Which Monkeytype Settings Actually Raise WPM?

Defaults are tuned for comfort

> Quick answer: The settings that actually change your improvement rate are test length, whether you punish errors, and word set difficulty. Longer tests build endurance rather than burst speed, stop-on-error forces accuracy at the cost of feeling slow, and the default 200 most common English words inflates your score against real writing. Quick restart is the setting most likely to be quietly harming you.

Monkeytype's defaults are chosen to make the site feel good on first use. Short test, common words, no punishment for mistakes, instant restart.

That's a sensible product decision and a poor training regime. Feeling good and improving are different objectives, and the default config optimises the first.

I've spent enough hours on it to have strong views about which toggles matter. Most of them don't. Three or four genuinely change what you're practising, and one is actively working against most people who use it.

Test length changes the skill

This is the biggest one and it's the one people fiddle with least.

The default is 15 or 30 seconds depending on when you started. At that length you're measuring burst speed. You can hold peak concentration for 15 seconds, you get an easy stretch of common words, and you post a number noticeably higher than you can sustain.

Move to 60 seconds and the number drops. Move to 120 and it drops again. That's not you getting worse, it's the test getting honest.

Which one you should use depends on what you're training.

Short tests, 15 to 30 seconds, are good for pushing your ceiling. You go faster than is comfortable, make mistakes, and gradually raise the top of your range. There's real value in that, and it's the most fun.

Long tests, 60 to 120 seconds, build the thing that actually matters at work: sustained speed without degradation. Almost nobody types in 15 second bursts in real life. If your 15 second number is 110 and your 2 minute number is 82, the 82 is the one describing your actual daily capability.

My suggestion is to alternate rather than pick. Short tests two days a week to push the ceiling, long tests the rest of the time to raise the floor. Tracking both separately is worth doing, and it stops the demoralising experience of watching your number fall when you switch formats. The test duration comparison goes into what each length actually measures in more depth.

Punishing errors, or not

Monkeytype gives you several ways to handle mistakes and they produce very different practice.

The default lets you type through errors, marking them red, and you can carry on or backspace. It's forgiving and it keeps the flow going.

Stop on error halts you at the mistake. You cannot continue until it's corrected. This is unpleasant and it works. It forces your hands to stop treating accuracy as optional, and it exposes exactly which letter combinations you fumble, because you sit there staring at them.

Confidence mode goes the other way and removes backspace entirely. You cannot fix anything. This trains commitment, and it's genuinely useful for people whose speed is destroyed by constant micro-corrections. Plenty of typists lose more time to unnecessary backspacing than to actual errors.

Which to use depends on your failure mode. If your accuracy is below about 95 percent, use stop on error for a couple of weeks. If your accuracy is fine but you backspace constantly out of anxiety, use confidence mode.

If both are fine, leave it on default and spend the effort elsewhere.

The reason this matters is that accuracy and speed aren't independent. Errors cost far more time than the keystroke you got wrong, because you have to notice, stop, correct and restart. That's the whole argument in why accuracy drops as speed increases, and it's why chasing a higher WPM with sloppy accuracy tends to plateau fast.

Word sets and the difficulty question

The default word set is the 200 most common English words. It is a very easy list.

Short words, familiar letter patterns, nothing that requires thought. Your score on it is genuinely higher than your score on real text, sometimes by 15 or 20 percent.

Switching to the 1000 word set immediately drops your number and immediately makes practice more useful, because you start meeting letter combinations you haven't automated. That's where the improvement lives.

Adding punctuation and numbers drops it further, and it's the single best change if you type real things for a living. Emails have commas. Code has brackets. Reports have figures. A number produced on lowercase common words with no punctuation describes a task nobody actually performs.

Quotes mode is the closest thing to real prose, since you're typing actual sentences with natural rhythm and punctuation. I use it more than word mode now, and my number there is roughly 12 percent below my clean word mode score, which I take to be the more honest figure. The official site at monkeytype.com has all of these under the test settings, and switching between them takes seconds.

The language setting is worth knowing about too, since it changes both word list and character set, and comparing your English number against your second language number is not a fair fight for reasons worth understanding separately.

One thing worth doing once: run the same length test on each of the four modes back to back on the same day. Common words, 1000 words, punctuation on, and quotes. Write the four numbers down. The spread between the easiest and hardest is your personal inflation factor, and knowing it means you can stop being confused about why your score moves so much between sessions. Mine ran about 20 points from top to bottom, which is a wider range than most of the day to day variation people worry about.

The setting I would turn off

Quick restart. Usually bound to tab or escape, it lets you abandon a test instantly and start a new one.

It sounds harmless. It is the single most common way people sabotage their own practice.

What happens is this: you start a test, fumble the third word, and reflexively restart. Then again. And again. Twenty minutes later you've done sixty attempts, forty of which lasted under five seconds, and you've practised the first three words of a test very thoroughly and nothing else.

Worse, you've trained a quitting reflex. Errors become a reason to abandon rather than to recover, and recovering from an error mid-flow is a real skill you now aren't building. In actual work you cannot restart the paragraph because you mistyped a word.

Turn quick restart off, or set it to require a deliberate action. Finish every test you start, even the bad ones, especially the bad ones. Your average will look worse and your improvement will be faster.

I'd pair that with not looking at the live WPM counter, which is the other attention thief. Watching the number while typing splits your focus and encourages you to chase the display rather than type well. Most people find their scores go up when they hide it, which tells you something about how much it was costing.

Then go race someone. Practice settings are worth tuning once and then leaving alone, and the fastest feedback on whether any of it worked comes from a live typing race against people who don't care about your config.