Your real typing speed is not your WPM. It is your WPM minus the time you spend fixing errors.
A typist at 70 words per minute with 92% accuracy is frequently slower in real work than one at 60 with 98%, because every error costs a backspace, a retype and a break in flow. This works out your effective speed and tells you which of the two is worth improving first.
A word, for typing measurement, is five characters including spaces — that convention is what makes WPM comparable between different passages. This charges keystroke-equivalents for every error: noticing it, moving to it, deleting and retyping.
That figure is an estimate, not a measured constant, and your real cost depends on how quickly you notice errors and whether you correct as you go. It is deliberately on the conservative side. The point of the number is the comparison between the two routes, which holds at any reasonable cost per error.
Below about 95% accuracy, correction time dominates your effective speed, and nothing else you do matters much. Above about 97%, corrections stop being the limiting factor and raw speed becomes worth chasing.
Most people trying to type faster are below 97% and attacking the wrong number. Running the figures usually makes that obvious immediately.
For accuracy: deliberately slow to the pace at which you are accurate, hold it until it feels automatic, and let speed return on its own. It always does, and the accuracy generally stays. The reverse — pushing speed and hoping accuracy follows — almost never works, because you practise the error at the same time.
Most people have two or three keys responsible for a disproportionate share of their errors. The Key Accuracy drill reports which are yours, and drilling those specifically is several times more productive than general practice.
For raw speed, the largest single factor is touch typing, and it is learnable at any age. If you look at your hands, that is the thing to fix before anything else.
Your gross words per minute minus the time spent correcting errors. It is the number that actually determines how long a document takes, which raw WPM does not.
Below about 97% accuracy, yes — correction time dominates. Above 97%, raw speed becomes the better thing to work on. Enter your own figures above and it will tell you which side you are on.
Above 97% on continuous prose is a reasonable working target. Below 95%, correction time swamps any speed gains.
Five characters including spaces. That standard is what makes WPM comparable across different passages, and it is what this calculator uses.
It is an estimate of keystroke-equivalents per error, covering noticing, deleting and retyping. It is not a measured constant, and it is deliberately conservative — the comparison it produces holds at any reasonable figure.