Packet Loss Impact Calculator
Turn a packet-loss percentage into real effects: expected packets lost, delivered goodput and the chance a whole multi-packet message arrives intact.
How it works
Goodput = 1 − p · Message intact = (1 − p)^(packets per message)
Loss looks small per packet but compounds fast across a message. If each packet independently survives with probability 1 − p, then all N packets of a message arrive only with probability (1 − p) raised to the N — so even 1% loss makes a 20-packet frame fail about 1 time in 5. Because a single missing packet forces the whole message to wait for a retransmit, loss hurts latency and throughput far more than its headline percentage suggests, which is why real-time media hates it.
Worked example
At 1% loss (p = 0.01) over 1000 packets you expect 10 losses. A 20-packet message survives with probability 0.99^20 = 0.818, so 18.2% of such messages lose at least one packet and must be resent — a big deal for latency-sensitive traffic.
Frequently asked questions
Is 1% packet loss bad?
For bulk downloads TCP absorbs it with retransmits, but for real-time voice, video and gaming even 1% is noticeable, and 2–3% becomes clearly disruptive because every lost packet in a stream shows up as a glitch or stall.
Why does loss hit throughput so hard?
TCP treats loss as congestion and cuts its sending window, and any lost packet blocks in-order delivery until it is resent. So loss both shrinks the window and adds round-trip delays, dropping effective throughput well below the raw link rate.
Related calculators
This calculator is for educational purposes. Double-check important results before acting on them.