See exactly which keys slow you down

Key accuracy

Correct presses divided by every attempt at that key. An attempt is either pressing the key — whether it was the right character or not — or being the key the passage expected while a different key was pressed. Both kinds of mistake land on the key you were meant to hit, which is why the map shows where your hands actually go rather than only which characters came out wrong.

Every press counts

A mistake you fixed with backspace still counts against the key. Correcting it cost you the time to notice, delete and retype, so hiding it would flatter the number. This is stricter than the character accuracy in the results panel, and deliberately so.

Shifted and unshifted forms are one key

Upper case letters and symbols that need shift are folded into the key underneath them. Pressing shift when the text wanted a lower case letter is a wrong press of that letter, not a separate key.

The bands

Green from 95% up, amber from 85% to 94.9%, red below 85%, grey for a key that has not come up yet. The thresholds are generous because a key pressed twice in a short test is not evidence of anything.

The layout

The board is a US QWERTY keyboard with the standard touch typing finger assignments. A few marks in the punctuation drill are not on that layout; they still count towards your accuracy, they are simply not drawn.

Start at the top of the keys to practise table. A key in the red band is not a key you are slow at — it is a key you are guessing at. Typing it faster will not help; finding it without looking will. Take the drill beside it, type it slowly enough to be right every time, and come back to the same test in a few days.

Then look at the finger, not only the key. Several weak keys that share a finger usually mean the hand is not returning to the home row between words. That is one habit to fix, not five. Thehome row lessonis the place to rebuild it, because every other key is reached from there.

Keep the tests comparable. A fifteen second test only touches a slice of the board, so a key that looks terrible after one short run may simply have come up four times. Two or three tests of a minute or longer give a picture stable enough to act on. The recent tests table above shows what went into the totals.

Finally, re-test after practising rather than during. Accuracy on a drilled key improves immediately and disappears again by the next day if the habit has not set. A weekly check is a better measure of whether the drill worked than a check five minutes later.