Prospira Tweaks
In August we read Polish forum threads about computers that are slow for no obvious reason, to see what the cause usually turns out to be. This page is that study done properly: every thread read in full, coded against the same rules, and listed at the end so anyone can check it.
What we found
The drive was the most common cause found, and four threads in ten never found one.
We read 72 threads from 10 Polish forums in which someone complained that a PC was slow, stuttered or froze, and coded how each one ended. 24 were fixed, and the person confirmed it, or the problem went away by itself. 6 more found the cause without fixing it. 31, 43 per cent, ended with nobody knowing why.
Where a cause was found, it was most often the drive: a dying disk, a drive that kept spinning down, games on a mechanical disk. Windows settings decided very few threads, and no program sold as an optimizer decided any.
How the threads ended
Strictly: a thread counts as fixed only when someone in it confirms the fix worked.
“Only partly better” means the person said things improved but the problem was still there. “Cause found, not fixed” covers a diagnosis everyone agreed on, such as a failing drive, with the replacement still to come.
What the cause turned out to be
Storage, by a wide margin. Everything else came up three times or fewer.
Everything else is one thread each: the display chain, the network, a background process, components reseated after a house move, hibernation switched off, the power plan, and two fixes that changed several things at once, so nobody can say which one helped.
What did not help
Reinstalling Windows, buying parts on a guess, and optimizer programs.
Reinstalling. In 28 threads, Windows had been reinstalled before the post. In 22 of them it was an attempt to fix the slowness, and it fixed none.
Buying a part first. In 13 threads someone replaced a part that turned out not to be the problem: a graphics card, a motherboard, a network card, a whole platform.
Optimizer programs. None of the fixes came from a cleaner, a debloat script or an optimizer. In one thread a program that promised to lower CPU use made the stutter worse.
The settings every guide starts with. The power plan, the page file and startup apps were part of the fix in 4 threads. In 1, switching to the High performance power plan alone ended a browser that lagged. In the others they were one item in a bundle of changes.
How we did it
Every thread read in full on 1 November 2026, and coded against written rules.
The threads were first collected on 30 August 2026 for our article on why a good PC is slow, from searches on Polish forums for a slow or stuttering computer with decent hardware. That first pass read summaries of the pages, not the posts themselves.
On 1 November 2026 we read every one again in full, every page including the last, and coded it against written rules. A thread counts if the person who opened it says the PC is slow, stutters or freezes. It counts as fixed only if someone in the thread confirms that a specific change made the problem go away. Each thread gets one main cause, and every quote we kept was checked against the page.
Of 80 threads we tried, 75 could be read. 2 were not about a slow PC, and 1 repeated another thread about the same computer, which leaves 72. The 5 we could not read were on a forum that was down for maintenance that day.
What this cannot tell you
It is what people wrote on forums, not a measurement of their PCs.
Polish forums only. 10 of them, so the sample follows their audiences.
Old and new threads mixed. They date from 2010 to 2026, and 25 are from 2023 or later. General slowness on older Windows dominates the older threads; stutter in games on current hardware, the newer ones.
Not all of it was good hardware. 39 of the PCs were current or decent for their time, 12 were weak or old, and for 21 the thread does not say.
Self-reported. A fix counts because the person said it worked, and some never came back to say. Threads that end well are also easier to find, so if anything the share with no answer is too low.
What changed from our first count
The first pass over-counted the fixes. This page replaces it.
In August we wrote that about sixty-five threads were read, that storage was the cause about thirteen times, that no thread was fixed by the power plan or the page file, and that in one thread a tweak was the whole cause. Read in full, the threads say less than the summaries did: 72 threads, storage 9 times, the power plan part of 4 fixes, and nothing in that one thread says who switched hibernation off. The article was corrected on 1 November 2026 to these numbers.
Every thread
Each one with its forum, the year it started, how it ended and the cause.
The same list with the hardware, the symptoms and what fixed each one: forum-study-2026.csv. Thread titles are as their authors wrote them.