How to Read Your Latency Test Results (Colours & Verdicts)
Each destination tile shows your average, best and worst latency next to the world average, low and high for that same destination. Colour and verdict depend on how you compare with the world average, not on a fixed millisecond limit. Flags such as Down, Slow and Unstable add detail, and the summary cards give an overall picture.
On this page
Quick answer: look at the verdict at the top first, then the tiles. In each tile, the You line is your own measurement and the World line is what other visitors typically get for that same server. Colour tells you how you compare with the world; best and worst tell you how stable the connection is; flags call out problems. A mostly green page with a few yellow tiles is a healthy connection.
The layout of a results page
Whether you are looking at a fresh test or a saved report, the page has the same structure: a headline verdict, a row of summary cards, a comparison bar, and one tile per destination, grouped by category or region where useful. Everything on the page comes from your passes, plus the world statistics stored daily for each destination.
The summary cards
| Card | Meaning |
|---|---|
| Your average | The mean of every destination's average, using only destinations that replied |
| Your best pass | The single lowest reading in the whole test |
| Your worst pass | The single highest reading in the whole test |
| Reached | How many destinations replied out of how many were tested, for example 19/20 |
| You vs world avg | Your overall average compared with the world average of the same destinations, as a percentage. Negative means faster than typical, positive means slower |
| Unstable | How many destinations raised the Unstable flag |
The verdict
The verdict compares your overall average with the world average of the same destinations:
- Excellent: clearly faster than the typical visitor.
- Good: about the same as typical, or slightly better or worse within normal variation.
- A little slower than usual: noticeably above the world average but not dramatically.
- Laggy: clearly and significantly slower than the world average.
- No replies: nothing answered; usually a connection or blocking problem.
The verdict deliberately avoids fixed limits. A user in a remote region may have a higher average than a user in a data-centre city and still be "Good" because the comparison is against the same destinations. The method is explained fully in The world comparison explained.
Reading one destination tile
A tile shows the destination name, its category or region, a big latency number, a colour, and two meta lines:
- You: your average, best and worst for this destination.
- World: the world average, lowest and highest recorded for this destination.
It also shows the percentage difference against the world average. For example "+12%" means you are 12% slower than the world average, and "-30%" means you are 30% faster. If you chose many passes, you also see p95 and p99 for that destination.
How colours are decided
The tile colour uses the ratio your average ÷ world average:
| Ratio | Band | Colour |
|---|---|---|
| Up to 0.7 | Great (30% or more faster) | Green |
| Up to 1.15 | Good | Green |
| Up to 1.5 | Slightly slower | Yellow |
| Above 1.5 | Clearly slower | Red |
Two protections avoid false alarms: a tile is never red when the difference is under 30 ms, or when the value itself is under 50 ms. Going from 15 ms to 28 ms is nearly double, but nobody can feel it, so it is not shown as a problem. When a destination has no world data yet, fixed bands are used instead: under 50 ms great, under 120 good, under 250 yellow, above that red.
Flags: Down, Slow and Unstable
- Down / No reply: not one pass succeeded within the 10 second timeout.
- Slow: the average is above 2000 ms.
- Unstable: the worst pass is more than three times the average.
Flags are independent from colour. A tile can be green and Unstable at the same time: the average is fine, but there was one large spike. For real-time use that spike is exactly what you feel, so treat the flag seriously. See jitter explained.
Average, best and worst: what each tells you
- Best is closest to the physical minimum of the route (distance, hops, your access type).
- Average is the day-to-day expectation, but one spike can drag it up. See mean, median, min and max.
- Worst shows how bad it gets. If worst is far above average, check stability.
- p95 / p99 describe the slow tail more reliably than worst when you have many passes. See percentiles explained.
Worked examples
Example A: healthy connection
Your overall average is 48 ms, world average 52 ms (-8%). Reached 20/20. Unstable 0. Verdict: Good. Nothing to do.
Example B: one distant destination is red
All European and North American tiles are green, but an Asia-Pacific tile is red at 420 ms against a world average of 240 ms (+75%). The rest of your connection is fine; that route is slower for you than for most visitors. This is usually a routing or peering matter at your ISP, not a fault in your home. See routing, BGP and peering.
Example C: good average, unstable flag
Average 38 ms, but best 22 ms and worst 190 ms, Unstable on six destinations. Wi-Fi interference or a busy local network is likely. Try a cable and re-run. See troubleshooting high latency.
Example D: everything slow
Almost every tile is 80% above the world average. The common factor is your own connection: Wi-Fi distance, a VPN, a saturated upload or a mobile signal. Start with how to reduce latency.
Common mistakes and myths
- Treating red as an outage. Red is "slower than usual here", not "broken".
- Ignoring best vs worst. A fine average hides the spikes that ruin calls and games.
- Comparing different destinations' milliseconds. 20 ms to a nearby server and 180 ms to a far one can both be perfect.
- Expecting zero difference from the world average. The world average includes fast fibre users and slow mobile users; being within about 15% is typical.
What to do next
Understand where the world numbers come from in The world comparison explained, share a report from Saved reports and sharing, or if numbers look poor go to troubleshoot high latency. You can also run a new test.
Frequently asked questions
What do the colours mean?
Green means your average is at or near the typical visitor for that destination, or better. Yellow means slightly slower than usual. Red means clearly slower than the world average, more than 1.5 times it, and at least 30 ms slower. Colour is relative to the same destination, not to a fixed millisecond number.
Why is 120 ms green for one server and red for another?
Because each destination is compared with its own world average. A server on another continent normally answers in 150 ms or more, so 120 ms is excellent there. A server in your own country that normally answers in 20 ms would make 120 ms look very slow.
What is the difference between average, best and worst?
Average is the mean of all successful passes. Best is the lowest single pass, close to what the path can do at its quietest. Worst is the highest single pass, which exposes spikes. A large gap between best and worst indicates an unstable connection.
What does the Unstable flag mean?
A destination is flagged Unstable when its worst pass is more than three times its average. It means the connection occasionally spikes badly, which matters for calls, games and anything real-time even when the average looks fine.
What does Slow mean?
Slow is shown when the average for a destination is above 2000 ms (two seconds). That is slow enough to be noticeable for any use, and usually points to heavy congestion, a failing link or a distant, overloaded server.