Frugal vs the closest competitor
Waiting
The numbers, with local flair
Local South America Usenet Latency
Six providers. One region. Let the numbers talk.
Loading dataChecking stored results
Published by Frugal Usenet using public RIPE Atlas measurements. Provider names are used for comparison only
Loading the reporting window…
Loading stored measurement results…
Each region uses its own probe group and test start date. Results are not combined into a worldwide ranking. All displayed periods roll forward with each stored update.
The result, minus the homework.
Checking the regional results…
Frugal’s rankN/Aout of six providers
Provider comparison
Typical response time
Lower numbers mean less delay.
How long a reply takes to get back. Measured in milliseconds (ms).
ProviderTypical response timeExtra delay vs #1
More detail p95 and test counts
“Typical response time” is the median: the middle of the comparable readings. P95 shows the slower end: about 95% of those readings were at or below this value.
Where Frugal had the lowest delay
WaitingFrugal vs each competitor
Frugal’s Local Usenet Advantage
Each comparison uses readings from the same test devices taken close together. “Lower” means less delay for Frugal.
Visualize The Wait
Same Tests, very different wait times
Animation slowed 30×. Same multiplier for every provider.
A simulation using this period’s stored median response times. Every provider starts together. The numbers show actual measured milliseconds. This illustrates response delay, not download speed.
Every square has a story
Who won at each test location?
One square per reporting test device. Its color and letters show the provider with the lowest median delay. This scoreboard uses the whole region for your selected period.
Hourly trend
Response time over time
Lower lines mean less delay. Tap a provider name to show or hide its line.
Your part of the world
Results by location and internet provider
Each row is a test device. Response times are in milliseconds (ms); lower means less delay. “N/A” means there are not enough comparable readings from all six providers yet.
What does “hourly comparisons” mean?
“Based on 23 hourly comparisons” means that device had 23 usable sets of readings from all six providers during the selected period. More comparisons mean more readings behind that result. Each competitor reading must be within 30 minutes of Frugal’s reading on the same device.
Filters apply to this table and its download.
Test settings and coverage
Recorded settings for each provider. A valid observation contains at least one successful ping; missing replies are not treated as zero latency.
| Provider | Assigned probes | Reporting probes | Valid observations | Packets | Packet interval | Run interval | Spread | Latest valid result |
|---|
How we test The nerdy bits. You asked.
Measurements
Public IPv4 ICMP ping tests run hourly. Packet count, interval, spread and current probe assignments are shown above. Each observation is the median of successful packet round-trip times. If multiple results occupy the same rounded UTC hour, the latest is retained.
Like-for-like statistics
Provider median, p95, trend lines and probe winners use complete observations shared by all six providers. Competitor observations are matched one-to-one to the nearest Frugal observation from the same probe within 30 minutes. They are close in time, not simultaneous. Both timestamps must be inside the selected period. “Extra delay vs #1” is the difference between unrounded provider medians on this shared set; it is separate from the paired differences below.
Paired differences
For each pair, the difference is competitor RTT minus Frugal RTT. The percentage is that difference divided by competitor RTT, times 100. We report medians of those differences and percentages separately. A positive value means lower latency for Frugal. Each two-provider comparison may have more observations than the six-provider ranking.
Probe wins & percentiles
The winner is the provider with the lowest unrounded median on each comparable probe. Ties within 0.000001 ms are reported separately and are not sole wins. P95 uses linear interpolation at the 95th percentile of observation medians, not individual packets. Unequal reporting time can give some probes more weight.
Coverage & periods
The six-provider comparison starts when Usenet.Farm’s measurement began in this region. Earlier readings remain stored, but are outside this comparison. The 24-hour, 7-day, 30-day and entire-test views use only the available shared history. Newly added probes appear once results arrive; earlier periods do not gain invented history. Missing results can reduce comparison coverage. Latency excludes timeouts and does not measure availability or throughput. These selected probes are not a representative sample of every home connection in a region.
Daily winners & replay
The daily strip always covers the last seven completed UTC days, independent of the selected period. We match readings afresh within each day, then rank the providers by their unrounded medians. A partial first test day is labeled; days before the shared test began and days without complete comparisons have no winner. Days can have missing hours. The replay uses the selected period’s medians. It starts at 30× slower; 10× and 100× are optional. The selected time multiplier and travel distance are identical for every provider. Providers tied for first use the same animation character. The animation illustrates response delay, not download speed.