Combo Ladder
Every correct character extends a running combo count; a single mistake drops it back to zero, so the final impression comes almost entirely from your one longest unbroken stretch rather than from your overall error rate.
QWERTY layout assumed. Backspace corrects; uncorrected errors count against net WPM.
Combo Ladder runs on a hundred-and-twenty-word generated passage, typed under the site's standard rules — a wrong character still advances the caret and backspace still corrects it, exactly like an ordinary test. What's different is what's tracked alongside that: a live combo counter that climbs by one for every correct character in a row and drops straight back to zero the instant one is wrong, plus a running record of the longest stretch you've managed and how many times the combo has broken.
Two typists can finish this passage with the identical number of total errors and still come away with very different combo numbers. Five mistakes spread evenly through the passage break the chain roughly every hundred characters or so; five mistakes clustered in the first minute leave a single uninterrupted run for everything after — same error count, a far longer 'Longest' figure for the second typist. The format is built to surface that difference, which a flat accuracy percentage averages away completely.
The progress bar under the combo counter tracks your current combo against the passage's full length in characters, so it fills in step with how much of the actual text your current unbroken streak covers. A full bar means exactly what it looks like: your combo has run the entire passage, first character to last, without a single break along the way.
How to Play
Type the passage as normal. Combo, Longest, and Breaks update live in the HUD above the passage, and the bar beneath them tracks your current combo as a share of the whole passage — watching it climb toward full is watching how much of the text you've covered without a single mistake.
Because the score here rewards the length of your best stretch far more than it punishes your total error count, the most valuable thing you can do partway through a long combo is recognise a word or letter pair that's likely to break it — a double letter, an awkward reach — and consciously ease into that specific spot rather than typing through it at full pace and risking the run you've already built.
Combo itself resets to whatever it happened to be the instant you stop, and isn't worth carrying forward — but Longest and Breaks are exactly what the results screen adds once you finish, sitting next to net WPM, gross WPM and accuracy instead of the live combo count that got you there.
What a Full Bar Actually Means
The bar is scaled against the passage's real length in characters, not its word count, so there is no shortcut version of a clean run. Reaching full means your current streak has covered the entire text, start to finish, without one character breaking it along the way — a literal readout of how much of the passage you've covered mistake-free right now, not an approximation that maxes out early.
Same Error Count, Different Story
A flat accuracy percentage treats every error identically and spreads the damage evenly across the whole passage in the number it reports. Combo Ladder does the opposite: it keeps the worst gap in your typing — how long you went before the first break, and the longest stretch after it — sitting right next to the total, so two runs with matching accuracy can still look nothing alike once you compare their Longest figures.
What Survives to the Results Screen and What Doesn't
Combo, the live number climbing while you type, is only ever a mid-run reading — whatever it shows the instant you stop rarely means anything on its own, so it isn't carried anywhere. Longest and Breaks are different: both land on the results screen once you finish, sitting next to the ordinary net WPM and accuracy figures every page reports. The number this whole format is actually built around, your single best stretch, is exactly the one that makes it to the final screen.
A Different Kind of Rhythm Problem Than a Steady Pace
The Steady Rhythm Drill measures how even your timing is across a whole passage; Combo Ladder doesn't care about timing at all, only about whether a character was right or wrong. A typist with perfectly even, unhurried timing and one bad habit that trips them up every few dozen characters will still see their combo reset constantly here, even though their rhythm, by the drill's own measure — or by the Typing Rhythm: Even Pacing vs. Bursts lesson's — would look fine.
Frequently Asked Questions
What does a full progress bar actually mean here?
That your current combo spans the entire passage, character for character, with nothing breaking it anywhere along the way. The bar is scaled against the passage's full length, so there is no way to fill it without a completely clean run from the first character to the last.
Does a mistake here end the run, the way it does in One Life?
No. The combo resets to zero and the break count goes up by one, but you keep typing normally afterward — backspace still works and the caret still advances, exactly as it does on a standard test. One Life is the format where a single mistake ends the attempt outright; this one just costs you your current streak.
Can I see my longest combo again after I finish the passage?
Yes — Longest run is one of the figures the results screen adds once you finish, reported in characters alongside your total Breaks and the site's ordinary net WPM and accuracy. It's the live Combo count, the number that resets constantly while you're playing, that doesn't carry over; Longest is the one built to survive.
Is this measuring the same thing as the Steady Rhythm Drill?
No. Steady Rhythm is about even timing between keystrokes regardless of correctness. Combo Ladder ignores timing entirely and only tracks whether each character was right or wrong in sequence, so a typist with very even timing but one recurring slip can still see their combo break constantly here.
Can two typists with the same number of total errors get very different scores?
Yes, and this is really the point of the format. Errors clustered together leave a long stretch untouched afterward; the same number of errors spread evenly breaks the combo repeatedly, producing a much lower Longest figure even with identical accuracy overall.