Every computer has a bottleneck. One component always runs out of capacity before the others — that is what it means for a system to have a maximum speed. A build with no bottleneck is not a well-balanced build; it is a build that is not doing anything.
So “do I have a bottleneck” is not a question with a meaningful answer. The questions worth asking are which component is the limit, at what frame rate it bites, and whether you would notice if it were gone.
The percentage problem
Most bottleneck calculators hand you a number, often with a colour. Ours does too. It is the least useful thing on the page, for three reasons.
It has no standard definition. There is no agreed formula for “bottleneck percentage”. Different sites compute different quantities and none publish the arithmetic, so a 23% from one calculator and a 23% from another are not the same claim and cannot be compared. The number looks precise, which is exactly what makes it misleading.
It is scale-blind. A 20% imbalance at 160 fps costs you about 32 frames you were never going to see on a 144 Hz monitor. The same 20% at 38 fps is the gap between playable and not. Identical percentage, completely different problem.
It implies an action it cannot justify. A percentage with a red bar next to it reads like a fault requiring a purchase. Often the honest advice is to change a setting, or to do nothing.
When it genuinely matters
There are real cases, and they share a shape: the limit lands below something you can perceive.
You are below your monitor’s refresh rate and want to be above it. This is the only common case where a bottleneck has an obvious cost. If you have a 144 Hz display and a competitive shooter runs at 90, the gap is visible and worth money.
You are below roughly 60 fps. Below that, frame-to-frame delays become perceptible as stutter rather than merely as lower smoothness. Getting a build from 45 to 70 is a bigger subjective improvement than getting one from 120 to 190.
The limit is the processor and you intend to upgrade your card. This is the forward-looking case. A graphics upgrade into a processor-limited system buys little, because the new card inherits the old ceiling. Worth knowing before you spend.
Frame times are inconsistent. A build that averages 90 fps but lurches is worse to play than one that holds a steady 70. Averages hide this; 1% low figures reveal it. Memory running below its rated speed and running short of video memory both show up here rather than in the average.
When it does not matter at all
You are already above your refresh rate. If your monitor is 144 Hz and you get 200 fps, the extra frames are discarded. A bottleneck that caps you at 180 instead of 240 costs you nothing you can see.
You play at 1440p or 4K. Higher resolutions shift work onto the graphics card and away from the processor. A pairing that looks badly mismatched at 1080p often becomes well balanced at 1440p without a single component changing. The same hardware, a different answer — which is why any bottleneck verdict that does not name a resolution is incomplete.
The mismatch is deliberate. Pairing a strong processor with a modest card is the sensible way to build if you plan to upgrade the card later. A calculator will call that a bottleneck. It is actually a plan.
You play games that do not stress the imbalance. Esports titles are light on the graphics card and heavy on the processor; large open-world games are the reverse. A pairing can be badly matched for one and fine for the other, and only your library decides which applies.
The question to replace it with
Instead of “how big is my bottleneck”, ask:
At the resolution I actually play, what frame rate do I get, and what is the cheapest change that moves it?
That has an answer you can act on. Sometimes it is a part. More often it is a setting — and it is worth knowing that dropping one quality preset step typically buys 20–35% on a graphics-limited build, which is more than most single-component upgrades and costs nothing.
Why the industry pushes the number
Worth saying plainly: a percentage with a red bar is good for business. It converts an abstract question into an apparent problem, and problems sell parts. A lot of bottleneck calculators exist to point you at an affiliate link, and a reassuring result does not do that.
We publish a percentage because people look for one and its absence would be its own kind of unhelpful. But the figures next to it — frame rate at each resolution, which side is the limit, what a specific change would buy — are the part worth reading, and our calculator leads with those deliberately.
If the answer for your build turns out to be “nothing worth buying”, that is a real result and you have saved money. That is a better outcome than a number that makes you feel you ought to spend.