How to Practice Robot Games

Playing a lot and practising are different activities, and only one of them is efficient. Deliberate practice in arcade games is simple to describe and almost nobody does it, largely because the instinct after a bad run is to restart immediately.

Isolate one variable per run

The central principle. Trying to fix your timing, your gaze and your target selection in the same run generally fixes none of them, because attention is finite and each requires conscious effort until it is automatic.

So pick one thing per run and let everything else be as bad as it wants to be. Your score will drop, sometimes substantially, and that is the correct outcome - a practice run is not a scoring run.

The usual objection is that this wastes runs. Runs are between forty five seconds and a few minutes, so a focused practice session is a handful of minutes. That is a small price for the difference between improving and plateauing.

Practise the band that beats you

Most players' runs end within a few seconds of the same point. That is not bad luck; it is a specific difficulty band their reading has not adapted to.

Once you have noticed where it is, stop practising the whole run. Play the earlier part loosely - conserve attention, skip pickups, take the safe option every time - and concentrate hard from just before the band where runs usually end.

This converts a two-minute run into deliberate practice for the fifteen seconds that matter, and it is the main reason some players improve quickly while others repeat the same run for weeks.

Stop restarting on instinct

The moment a run ends, the information about why it ended is available and perishable. Restarting immediately discards it.

Three seconds is enough. What killed you, where you were, what you were trying to do, and whether it is the same thing as last time. That last question is the valuable one, because the answer is usually yes and most players never notice.

The results screen helps here if the game has one worth reading. Ratios are more informative than totals: accuracy against throughput, collected against delivered, waves survived against enemies destroyed. Two runs with identical scores can represent completely different problems.

Five drills that work

Each isolates one thing, and each takes a few minutes.

• The no-attack run. In a combat game, survive without attacking at all. Builds telegraph recognition and the waiting habit faster than anything else.
• The fixed-gaze run. Choose a point ahead of your machine and refuse to move your eyes. Uncomfortable, and the fastest way to build the most valuable habit.
• The collect-nothing run. Take no pickups and make no detours. Isolates reading from greed, and the length usually reveals how much of a normal run is avoidable risk.
• The full-power run. Never let your resource meter reach zero. Nothing else changes. Most players attack less and last considerably longer.
• The name-the-opening run. Before each attack, consciously name which opening you are taking. Forces the light-versus-heavy distinction into awareness.

When to stop

Practice quality degrades with frustration considerably faster than it degrades with fatigue, and practising badly builds bad habits rather than none.

The signal is usually restarting without thinking. When you notice you have started three runs in a row without registering how the previous one ended, the session has stopped being practice.

Stop there. Arcade games are short enough that coming back in twenty minutes costs nothing, and the difference between a focused ten minutes and an unfocused hour is large and in the ten minutes' favour.

Why constraint runs work

The drills in this guide all take the same form - play normally but with one artificial restriction - and it is worth understanding why that shape is so effective.

A constraint removes decisions. In a normal run you are choosing between several things at every moment, and attention is split across all of them. Forbidding one option collapses that: with 'take no pickups' in force, the entire question of whether a pickup is worth it disappears and the freed attention goes somewhere useful.

It also makes the effect of that decision visible. If your runs get substantially longer when you stop collecting, you have learned something specific and quantified about how much your normal play costs you. That is far more convincing than being told, and it sticks.

And it produces a clean success condition. 'Complete a run without attacking' either happened or did not, which means a practice session has an outcome independent of score - and a session with an outcome feels finished in a way that a session of ordinary play often does not.

Practising when you cannot play

A small point, and more useful than it sounds.

A meaningful part of what makes you better at these games is not physical. Knowing that a heavy attack recovers in roughly a third of a second, that turrets are weaker up close, that the multiplier is worth more than any exchange - all of that is knowledge, and knowledge can be acquired away from the game.

This is why reading a guide with the game closed is not wasted, and why the specific numbers in these guides are quoted rather than approximated. 'Wait a moment longer' is advice; 'the recovery is 0.34 seconds and your light attack takes 0.16 to land' is a fact you can act on precisely.

What cannot be acquired this way is calibration - the felt sense of a window - and that is what the drills are for. The division is fairly clean: read to learn what to do, play to learn when.

Frequently asked questions