Download Time Calculator

How download time is actually calculated

Every download estimate is one division: file size divided by transfer speed. The trap is that the two numbers almost never arrive in the same unit. Your internet plan is sold in megabits per second, your file manager reports gigabytes, and the two differ by a factor of eight before anything else goes wrong. Get the conversion right and the arithmetic is trivial; get it wrong and your estimate is off by 800%.

Worked example: a 50 GB game on a 100 Mbps line

Start with the connection. 100 Mbps means 100 million bits per second. Eight bits make one byte, so divide by eight: 12.5 megabytes per second. That is the theoretical ceiling of the line, and it is the number you should compare against the "12.4 MB/s" your browser shows during a download.

Now the file. 50 GB is 50,000 MB. At the theoretical 12.5 MB/s the download takes 50,000 ÷ 12.5 = 4,000 seconds, which is 1 hour 6 minutes 40 seconds. Switch to the realistic setting and 10% comes off the top: 12.5 × 0.9 = 11.25 MB/s, so 50,000 ÷ 11.25 = 4,444 seconds, or about 1 hour 14 minutes. That five-to-eight minute gap is the protocol tax, and on a big file it is real time you will wait.

The eight-times confusion, and the bill it produces

A customer upgrades from 100 Mbps to 500 Mbps and complains the download manager only shows 56 MB/s. Support explains, again, that 500 Mbps is 62.5 MB/s at best, and after overhead about 56 — the line is performing exactly as sold. Nothing is broken. The same confusion runs the other way: people budget "1 GB per second" for a gigabit line when the honest figure is 125 MB/s before overhead, roughly 112 MB/s after.

A quick rule for anything a UK or US ISP advertises: drop the "b", divide by eight. 25 Mbps is about 3 MB/s, 100 Mbps about 12.5 MB/s, 500 Mbps about 62.5 MB/s, 1 Gbps about 125 MB/s. Lowercase b means bits, uppercase B means bytes — the whole industry hangs on one letter of case.

Why 10% overhead is the honest default

TCP/IP headers, TLS handshakes, DNS lookups, retransmitted packets and acknowledgement traffic all ride the same pipe as your file. Measured overhead on a typical HTTPS transfer sits between 5% and 10%; the calculator uses the pessimistic end because most real sessions also lose a slice to something else. Wi-Fi is the usual suspect: a 2.4 GHz link in a busy flat rarely sustains 100 Mbps regardless of what fibre reaches the router. Then there is the far end, where the server may cap each connection to spread load fairly — the reason a torrent with 30 peers can outrun a single HTTPS stream on the same line.

The same file across speed tiers

The comparison table in the tool is where the upgrade question gets settled. A 100 GB console game at 10% overhead takes roughly 9 hours 53 minutes on 25 Mbps, 2 hours 28 minutes on 100 Mbps, 30 minutes on 500 Mbps and 15 minutes on 1 Gbps. Notice the shape of the curve: the jump from 25 to 100 Mbps saves seven and a half hours, while the jump from 500 Mbps to gigabit saves fifteen minutes. Most households feel the first upgrade dramatically and the last one barely at all, which is worth knowing before signing a two-year contract.

The same shape explains why streaming tiers stopped mattering. 4K video needs about 25 Mbps sustained; anything above that changes nothing about picture quality and only shortens large downloads. If nobody in the house downloads 100 GB games, the practical benefit of gigabit is mostly latency and simultaneous users, not the headline number.

Limitations worth knowing

This calculator uses decimal units — 1 GB = 1,000 MB — which matches how drives and ISPs are sold. Windows reports the same file as 46.6 GiB using binary units, so a size copied from Explorer runs about 7% high here. Steam, PlayStation and Xbox all display decimal sizes, so pasting a store page number needs no adjustment.

The estimate also assumes a steady rate. Real transfers ramp up over the first few seconds (TCP slow start), dip when a shared network gets busy, and sometimes stall while a server thinks. Treat the result as the best case with a realistic haircut applied: if your actual download finishes within about 20% of it, your connection is behaving normally. If it lands at one eighth of the predicted speed, you converted bits to bytes in the wrong direction.