in skill.

A1600 is twice as good as a 1400.

@CatPetter

There are several people who have become GM's at late ages. GM Ben Fiengold for example, was 40 when he got his GM norm. But as in most cases (like Ben) they have been playing since childhood and gradually attained that level. Those who ascend to GM status quickly, and without making it a lifelong professional pursuit, are gifted savants who were born to play chess like Gretzky was born to play hockey. 

@kindaspongy

There are a number of ways to "test" that, all of which have varying degrees of relevance and accuracy. 

1.) Hold a series of games between an 1800 and a 1400, with both sides alternating colors. The results should bear out the comparison, and become more indicative of accuracy based on the number of games played. (e.g. If 100 games are played with 1800 winning 40, 1400 winning 10, and 50 draws, or 10 games are played with 1800 winning 5, 1400 winning 1, and 4 draws.)

2.) Hold a simultaneous match between 4 1400's and an 1800 with the 1800 given 2 hours, and the 1400's given 2 hours each. The 1800 should win all 4 games with white, or at least not loose any.

Again, this is a rough way of approximating skill, and doesn't account for quite a few factors that are relevant to the discussion, but it would give a decent idea. 

I think it is also prudent to point out that Arpad Elo, the mathematician who invented the Elo rating system used to rate players in chess and other 0-sum games, once described it as, "the measurement of the position of a cork bobbing up and down on the surface of agitated water with a yard stick tied to a rope and which is swaying in the wind". In other words, it is at best a rough and vague approximation of something which is ever changing, dynamic, and affected by a litany of un-calculated variables. 

Simply put, "how good are you at chess?" is not a question that is scientifically quantifiable with a single monolithic rating number. What kind of chess? (Slow, bullet, blitz, bughouse, 3d?) In what circumstances? (In your living room with your dad, casually at a chess club, in an international tournament?) With what adversity? (Playing black, pawn odds, rook and move odds, blindfolded, simultaneous?). Kasparov at his peak was objectively better than Anand at his peak, but Anand is a masterful blitz player, and possibly better than Kasparov given their respective strengths. 

That's my 2 cents on ratings.

@CatPetter

There are several people who have become GM's at late ages. GM Ben Fiengold for example, was 40 when he got his GM norm. But as in most cases (like Ben) they have been playing since childhood and gradually attained that level. Those who ascend to GM status quickly, and without making it a lifelong professional pursuit, are gifted savants who were born to play chess like Gretzky was born to play hockey. 

@kindaspongy

There are a number of ways to "test" that, all of which have varying degrees of relevance and accuracy. 

1.) Hold a series of games between an 1800 and a 1400, with both sides alternating colors. The results should bear out the comparison, and become more indicative of accuracy based on the number of games played. (e.g. If 100 games are played with 1800 winning 40, 1400 winning 10, and 50 draws, or 10 games are played with 1800 winning 5, 1400 winning 1, and 4 draws.)

2.) Hold a simultaneous match between 4 1400's and an 1800 with the 1800 given 2 hours, and the 1400's given 2 hours each. The 1800 should win all 4 games with white, or at least not loose any.

Again, this is a rough way of approximating skill, and doesn't account for quite a few factors that are relevant to the discussion, but it would give a decent idea. 

I think it is also prudent to point out that Arpad Elo, the mathematician who invented the Elo rating system used to rate players in chess and other 0-sum games, once described it as, "the measurement of the position of a cork bobbing up and down on the surface of agitated water with a yard stick tied to a rope and which is swaying in the wind". In other words, it is at best a rough and vague approximation of something which is ever changing, dynamic, and affected by a litany of un-calculated variables. 

Simply put, "how good are you at chess?" is not a question that is scientifically quantifiable with a single monolithic rating number. What kind of chess? (Slow, bullet, blitz, bughouse, 3d?) In what circumstances? (In your living room with your dad, casually at a chess club, in an international tournament?) With what adversity? (Playing black, pawn odds, rook and move odds, blindfolded, simultaneous?). Kasparov at his peak was objectively better than Anand at his peak, but Anand is a masterful blitz player, and possibly better than Kasparov given their respective strengths. 

That's my 2 cents on ratings.

Kids do learn faster, but I think that's because there is less commitment, at a later age. People simply don't have 5 hours a day to play when they get older. I've done some calculations, and if I just improve my rating by 100 every year for 15 years, I would be able to get there. 

… @kindaspongy

There are a number of ways to "test" that, all of which have varying degrees of relevance and accuracy. 

1.) Hold a series of games between an 1800 and a 1400, with both sides alternating colors. The results should bear out the comparison, and become more indicative of accuracy based on the number of games played. (e.g. If 100 games are played with 1800 winning 40, 1400 winning 10, and 50 draws, or 10 games are played with 1800 winning 5, 1400 winning 1, and 4 draws.) ...

@darkunorthodox

Not misleading at all. I think perhaps you misunderstand the point of my statement.

My point, which that statement was intended to communicate, and evidently failed to do for you, was that:

A. It is not impossible to become a GM at a late age in life.

B. Those who do so are typically not newcomers to chess who attain greatness without having a tremendous natural aptitude.

C. Those who do become GM's late in life are usually lifelong chess players, and often would have attained that rating earlier in life had something not delayed them (see my earlier post). It is tremendously difficult to attain a GM rating if you start playing in your 30's. 

D. Simply put, becoming a chess GM (around 1500 in the world currently) is a feat on par with being an NFL, MLB, NBA, NHL, etc... starting player. This is a tremendous and rare feat that only a small percentage of the population can attain. In most cases, those elite athletes play and practice from childhood, and have natural aptitude in their sport. That is not to say that rare untrained walk-on's in their late 20's don't occasionally make it but that is the exception, not the rule.