In order to improve, is it better to play handicap or even games against stronger players?

The purpose of handicap stones is to make the game competitive. If your opponent is stronger than you AND is willing to play a non-competitive, or less competitive game, then by all means forego the stones.

I'd say that handicap stones are also (mainly?) there to provide interesting learning situations.

A few weeks ago, I played an even game against a player I could give 8-9 stones to. When I started doing the review I realized that it would be very hard to comment on his moves because I would have to comment on almost every one of them. Handicap games on the other hand make it possible to focus on the main ideas: "separate white", "connect your stones", "reduce white's eyeshape", etc.

If you are asking, you need to play handicap games. For example, Black with a handicap of four or five stones has a "won game". In an even game against a opponent of the same level, if you have that sort of lead, your opponent will try to trick you, or resign. A stronger player can rely more on incremental methods, and genuine fighting skills. So that is instructive.

Of course not all stronger players have the same approach. But if they are relying on tricks, then it hardly matters about the handicap.

I hate handicap games. You learn nothing from them since the stronger player needs to play overplays in order to catch up.

I would rather play a strong player and lose by 100 points than play them in high handicap stones.

No they don't. Playing overplays and being successful in them certainly means you can catch up faster, but it is possible to win correctly handicapped games with patient non-overplay moves, waiting for the opponent's mistakes. Remember that to win by 0.5 on move 300 you can still be 5 of the 9 stones behind on move 100. For what it's worth I will sometimes play this patient style in 9 stone games, sometimes lots of overplay, sometimes in between depending on my mood, the opponent, the situation (competitive vs teaching game) and so on.

I disagree. The stronger player does not need to overplay at all. He just plays the best move available and will slowly catch up due to strength difference. You should play games at the correct handicap most of the time if you want to learn. Playing an even game and losing by 100 points teaches almost nothing. There is nothing wrong with an occasional even game as an experiment (maybe one in five) but it should be no more than that. There is a reason the handicap system was created.

Also, you are unlikely to lose by 100 unless the stronger player decides to do so. He may just as easily coast to an easy win. In such a situation I usually just play to win and once the game is decided I don't push it any further. A 15 point win is still a win, even if I could give the player 9 stones. Of course, he probably will make so many mistakes that I will win by more, but what has he learned when I wasn't even playing optimal moves anyway?

_________________
Still officially AGA 5d but I play so irregularly these days that I am probably only 3d or 4d over the board (but hopefully still 5d in terms of knowledge, theory and the ability to contribute).

If you catch and correctly answer their constant overplays you won't learn much, except what it's like to get accused of sandbagging. If the overplays are working some of the time it's fairly hard to argue you can't learn from the game. It's not like you won't face overplays when playing even games at your own level.

I won't play games with more than a 2-3 stone difference against anonymous opponents, but high handicap games in a face to face context or against people you know can be really fun. It's helpful if someone is willing to review the game with you after you play, of course.

I've tried playing as the stronger player against people 9+ stones weaker than me without a handicap. It's really hard to create interesting positions for someone to learn from when there is that great of a discrepancy in strength. I'm sure it's a skill that can be learned, but I think you'll learn more from most opponents by playing with a proper handicap.

Generally speaking playing with a proper handicap is best for learning. It sets the challenge at about the right level.

Consider the extremes. If the handicap is so small that you are destined to lose the game, no matter what you do you will lose, so how can you compare moves? (In go, yes, you can compare how much you lose by, but that is so no matter what the handicap.) OTOH, if the handicap is so large that you are destined to win the game, how can you compare moves? You get the most information from winning or losing when you have a 50-50 chance of winning. That is, when you have a proper handicap.

_________________
The Adkins Principle:
At some point, doesn't thinking have to go on?
— Winona Adkins

Visualize whirled peas.

Everything with love. Stay safe.

I fully agree with this and this is the reason I prefer handicap games nowadays. Handicap games tend to be competitive and tense for a longer time, which leads to learning throughout the game. There's nothing wrong with an occasional even game to study fuseki and early chuban and then resigning as needed, but there's nothing worse than a game dragging on where you can't bring yourself to play sharply because you lead by 30+.

_________________
La victoire est un hasard, la défaite une nécessité.

If your goal is to get stronger, then your goal is to increase your ability to win games.

You are not efficiently exercising your ability to win games if either of the following are true:
1. It is not very challenging to win the game - you do not practice skills necessary to win games against challenging opponents.
2. It is too challenging to win the game (e.g. you've lost hope) - you don't get to practice exercising whatever it takes to actually win a game.

In order to efficiently exercise your ability to win games, I'd recommend that you aim to strike a balance between having a challenge (will increase your ability to stretch your skill) and winning the game (will increase your ability to close out a game and win it). If you are good at one, but not the other, there's an area to improve.

Therefore, playing with the appropriate handicap is one way to achieve this balance, but you can experiment with different handicaps to become more well rounded.
1. If you don't get much experience in winning games, or (for example) in playing out the endgame, practice playing with a higher handicap so that you can become skilled in this aspect of the game.
2. If you win a lot, and don't get much experience in stretching yourself to get a win, practice playing with a lower handicap in order to increase this skill.

---

I started out this post with the idea that always playing with the appropriate handicap is the way to go. Now, I think that it might be nice to mix things up and challenge yourself in the areas where you are weak.

_________________
be immersed

I think it is better to play agreed-upon teaching games with handicap. This handicap may or may not be less than that recommended by your relative strength differences in order for the stronger player to attempt to work with you on certain lacking concepts, especially those already revealed by a previous game, but there is a very real limit that even games with a stronger player can possibly provide. If the question is about the opening, then maybe it is better to have some even games ready to ask questions about rather than play a new game with a much stronger player (6+ stones difference is my cutoff here).

In my experience, non-handicap teaching games are not especially instructive, and neither are handicap games without prior agreement to be a teaching game.

I think some people just don't give or receive handicap a lot, so they don't know how to handle it, so they discourage it in others, which feeds this cycle of the "even" teaching game in which white isn't even trying and black feels totally bullied.

The difference is not entirely unrelated to life and death problem difficulties. You might say: the difference between the 4d problem and the 10k problem is the number of moves the 10k problem has had moved forward along some sequence the 4d problem only has as a possibility. It isn't that a 10k couldn't understand someone explaining the 4d solution (no new rules of go are introduced after 30k), it's that they could never see it. Then, a handicap game is like an even game with some moves put forward in all the local positions.

To say that there are lots of beginners and new players on OGS, there aren't many handicap games that are played. It seems the opposite to what would happen in a club, or even on KGS. Certainly handicap games train the player with the black stones to secure territory while attacking. If a player starts with 4 stones and loses by 10 points, it is going to teach them more than if they lose by 100 points and spend the game being bullied.

_________________
12k: 2015.08.11; 11k: 2015.09.13; 10k: 2015.09.27; 9k: 2015.10.10; 8k: 2015.11.08; 7k: 2016.07.10 6k: 2016.07.24 5k: 2018.05.14 4k: 2018.09.03 3k: who knows?

Yes.

Just my feeling, but I suspect that is because nobody trusts the handicap system on OGS (or rather, people don't seem to trust its ranking system to divide people such that one rank is reliably worth one handicap stone) and almost all games are ranked.

If I am a 10k and I give a 15k 5 stones, and rank is at stake, then I want that 5-stone game to be an even contest. There doesn't seem to be a feeling that that's the case, yet.

(I play unranked handicap games on OGS, though not many - I would happily play more.)

_________________
Confucius in the Analects says "even playing go is better than eating chips in front of tv all day." -- kivi

If you are right about that, JC, then it will continue that way until people start playing ranked games and stabilize the system. If human pride is getting in the way it will never work.

_________________
Still officially AGA 5d but I play so irregularly these days that I am probably only 3d or 4d over the board (but hopefully still 5d in terms of knowledge, theory and the ability to contribute).

One of the problems with the preponderance of non-handicap games between differently ranked players on OGS is it contributes to unrealistic ranks. If a 1d plays a 8k in a properly handicap game and one player times out or their browser hangs etc and loses then that game result of 1d beats or loses to 8k with 8 stones is not a surprise and neither rank changes much. But if it is an even game and it's the 1d who times out or loses for whatever non-game reason then that result is a big surprise and both players' ranks will change a lot (but not in the reverse situation, an asymmetry is introduced). This was a problem with me playing real-time games on my phone on new OGS, as I would often misclick in even games against much weaker opponents and then time out waiting for them to give the undo (the undo system on OGS is poor, you can't request it on your move so must resort to sending chat messages, which the opponent might not see/comprehend), or resign if they refused undos for misclicks, or time-out due to bugs necessitating a (slow) page reload.

(I believe this "even game default" is a carry-over from the old OGS days where there were only correspondence games, which tend to be more serious and suffer from fewer non playing-strength related outcomes than live/blitz games; there was also a system so that many time-outs when a player left the site would become unranked.)

That is an excellent point, well made.

We (OGS people) all need to play more rated handicap games.

_________________
Confucius in the Analects says "even playing go is better than eating chips in front of tv all day." -- kivi