← osu! guides

Reading your osu! top 100: what your pp curve says about you

An osu! top 100 is the ranked list of a player's 100 best-performing scores, and plotting pp value against rank produces a curve whose shape carries information the total pp figure hides. A steep curve means the profile leans on a few standout scores; a flat curve means broad consistency with no breakthrough; a single score far above the rest usually marks a one-off rather than a repeatable standard. Because profile pp weights the n-th score by 0.95 to the power of n minus one, two players with identical pp totals can hold very different curves, and each shape points to a different thing to work on next.

What the curve is

Your osu! profile lists your best-performing scores in descending pp order, capped at 100. Plot pp value on the vertical axis against rank 1 to 100 on the horizontal axis and you get a curve that always falls from left to right. What varies between players is how fast it falls, and that rate of fall is the diagnostic. Two players can share an identical profile pp figure while one holds a 400 pp best score above a thin tail and the other holds a hundred scores clustered near 100 pp each.

The shape matters because of how profile pp is assembled. Each score is weighted 0.95ⁿ⁻¹ by its rank, so a drop across the first ten places costs far more total pp than the same drop between ranks 60 and 70. Reading the curve is really reading where your pp is concentrated, and concentration determines which kind of play will move your profile next.

Where the weight actually sits

Before diagnosing any shape it helps to know how much of the total each region of the list can contribute. On a hypothetical top 100 of identical scores, the weights distribute like this:

Nearly three quarters of a profile lives in the top 25, and the entire bottom half is worth less than the best two scores combined. That frames everything below: a shape problem in ranks 1 to 25 is a real problem, while a gap in ranks 70 to 100 is almost never worth targeting directly. It also explains why grinding scores that land around rank 90 feels unrewarding — because it is.

The steep curve: a top-heavy profile

A steep curve drops sharply over the first ten or twenty ranks and then flattens into a long low tail. Typically the best score is two or three times the value of the twentieth. It usually means one of two things: a recent genuine jump in skill that the rest of the list has not caught up to, or a handful of scores set on maps that reward this particular player disproportionately.

The diagnosis is that your ceiling is ahead of your floor. Your best score proves you can do something the rest of your list does not reflect, so the fastest available pp is not another record — it is repeating something close to your current top ten across more maps. Every score converted from the low tail into the top 25 region moves a large weight. In practice: return to the difficulty band your best top plays sit in and set several more scores there before chasing a new personal ceiling.

The flat curve: consistent but capped

A flat curve falls gently — the best score might be only 30% to 50% above the fiftieth. This is the profile of a consistent player who reliably converts maps inside a comfortable band and rarely plays outside it. Accuracy is usually good and misses are rare.

The weakness is that consistency alone has a hard ceiling. A new best score adds roughly its own value minus 5% of your weighted total, so on a flat profile almost every score you set lands near break-even: you are running to stand still. Once the list is full, filling it with more scores at your existing level barely registers.

The target here is difficulty rather than repetition. Something has to be worth meaningfully more than your current best — a higher star rating, a mod you do not normally use, or a longer map that accumulates more difficulty over its length. Expect the first attempts to score below break-even; that is the price of moving the band.

The single outlier

An outlier profile has one score standing far above everything else — sometimes double the second-place value — with an otherwise ordinary curve beneath it. It is the shape produced by a lucky run: a map that suited the player exactly, or an unusually clean attempt that has never been repeated.

Two things follow. That score is doing real work, since rank 1 carries full weight and the play is counted at 100% of its value. But it is also the least useful score in the list going forward, because improving on your best now means beating a number you have never reproduced. An outlier raises your pp once and raises your bar permanently.

The right response is to treat it as evidence rather than as a standard. Work out what made it possible — a pattern type, a BPM range, a map length, a mod — then look for the second and third instances of that. Turning one outlier into a cluster is what converts a fluke into a level.

Concentration in one mod

Scan the mod tags down your top 100 and count them. A list that is 90% no-mod, 90% Hidden or 90% Double Time tells you where your pp comes from and, by omission, where it does not.

Each mod loads a different part of the difficulty calculation. Double Time raises the rate at which objects arrive, pushing the speed and strain components; Hard Rock tightens circle size and approach and mostly loads aim; Hidden and Flashlight change what you can see rather than what you must hit, and are also the mods that produce silver S and SS grades instead of gold ones (along with Fade In in osu!mania). Easy and Half Time reduce a map's rating and therefore the pp a clean play returns. Exact mod multipliers are revised from time to time, so treat the direction as stable and the magnitude as current.

Heavy concentration is not automatically a fault; specialising is a legitimate route. It becomes a problem when your chosen mod has run out of headroom in your map pool, at which point the cheapest untapped pp is usually the adjacent mod rather than a harder map with the same one.

Concentration in one kind of map

Mods are the easy axis to check; map type is the more informative one. Look down your top 100 for star rating, BPM, length and pattern character. Four patterns come up repeatedly:

• A narrow star band. Nearly every score within half a star of the others. The list is a snapshot of one skill level and will only move when that level moves.
• Short maps only. Length contributes to difficulty, so a top 100 of 90-second maps has a lower achievable ceiling than one that includes long maps. Endurance is a trainable weakness.
• One pattern archetype. All streams, or all jumps. Highly efficient right up until the suitable maps run out.
• Repeated beatmapsets. The same few sets appearing many times means you are refining rather than expanding.

None of these is wrong, but each predicts a different plateau: a narrow star band plateaus on skill, a short-map list plateaus on stamina, and a single archetype plateaus on supply.

Accuracy across the list

Profile accuracy is a weighted average of your best-performance scores using the same 0.95ⁿ⁻¹ curve as pp, so it is dominated by the same handful of top scores. Reading accuracy down the list separates two very different players who may share a pp total.

A top 100 of high accuracy on modest star ratings says your execution is clean and difficulty is the limiter, so raising star rating should convert fairly directly into pp. A top 100 of low accuracy on high star ratings says the opposite: you are playing above your control, and because the pp algorithm weights accuracy heavily, cleaning up maps you already play is usually the larger and faster gain.

One caveat when comparing profiles or your own history. osu!(lazer) counts slider tails as judgements contributing to accuracy, while osu!(stable) does not, so the same play can report a noticeably different accuracy figure on each client. Compare like with like before drawing conclusions.

Shape to action, summarised

Collecting the diagnoses into one place:

Treat these as starting hypotheses rather than verdicts. The reliable test of any of them is whether the change produces scores that enter your top 25, because that is where nearly three quarters of the weight sits.

Watching the shape change over time

A curve read once is a diagnosis; a curve read repeatedly is feedback. After changing your approach, the useful question is not whether pp rose but where the new scores landed. Scores entering around rank 80 confirm you are still working inside your existing band. Scores entering the top 20 confirm the band itself has moved.

TrackMyOsu records exactly that. Each session snapshots your full top 100 at the start, refreshes every 60 seconds while you play, flags new top plays and new medals as they land, and saves the before-and-after comparison permanently, so you can see which score entered and which was pushed out. Public profiles at /u/<username> make the same view available for other players, which is a practical way to see what a top 100 looks like a thousand pp above your own.

By carrot12_. Last updated 2026-08-12.