How to Play Chess Online: time controls, etiquette and real improvement

Online chess means picking a time control, starting a clock and accepting that time weighs as much as the pieces do. Beginners lose more points to three things than to shaky opening knowledge: they play faster than they can handle, they cannot convert a won position before the clock runs out, and they never look at their own games again.

Time controls, and what each one trains

Every time control is written as two numbers. 3+2 is three minutes of game time per side, plus two seconds added back to the clock on every move. 10+0 is ten minutes and no bonus. The first number decides whether you have time to think; the second, whether you have time to finish.

• Bullet (1+0, 2+1): the whole game in one or two minutes. It trains reflexes and patterns, not calculation: there is no calculation in half a second.
• Blitz (3+0 to 5+3): enough time to spot a tactic, not enough to check it. Great for testing an opening, terrible for learning endgames.
• Rapid (10+0 to 15+10): the sweet spot if you want to get better. Room for a plan and a three move line, and it ends inside twenty minutes.
• Classical (30+20 and up): every move can be checked before you play it. This is where the mistake is yours, not the clock.

Increment vs delay

Increment adds seconds to your clock after every move, and it accumulates: move quickly and your time goes up. Delay only postpones the countdown, so you always have time to push a piece but never bank any. Increment is what keeps endgames playable: with two seconds a move you can mate with king and rook; with zero, rarely.

The clock wins games you had already won

Flagging, losing on time, is the most common way to throw away a won position, and it is not bad luck: converting an advantage costs moves, and moves cost seconds.

One rule protects you here: if your clock runs out and your opponent cannot possibly mate with the material left, the result is a draw, not a loss. A bare king cannot win on time, and neither can king and knight or king and bishop. King and bishop against king and pawn does win, because with a pawn on the board mate is still possible.

The premove was built for that squeeze: you queue a move and it fires the instant your opponent moves, costing almost nothing on your clock. It is what makes bullet playable, and how good players hang pieces for free, because a premove does not think.

• Good: forced recaptures, walking your king in an endgame already decided, forcing lines you have already calculated.
• Bad: positions where your opponent has more than one reasonable capture. They pick the one your premove does not answer.
• Fatal: a capture that is still legal after an in-between move. The premove fires, it is legal, and it hands over a piece.

How online ratings work

Almost every site uses a system in the Elo family, usually Glicko, which stores two numbers: your rating and how confident it is in that rating. Early on the confidence is low and the number jumps thirty or forty points per game. After a few hundred games, moving fifty points takes weeks.

• Every time control has its own rating. Your 1200 bullet and your 1200 rapid are not the same skill.
• Ratings do not compare across sites, or with FIDE ratings. Each number only means something inside its own pool.
• Swinging is normal. Six blitz losses in a row do not say you got worse, they say variance, fatigue or tilt.
• The number measures results, not understanding. You can gain 200 bullet points without learning anything.

Etiquette and fair play

You are playing a person, even when all you can see is an avatar. Almost all of the etiquette is about respecting the time of whoever is on the other side.

• Do not stall a lost position. Playing on until mate is fine and winning on time is part of blitz. Burning your clock on purpose to make the other player wait is not.
• Offer a draw once. If the answer is no, play: repeating it every move is noise.
• Accept a draw when it is a draw. King and bishop against a bare king does not become a win because you kept playing.
• No engine. Not "just this position", not "just to check". A game played with help erases the result of someone who played it straight.
• Say hello, then play. Using the chat to point out an opponent mistake never goes well.

And engine use is detectable. No site flags anyone over one brilliant move: what catches people is statistics across many games. How closely your choices match the engine, whether that agreement is too high for your history, and how your time per move behaves.

The improvement loop that actually works

There is no shortcut, but there is an order. Four habits, in this priority:

1. Play longer time controls. You cannot learn to calculate where there was never time to calculate. One 15+10 game a day teaches more than thirty bullet games.
2. Review your own games. Before you open an engine, find the move where the game turned yourself. The engine says what was better, only you know why you did not see it.
3. Drill tactics daily. Twenty minutes a day beats two hours on Sunday, and repeat the patterns you failed until they come out instantly.
4. Learn the basic endgames. King and queen, king and rook, opposition, Lucena and Philidor. That is what decides games when the clock is short.

That is the back rank mate, and it sums up what "drilling tactics" means: you do not calculate it, you recognize it in half a second, both to deliver and to avoid.

The first you recognize, the second you work out from a rule. Endgames are the part of the game almost nobody studies, because bullet never gets there.