How to beat a much better player?

Well, the question came to my mind while watching a video annotation of a US Champ game between So and a junior GM whose game I don't remember. The 14 year old guy defeated So (with around 200 points difference in rating, which is more significant at higher levels I guess).

Ok, the question is simple. I am 1450 in rapid on Chess.com; would it be possible to win (or draw) against a, let's say, 1800 player without relying on a big blunder by him?

If no, what would be the maximum difference in rating for a game with a result which is not completely predictable?

If yes, how to play it? Try a rare, puzzling opening where innacuracies are more likely? Play solid and defensive in the opening and hope for a draw? Try an open game with lots of tactics, hoping to outcalculate him? Try to defeat him positionally?

All seem pretty impossible, but even if the odds were very small, I would try to grasp them and maximise them by steering the type of game into some specific direction.

So, any opinion is welcome.

Thank you very much

Sam Sevian!

First maybe it's useful to consider that all games are won because one side makes mistakes.

So as you might expect my answer here is no. You could only win after a big blunder (or multiple smaller blunders).

This is an interesting question. First of all consider this chart:

Notice a difference of 200 points means that the expected result is 76% to 24%. In a 10 game match this means the stronger player would be expected to score 7.5 to the weaker player's 2.5 (any mix of wins and draws). So as you can see, it's never "completely predictable."

In my personal experience, I'm still optimistic against an opponent rated 200 above me. 300 not so much, and at 400 I play my best, but honestly I expect to lose.

The chart is just what the math says. People have crunched some real life data to see how it compares. What they found is when the gap gets much wider (say, a 700 point difference) the lower rated players score better than they should. Something like 5% instead of 1%

Against a much stronger player you have less chances of winning in a dry technical position. Their experience and knowledge will really outshine yours.

Because much of their strength comes from study and experience, you have better chances of winning if you minimize that and make them rely on calculation. You do this by playing a sharp, aggressive, chaotic game. It gives you a sort of puncher's chance. If they miscalculate you can win.

Don't go nuts from the opening, develop all your pieces for example, but yes, essentially play for tactics.

The above is what you'd do if you must win (lets say the last round of a tournament and you want the cash prize or a title). However my advice for most of your games against much stronger opponents is to play what you know best. Yes, even if it's a stale defensive positional game where you're more likely to lose. Because that way, you maximize what you can learn from the game. If you play outside of your usual style and positions, your mistakes are more likely to be simple oversights due to inexperience.

[q]First maybe it's useful to consider that all games are won because one side makes mistakes.

So as you might expect my answer here is no. You could only win after a big blunder (or multiple smaller blunders).[/q]

I don't agree with this. Unless you consider every inaccuracy as a mistake. But many games are won due to a sum of small inaccuracies.

So, couldn't I win simply by playing more accurately and reaching a better endgame? The question is if I could outplay him, on one of every 10 games

Yes, I was regarding any move that changes the evaluation a mistake.

But ok, so your question is if you could outplay them slowly over many moves even if their rating is much higher.

Sure, it's possible. There are many different ways to get a rating i.e. each player has their own strengths and weaknesses. If you get a position where your strengths match up with their weakness, then you'll have a good chance.

Also, just as in sports, sometimes a person plays better than usual. There's a quote that goes something like this: second class players can beat first class players because sometimes second class players will play a first class game.

If you have to beat a much better player, then your best shot is to play some wild line of an opening and steer into unclear tactical positions at every turn.  This is somewhat akin to to closing your eyes and swinging wildly at a boxing champion, but it's still a much better chance than going toe to toe.

Conversely, if you just need to draw a much better player, you have a better shot by steering into the dryest open game you can, one where exchanges are hard to avoid, then trade pieces often and simplify the position at every opportunity.

If you are trying to draw and you get a couple pawns down with queens still on the board, try for some kind of sacrifice that allows for perpetual check.  Players will often correctly discern that your attack has no mating potential and dismiss it for the moment, but can often overlook that you can force perpetual check.

Play the man, not the board.

I have a blog post on this topic. http://chess-for-engineers.blogspot.com/2015/01/chess-graphs-risk-vs-reward.html

Against a stronger player, do not be afriad to go into a complex position where sharp chess abounds. As strange an advice as that sounds. In technical positions stronger players have a more finely tuned auto-pilot than you do.

A big problem lies in the title. If you believe the opponent is "better" you are well on the way to being beat anyway. A rating is a past indication of performance and doesn't represent the present. We should live in the present and deal with THAT. That's the plan ! :-)

Like the others said, you have the best chances in a chaotic, tactical position because it is easier for both players to miss something... This could mean, for example, to sacrifice a pawn in order to start a wild kingside attack. Simon Webb wrote a nice chapter on this in "Chess for Tigers".

It might depend on the rating though... I guess in the 1000-range it might be more a question of keeping your pieces together.

If you are playing more accurately then they are not a much better player; either you would be underrated or they would be overrated... or both.

The way to get a stronger player to make a mistake is to play a sharp, complicated battle and hope he goes astray in the forest. 

If you play a simple positional game, you will almost always lose without a fight. The stronger player is almost always going to be better in those positions. 

So, in order to pull off the big upset eat your Wheaties, play aggressively, and look for that great knock-out blow!