Should You Fix Typos or Just Keep Typing?

Short Answer

> Quick answer: Fix it. A single typo left uncorrected costs you more than the four to six keystrokes it takes to repair, because scoring counts net words per minute rather than raw keystrokes, and real writing needs the text to be right. The exception is a typo you notice several words later, where the backspacing cost grows fast and a note to fix it at the end is cheaper.

The more useful point sits underneath the question. A high backspace rate isn't a habit problem, it's a signal that you're typing above the speed your accuracy supports.

What a Correction Actually Costs

Count the keystrokes properly and the decision gets easier.

A typo in the middle of a short word costs you the noticing, then a backspace for every character since the error, then retyping those characters. For a mistake three characters back that's six keystrokes to recover from one wrong one. The penalty scales with how late you catch it, not with how bad it was.

Catching it immediately is dramatically cheaper. One backspace, one correct character, done, and the whole thing costs about as much as two normal keystrokes. That's why fast typists correct instantly or not at all, and why the middle case of noticing five words later is the expensive one.

The rhythm cost is real too and it's harder to measure. Stopping to correct breaks the flow of a sentence you were partway through composing, and picking the thread back up takes a moment that no keystroke counter records.

None of that changes the answer for most situations, it just explains why the answer isn't unanimous.

How Scoring Treats Your Mistakes

Understanding this settles the argument for anyone who cares about their number.

Gross words per minute counts everything you typed. Net words per minute subtracts a penalty for errors, and net is what almost every serious test reports, because gross rewards typing quickly and badly. The definition of words per minute covers how the standard five character word and the error penalty combine.

Under net scoring, leaving an error in place is worse than spending keystrokes to fix it in nearly every case. A correction costs you a little raw speed while an uncorrected error costs you a full penalty, and the penalty is larger.

That's the arithmetic reason the answer is fix it. Our breakdown of how accuracy affects real WPM works through the numbers if you want to see the crossover point rather than take it on trust.

Watch out for tests that only report gross speed. They flatter sloppy typing and they'll teach you a habit that costs you the moment you do real work with it.

When Pushing Through Is Right

Three situations genuinely favour leaving the mistake alone.

The first is a typo you spot several words later. Backspacing back to it destroys correct text you'd then have to retype, and the cost climbs with every word. Finishing the sentence and fixing it afterwards is cheaper, assuming you remember.

Second is drafting, where the text is going to be edited anyway. Stopping to fix spelling while you're trying to hold an argument together is a poor trade, and the editing pass will catch it at almost no marginal cost.

Third is anywhere a spell checker will handle it. A single wrong character in an ordinary word gets flagged and fixed later in bulk, so interrupting yourself for it buys nothing.

What all three have in common is that the correction is deferred rather than skipped. Genuinely leaving errors in finished text isn't a speed strategy, it's just producing worse work faster, and nobody who types for a living treats it as an option.

Your Backspace Rate Is a Symptom

This is the part worth taking away, and it reframes the whole question.

If you're backspacing constantly, the problem isn't your correction policy. You're typing faster than your accuracy currently supports, and every correction is the system telling you so. Changing whether you fix the errors doesn't address the rate at which you're producing them.

Accuracy tends to fall away sharply past a personal threshold, and that threshold moves with practice rather than with effort. Pushing past it produces a burst of speed followed by a pile of corrections that costs more than the burst gained, which is why our piece on why accuracy drops as speed increases matters more here than any rule about backspacing.

The fix is uncomfortable and it works. Slow down until your error rate is very low, hold that pace until it feels easy, then let the speed come back on its own. Speed built on top of accuracy is stable, and speed built by pushing is not.

I tried the opposite approach for months, chasing a higher number and correcting furiously, and my net score barely moved the whole time. Dropping ten words per minute deliberately for two weeks moved it more than any amount of pushing had.

Where the Errors Actually Come From

Before deciding what to do about mistakes it helps to know which ones you make, because they are not evenly spread.

Most typists have a handful of letter pairs that go wrong far more than the rest. Combinations where the same hand has to move twice in a row are common offenders, as are pairs where a weaker finger follows a stronger one and arrives slightly late. The result is transposed letters rather than wrong ones, which is why so many typos are the right characters in the wrong order.

Rare words cause a second cluster. Familiar words are typed as single motor patterns, while an unfamiliar word gets assembled letter by letter, and the switch between those two modes is where the rhythm breaks. Technical writing and names produce far more corrections than ordinary prose for exactly this reason.

Then there is the tail end of a long session, where error rate climbs without speed dropping much. I logged my own tests across a working day once and the afternoon runs had noticeably worse accuracy at the same speed, which is the clearest argument for treating fatigue as a real variable rather than an excuse.

Knowing which of the three is your main source tells you what to fix. Problem pairs respond to drills, rare words respond to slowing down deliberately on them, and fatigue responds to breaks rather than to any amount of practice.

Tracking It Properly

Most people never look at their correction rate at all, which is odd given how much of their typing time it accounts for.

Watch your accuracy percentage as a leading indicator rather than your speed. If accuracy sits below the mid nineties in ordinary typing, corrections are eating meaningfully into your output and that's the number to move first.

Run a test where you deliberately don't correct anything and compare the gross and net figures. The gap between them is the cost of your error rate expressed in words per minute, and seeing it as a single number tends to be more persuasive than any argument about habits.

Repeat the comparison every few weeks rather than daily. Error rate is noisy from session to session and a single bad run means nothing, while a trend over a month is real. Our typing test keeps the history so the trend is visible without you recording anything.

Then target the specific errors rather than errors in general. Most people's mistakes cluster on a small number of letter combinations, and drilling those clears a disproportionate share of the problem.

Building the Better Habit

Two changes cover nearly all of the improvement available here.

Correct immediately or not at all. The middle ground of noticing something four words back and going after it is where the real time goes, and simply deciding in advance that late errors get fixed at the end removes the worst case entirely.

Then work on producing fewer errors instead of managing them better. Targeted drills on the combinations you actually get wrong do more than general practice, and our accuracy drills guide covers how to find your own problem pairs rather than guessing at them.

Give either change a fortnight before judging it. Correction habits are automatic by the time you're any good, so changing one feels awkward for longer than you'd expect, and the awkward phase is not evidence that it isn't working.

The typists with the highest sustained output almost never look like they're trying hard. They're not correcting much because they're not making many mistakes, and that ordering is the whole lesson.