Map

Train punctuality in Europe and beyond

We measure arrivals at the destination from open data. Every figure comes with its threshold, its coverage (how many of the scheduled trains we could observe) and its confidence interval. When two countries cannot be told apart statistically, we say so.

Data: 21 Nov 2026, 02:13 CET

One coverage rule across the page. A train "has data" when its realtime feed reports a valid delay (between −60 and +120 min). Punctuality is computed over those trains only: trains without data never count as on time. A country enters a ranking when at least 30% of its trains have data and there are 20 or more; a station, with 10 or more. On time = under 3 min late.

Monthly punctuality report → · Operator and line rankings, month by month

Right now 21 Nov 2026, 02:13 CET Snapshot; biased towards long and delayed trains
US412 trains running +1.2 min current median delay
DE405 trains running 0.0 min current median delay
AU374 trains running 0.0 min current median delay
JP220 trains running 0.0 min current median delay
MY163 trains running no median few trains with data
GB114 trains running +1.0 min current median delay
CH71 trains running 0.0 min current median delay
RO51 trains running no median few trains with data
CA47 trains running no median meaning of 0 unknown
DK42 trains running no median few trains with data
PL40 trains running 0.0 min current median delay
NL25 trains running 0.0 min current median delay

Trains running: those with a position or prediction right now. Current median delay: median of the last known delay of the trains with data. Not comparable with punctuality at destination.

Right now

A snapshot of the trains running at this moment; it does not depend on the selected period. Trains that spend longer running (long journeys, or delayed ones) weigh more in the snapshot than over the whole day: this is length bias. To compare periods, use the trend and the historical punctuality.

Current punctuality (<3 min)

Current punctuality Fair
78%
based on 1,522 of 2,096 trains with RT data
95 % CI: 76.2–80.3 %
Punctual = less than 3 min late. The 574 trains without data do not count.
Average delay
3.3min
mean over the 1,522 trains with data
Includes on-time (0) and early (negative) trains. It is the only definition of “average delay” on this page.
% of trains delayed
21.7%
330 of 1,522 trains with data are 3 min or more late
The complement of punctuality: 78 % + 22 % = 100 % of trains with data.
Average delay of delayed trains
15.6min
mean over the 330 trains 3 min or more late only
Measures severity when there is a delay. Not comparable with the overall “average delay”. Maximum now: 108.0 min.

Source of these cards and of the «right now» league: positions at this moment (current_positions), with the last delay of each visible train. It is not the punctuality at destination of the «over scheduled trains» table below, which may give another figure.

Closed period · from run_facts

Punctuality at destination

Last 30 days: 13 Nov to 19 Nov · only 7 days computed
Threshold
common: 5 min
Punctuality at destination < 5 min 80.7% 95 % CI 80.5–80.9 % · 179,992 arrivals in 4 countries threshold 5 minleague countries only
Cancelled and partial 2.4% 95 % CI 2.4–2.5 % · 4,889 cancelled + 19,573 partial of 999,567 scheduledonly where the feed reports them; not part of punctuality
Coverage of scheduled trains 63% 95 % CI 62.5–62.7 % · 999,567 scheduled with data / scheduledmin. 30 % to publish
Extreme delay (≥ 2 h) 0% 95 % CI 0.0–0.0 % · 996,347 runs seen plausible: counts as late6 artefacts excluded

League by country

Threshold 5 min · 13 Nov–19 Nov · sorted by punctuality; two countries are only ordered if the CI of their difference does not cross 0; with more than 20 % of arrivals unobserved a country is not ranked
RankCountryPunctuality < 5 min · 95 % CIObserved to destination / scheduledCoverageOfficial threshold and its figureDelay semantics
1NO 🇳🇴 Norway
87.2 % 86.3–88.1
5,149 / 13,072 48 % no official threshold on recordseconds
2PL 🇵🇱 Poland
84.3 % 84.0–84.7
42,144 / 44,657 95 % no official threshold on recordwhole minutes
3GB 🇬🇧 United Kingdom
79.4 % 79.1–79.6
130,032 / 150,286 81 % PPM 72.8–81.5 %PPM: < 5 min regionales, < 10 min larga distancia; cancelación = fallowhole minutes
4BG 🇧🇬 Bulgaria
76.6 % 75.0–78.2
2,667 / 2,718 98 % no official threshold on recordwhole minutes
–CH 🇨🇭 Switzerland
96.7 % 96.6–96.9
dimmed bounds 70.8–97.6 % ◆ 27 % arrivals unobserved
75,393 / 130,544 58 % ▼ 13 pp92.7 % · < 3 minSBB: puntualidad de clientes, < 3 minthreshold not verifiedseconds · arrival sometimes missing
–FI 🇫🇮 Finland
96.0 % 95.5–96.5
bounds 75.2–96.9 % ◆ 22 % arrivals unobserved
6,353 / 8,084 99 % 96.0 % · < 5 minLarga distancia < 5 min; cercanías < 3 minthreshold not verifiedseconds · arrival sometimes missing
–AU 🇦🇺 Australia
93.0 % 92.5–93.4
bounds 27.7–97.9 % ◆ 70 % arrivals unobserved
12,232 / 32,443 57 % no official threshold on recordseconds · arrival sometimes missing
–US 🇺🇸 United States
84.9 % 84.1–85.6
bounds 46.5–91.7 % ◆ 45 % arrivals unobserved
8,541 / 24,216 43 % no official threshold on recordseconds · arrival sometimes missing
–DE 🇩🇪 Germany
76.3 % 76.1–76.5
bounds 53.8–83.3 % ◆ 29 % arrivals unobserved
196,093 / 243,565 82 % threshold 6 min (not computed here)DB: < 6 minthreshold not verifiedseconds · arrival sometimes missing
–ES 🇪🇸 Spain
51.8 % 51.0–52.6
bounds 20.8–80.7 % ◆ 60 % arrivals unobserved
16,017 / 32,045 67 % no official threshold on recordseconds · arrival sometimes missing
1 NO 🇳🇴 Norway 87.2 %
95 % CI 86.3–88.1 %
More about this country
Observed to destination / scheduled
5,149 / 13,072
Coverage
48 %
Official threshold and its figure
no official threshold on record
Delay semantics
seconds
2 PL 🇵🇱 Poland 84.3 %
95 % CI 84.0–84.7 %
More about this country
Observed to destination / scheduled
42,144 / 44,657
Coverage
95 %
Official threshold and its figure
no official threshold on record
Delay semantics
whole minutes
3 GB 🇬🇧 United Kingdom 79.4 %
95 % CI 79.1–79.6 %
More about this country
Observed to destination / scheduled
130,032 / 150,286
Coverage
81 %
Official threshold and its figure
PPM 72.8–81.5 %· PPM: < 5 min regionales, < 10 min larga distancia; cancelación = fallo
Delay semantics
whole minutes
4 BG 🇧🇬 Bulgaria 76.6 %
95 % CI 75.0–78.2 %
More about this country
Observed to destination / scheduled
2,667 / 2,718
Coverage
98 %
Official threshold and its figure
no official threshold on record
Delay semantics
whole minutes
– CH 🇨🇭 Switzerland 96.7 %
95 % CI 96.6–96.9 %
dimmed bounds 70.8–97.6 % ◆ 27 % arrivals unobserved
More about this country
Observed to destination / scheduled
75,393 / 130,544
Coverage
58 %
Official threshold and its figure
92.7 % · < 3 min· SBB: puntualidad de clientes, < 3 min
Delay semantics
seconds · arrival sometimes missing
– FI 🇫🇮 Finland 96.0 %
95 % CI 95.5–96.5 %
bounds 75.2–96.9 % ◆ 22 % arrivals unobserved
More about this country
Observed to destination / scheduled
6,353 / 8,084
Coverage
99 %
Official threshold and its figure
96.0 % · < 5 min· Larga distancia < 5 min; cercanías < 3 min
Delay semantics
seconds · arrival sometimes missing
– AU 🇦🇺 Australia 93.0 %
95 % CI 92.5–93.4 %
bounds 27.7–97.9 % ◆ 70 % arrivals unobserved
More about this country
Observed to destination / scheduled
12,232 / 32,443
Coverage
57 %
Official threshold and its figure
no official threshold on record
Delay semantics
seconds · arrival sometimes missing
– US 🇺🇸 United States 84.9 %
95 % CI 84.1–85.6 %
bounds 46.5–91.7 % ◆ 45 % arrivals unobserved
More about this country
Observed to destination / scheduled
8,541 / 24,216
Coverage
43 %
Official threshold and its figure
no official threshold on record
Delay semantics
seconds · arrival sometimes missing
– DE 🇩🇪 Germany 76.3 %
95 % CI 76.1–76.5 %
bounds 53.8–83.3 % ◆ 29 % arrivals unobserved
More about this country
Observed to destination / scheduled
196,093 / 243,565
Coverage
82 %
Official threshold and its figure
threshold 6 min (not computed here)· DB: < 6 min
Delay semantics
seconds · arrival sometimes missing
– ES 🇪🇸 Spain 51.8 %
95 % CI 51.0–52.6 %
bounds 20.8–80.7 % ◆ 60 % arrivals unobserved
More about this country
Observed to destination / scheduled
16,017 / 32,045
Coverage
67 %
Official threshold and its figure
no official threshold on record
Delay semantics
seconds · arrival sometimes missing
value and 95 % CI (Wilson); scale 20–100 % cannot be told apart the CI of the difference with the country above crosses 0 dimmed coverage changed by more than 10 pp in a day: not ranked with confidence

Wilson treats each train as independent; trains on the same day and line are not, so the real uncertainty is larger (the day-and-line block bootstrap will come with more history).

🇨🇭 Switzerland: 27 % of the observed runs have no observed arrival. The feed stops reporting some trains before the destination. Those runs are not in the percentage (we do not know whether they arrived on time) but they do count in the coverage. The figure only describes the trains whose arrival we saw and could be biased: we show pessimistic and optimistic bounds and do not rank it in the league (more than 20 % unobserved).
🇫🇮 Finland: 22 % of the observed runs have no observed arrival. The feed stops reporting some trains before the destination. Those runs are not in the percentage (we do not know whether they arrived on time) but they do count in the coverage. The figure only describes the trains whose arrival we saw and could be biased: we show pessimistic and optimistic bounds and do not rank it in the league (more than 20 % unobserved).
🇦🇺 Australia: 70 % of the observed runs have no observed arrival. The feed stops reporting some trains before the destination. Those runs are not in the percentage (we do not know whether they arrived on time) but they do count in the coverage. The figure only describes the trains whose arrival we saw and could be biased: we show pessimistic and optimistic bounds and do not rank it in the league (more than 20 % unobserved).
🇺🇸 United States: 45 % of the observed runs have no observed arrival. The feed stops reporting some trains before the destination. Those runs are not in the percentage (we do not know whether they arrived on time) but they do count in the coverage. The figure only describes the trains whose arrival we saw and could be biased: we show pessimistic and optimistic bounds and do not rank it in the league (more than 20 % unobserved).
🇩🇪 Germany: 29 % of the observed runs have no observed arrival. The feed stops reporting some trains before the destination. Those runs are not in the percentage (we do not know whether they arrived on time) but they do count in the coverage. The figure only describes the trains whose arrival we saw and could be biased: we show pessimistic and optimistic bounds and do not rank it in the league (more than 20 % unobserved).
🇪🇸 Spain: 60 % of the observed runs have no observed arrival. The feed stops reporting some trains before the destination. Those runs are not in the percentage (we do not know whether they arrived on time) but they do count in the coverage. The figure only describes the trains whose arrival we saw and could be biased: we show pessimistic and optimistic bounds and do not rank it in the league (more than 20 % unobserved).

Countries where 0 means «no reported incident»

outside the league
FR 🇫🇷 France

The feed only reports delays of 5 min or more; nothing is shown below that. A train without a notice counts as «no incident», which is not the same as «arrived on time».

No incident < 5 min 72.6 % 95 % CI 72.2–72.9 % · n 53,314
coverage 94 %semantics: only incidents ≥ 5 min

Not published

rule: coverage ≥ 30 % of scheduled trains
  • CZ 🇨🇿 Czechia

    Coverage 29.4 % (< 30 %): 15,734 of 53,457 scheduled with data.

    Not published
  • BE 🇧🇪 Belgium

    Coverage 0.0 % (< 30 %): 0 of 23,985 scheduled with data.

    Not published
  • CA 🇨🇦 Canada

    We do not know what a zero delay means in this source.

    Not published
  • DK 🇩🇰 Denmark

    Coverage 0.0 % (< 30 %): 0 of 23,589 scheduled with data.

    Not published
  • EE 🇪🇪 Estonia

    Coverage 0.0 % (< 30 %): 0 of 1,492 scheduled with data.

    Not published
Show 14 more
  • HR 🇭🇷 Croatia

    Coverage 0.0 % (< 30 %): 0 of 3,986 scheduled with data.

    Not published
  • IE 🇮🇪 Ireland

    Only 0 observed arrivals (minimum 50).

    Not published
  • IL 🇮🇱 Israel

    Coverage 0.0 % (< 30 %): 0 of 3,076 scheduled with data.

    Not published
  • IT 🇮🇹 Italy

    Coverage 5.9 % (< 30 %): 1,310 of 22,278 scheduled with data.

    Not published
  • JP 🇯🇵 Japan

    No timetable to compare with (18,175 runs seen).

    Not published
  • LU 🇱🇺 Luxembourg

    Coverage 0.0 % (< 30 %): 0 of 6,083 scheduled with data.

    Not published
  • LV 🇱🇻 Latvia

    Coverage 0.0 % (< 30 %): 0 of 1,944 scheduled with data.

    Not published
  • MY 🇲🇾 Malaysia

    Coverage 0.0 % (< 30 %): 0 of 17,980 scheduled with data.

    Not published
  • NL 🇳🇱 Netherlands

    Only 18 observed arrivals (minimum 50).

    Not published
  • PT 🇵🇹 Portugal

    Coverage 0.0 % (< 30 %): 0 of 8,465 scheduled with data.

    Not published
  • RO 🇷🇴 Romania

    Coverage 0.0 % (< 30 %): 0 of 11,381 scheduled with data.

    Not published
  • SE 🇸🇪 Sweden

    Only 0 observed arrivals (minimum 50).

    Not published
  • SI 🇸🇮 Slovenia

    Coverage 0.0 % (< 30 %): 0 of 3,817 scheduled with data.

    Not published
  • SK 🇸🇰 Slovakia

    Coverage 0.0 % (< 30 %): 0 of 13,148 scheduled with data.

    Not published

Daily trend

Punctuality at destination < 5 min per day · 13 Nov to 19 Nov · up to 4 highlighted countries
other countries unstable coverage (hollow point)
40 %60 %80 %100 %13 Nov14 Nov15 Nov16 Nov17 Nov18 Nov19 Nov
View as table
DayAUBGCHDEESFIFRGBNOPLUS
13 Nov80.7 %–96.9 %–46.5 %96.3 %77.1 %73.8 %82.1 %85.0 %82.9 %
14 Nov94.7 %75.9 %98.4 %83.5 %58.2 %96.6 %81.9 %77.2 %84.3 %84.7 %89.9 %
15 Nov91.2 %73.2 %98.2 %86.4 %62.6 %96.6 %73.8 %80.5 %92.0 %83.9 %88.6 %
16 Nov91.0 %80.9 %95.3 %73.3 %51.2 %97.2 %72.0 %79.8 %90.4 %84.6 %79.8 %
17 Nov93.1 %–96.3 %73.1 %51.5 %96.9 %68.1 %77.8 %91.2 %83.3 %85.9 %
18 Nov93.2 %78.0 %95.8 %71.3 %46.7 %95.5 %70.6 %85.2 %84.5 %85.2 %83.8 %
19 Nov94.9 %74.9 %95.8 %75.0 %49.6 %93.2 %68.2 %81.2 %85.2 %83.5 %84.1 %

GB · industry standards

computed by TrainsRadar from TRUST; not Network Rail's official figures
observed

PPM by sector

Regional and commuter (< 5 min) 72.5 % n 129,115
Long distance (< 10 min) 75.0 % n 16,443

PPM without cancellations (TRUST does not store them here): arrival at destination less than 5 min late (10 min long distance); no observed arrival counts as a failure. Black bar: 95 % CI (Wilson). Dashed line: overall PPM (72.8 %).

CaSL0.9%≥ 30 min late, partial or no arrival; without cancellations · between 0.9 and 11.5 %
On Time44.7%early or within its minute at every recorded stop · CI 44.5–44.9 · n 278,472 stops
T-371.9%< 3 min at every recorded stop · CI 71.8–72.1 · n 278,472 stops
T-1596.1%< 15 min at every recorded stop · CI 96.1–96.2 · n 278,472 stops

Stops of a single day (19 Nov) recorded in TRUST, which gives the variation in whole minutes; stops of the same train are not independent.

Where delay is generated

countries with an occupancy engine: FI and GB · 13 Nov to 19 Nov
primary (not explained by another train) reactionary (lost behind another train)
GB Northern 80 % 20 % 8,875 late arrivals · 49,379 min
GB Govia Thameslink Railway 79 % 21 % 7,501 late arrivals · 41,836 min
GB GWR 85 % 15 % 4,983 late arrivals · 32,596 min
GB Southeastern 81 % 19 % 4,519 late arrivals · 28,031 min
GB ScotRail 86 % 14 % 6,073 late arrivals · 25,160 min
GB Avanti West Coast 80 % 20 % 1,802 late arrivals · 20,151 min
GB West Midlands Trains 79 % 21 % 3,799 late arrivals · 19,773 min
GB South Western Railway 83 % 17 % 3,031 late arrivals · 19,366 min
GB Transport for Wales 88 % 12 % 2,868 late arrivals · 18,993 min
Approximate attribution: reactionary = minutes lost in observed hindrances by another train (section occupancy in FI, berths in GB); the rest counts as primary. Not the official attribution. In GB we will check it against the published delay codes once their licence is confirmed.
View as table
GroupReactionarynmin
GB Northern19.9 %8,87549,379
GB Govia Thameslink Railway20.8 %7,50141,836
GB GWR14.8 %4,98332,596
GB Southeastern19.2 %4,51928,031
GB ScotRail13.8 %6,07325,160
GB Avanti West Coast20.2 %1,80220,151
GB West Midlands Trains20.7 %3,79919,773
GB South Western Railway17.3 %3,03119,366
GB Transport for Wales11.9 %2,86818,993
Full table: every country and threshold

Punctuality over scheduled trains

Service days from 2026-11-13 to 2026-11-19. One row per run in the day's timetable, with or without data (source: run_facts). Punctuality is measured at the destination, over the runs whose arrival we saw.

RankCountryScheduledWith dataArrival seenWithin 1 minWithin 3 minWithin 5 minWithin 10 minWithin 15 minWithin 30 minMean delay at destinationCancelled
1🇳🇴 Norway !13,07247.9 %47.0–48.7 ≤ 48.7 % by count80.9 %64.0 %62.7–65.380.4 %79.3–81.487.2 %86.3–88.195.2 %94.6–95.898.0 %97.6–98.499.6 %99.3–99.72.0 min236
2🇵🇱 Poland !44,65794.8 %94.6–95.0 ≤ 94.8 % by count94.9 %61.6 %61.1–62.075.9 %75.5–76.384.3 %84.0–84.792.5 %92.2–92.795.4 %95.2–95.698.3 %98.1–98.42.9 min81
3🇬🇧 United Kingdom !150,28681.1 %80.9–81.3 ≤ 96.8 % by count89.3 %46.7 %46.4–47.068.0 %67.7–68.379.4 %79.1–79.691.4 %91.2–91.695.6 %95.5–95.799.0 %99.0–99.13.1 min–
4🇧🇬 Bulgaria !2,71898.3 %97.8–98.7 ≤ 100.0 % by count82.4 %64.2 %62.4–66.069.3 %67.5–71.076.6 %75.0–78.287.2 %85.9–88.492.6 %91.5–93.597.5 %96.8–98.03.9 min–
5🇫🇷 France !57,00994.1 %93.9–94.2 ≤ 94.5 % by count95.3 %––72.6 %72.2–72.984.7 %84.4–85.089.7 %89.4–89.995.4 %95.2–95.54.5 min1,461
🇦🇺 Australia Not ranked !32,44356.6 %56.1–57.2 ≤ 95.4 % by count29.7 %between 19.9 and 90.2 %between 25.6 and 95.8 %between 27.7 and 97.9 %between 29.2 and 99.5 %between 29.5 and 99.8 %between 29.7 and 100.0 %1.4 min–
🇧🇪 Belgium Not ranked !23,9850.0 %0.0–0.0 ≤ 0.0 % by count0.0 %––––––––
🇨🇦 Canada Not ranked !5,2577.3 %6.6–8.0 ≤ 7.3 % by count6.7 %––––––26.0 min–
🇨🇭 Switzerland Not ranked !130,54458.2 %57.9–58.5 ≤ 59.8 % by count73.2 %between 54.2 and 81.0 %between 67.9 and 94.7 %between 70.8 and 97.6 %between 72.4 and 99.2 %between 72.7 and 99.6 %between 73.1 and 99.9 %0.9 min–
🇨🇿 Czechia Not ranked !53,45729.4 %29.0–29.8 ≤ 29.5 % by count29.1 %between 12.4 and 83.3 %between 21.1 and 92.0 %between 24.4 and 95.3 %between 27.6 and 98.5 %between 28.5 and 99.4 %between 29.1 and 99.9 %2.3 min–
🇩🇪 Germany Not ranked !243,56581.6 %81.4–81.7 ≤ 96.2 % by count70.6 %between 35.1 and 64.5 %between 46.8 and 76.2 %between 53.8 and 83.3 %between 62.0 and 91.5 %between 65.4 and 94.8 %between 68.8 and 98.2 %4.1 min–
🇩🇰 Denmark Not ranked !23,5890.0 %0.0–0.0 ≤ 0.0 % by count0.0 %––––––––
🇪🇪 Estonia Not ranked !1,4920.0 %0.0–0.3 ≤ 0.0 % by count0.0 %––––––––
🇪🇸 Spain Not ranked !32,04567.0 %66.5–67.5 ≤ 91.2 % by count40.1 %between 9.3 and 69.2 %between 15.6 and 75.5 %between 20.8 and 80.7 %between 28.1 and 88.1 %between 32.2 and 92.2 %between 37.4 and 97.3 %9.2 min3,111
🇫🇮 Finland Not ranked !8,08499.3 %99.1–99.4 ≤ 100.0 % by count78.3 %between 61.0 and 82.7 %between 72.2 and 93.9 %between 75.2 and 96.9 %between 77.0 and 98.6 %between 77.5 and 99.2 %between 78.1 and 99.8 %1.1 min–
🇭🇷 Croatia Not ranked !3,9860.0 %0.0–0.1 ≤ 0.0 % by count0.0 %––––––––
🇮🇪 Ireland Not ranked !5,50074.4 %73.2–75.5 ≤ 74.6 % by count0.0 %––––––––
🇮🇱 Israel Not ranked !3,0760.0 %0.0–0.1 ≤ 0.0 % by count0.0 %––––––––
🇮🇹 Italy Not ranked !22,2785.9 %5.6–6.2 ≤ 38.6 % by count0.0 %––––––––
🇯🇵 Japan Not ranked ––0.0 %––––––––
🇱🇺 Luxembourg Not ranked !6,0830.0 %0.0–0.1 ≤ 0.0 % by count0.0 %––––––––
🇱🇻 Latvia Not ranked !1,9440.0 %0.0–0.2 ≤ 0.0 % by count0.0 %––––––––
🇲🇾 Malaysia Not ranked !17,9800.0 %0.0–0.0 ≤ 0.0 % by count0.0 %––––––––
🇳🇱 Netherlands Not ranked !42,67789.1 %88.8–89.4 ≤ 89.4 % by count0.0 %––––––10.1 min–
🇵🇹 Portugal Not ranked !8,4650.0 %0.0–0.0 ≤ 0.0 % by count0.0 %––––––––
🇷🇴 Romania Not ranked !11,3810.0 %0.0–0.0 ≤ 0.0 % by count0.0 %––––––––
🇸🇪 Sweden Not ranked !12,81355.1 %54.3–56.0 ≤ 55.6 % by count0.0 %––––––––
🇸🇮 Slovenia Not ranked !3,8170.0 %0.0–0.1 ≤ 0.0 % by count0.0 %––––––––
🇸🇰 Slovakia Not ranked !13,1480.0 %0.0–0.0 ≤ 0.0 % by count0.0 %––––––––
🇺🇸 United States Not ranked !24,21642.6 %42.0–43.2 ≤ 51.2 % by count54.8 %between 26.4 and 71.6 %between 40.2 and 85.4 %between 46.5 and 91.7 %between 51.3 and 96.5 %between 52.5 and 97.7 %between 53.8 and 99.1 %3.4 min–

Great Britain: PPM and CaSL (approximation)

PPM without cancellations: 72.8 % (no arrival = failure) · 81.5 % (no arrival, left out of the base) · CaSL without cancellations: 0.9 % · 11.5 %

Over 145,558 passenger trains activated in TRUST. Approximation with open data, not an official Network Rail figure. PPM: arrival at the destination less than 5 min late (regional) or 10 min (long distance, approximated by operator). The two figures do NOT bound the real value: we do not have the cancellations (TRUST does not store them here) and trains cancelled before activation are not even in the base, so the real PPM would tend to be below the low figure and the real CaSL above the high one. The plan asked to divide by scheduled trains; Network Rail's open timetable is the permanent one (without the day's changes), which is why the base is the activated trains.

Where the source does not publish delays below a threshold (France: 5 min), that column is left empty instead of showing a misleading number. 95 % Wilson intervals: trains on the same day are not independent (one incident delays many), so the real uncertainty is larger.

More detail: live and history

Punctuality league · now

Trains running right now. Sortable by column. The vertical line on each bar marks the 30 % threshold.

Rank
=1🇯🇵 Japan 191 / 220 100 %98–100 %0.2 min
=1: Statistical tie: it shares the rank because it cannot be told apart from the countries above with the same number (the 95 % interval of the difference includes 0).🇳🇱 Netherlands 96 % at 0 s: 22 of 23 trains report exactly 0 s: the feed may be padding the delay with 0. 23 / 25 100 %85.7–100 %0.1 min
=2🇨🇭 Switzerland 66 / 71 94 %85.4–97.6 %0.5 min
=2: Statistical tie: it shares the rank because it cannot be told apart from the countries above with the same number (the 95 % interval of the difference includes 0).🇸🇪 Sweden 23 / 23 91 %73.2–97.6 %0.2 min
=2: Statistical tie: it shares the rank because it cannot be told apart from the countries above with the same number (the 95 % interval of the difference includes 0).🇦🇺 Australia 311 / 374 86 %81.2–89 %1.2 min
=5🇩🇪 Germany 343 / 405 79 %74.1–82.7 %2.6 min
=5: Statistical tie: it shares the rank because it cannot be told apart from the countries above with the same number (the 95 % interval of the difference includes 0).🇵🇱 Poland 27 / 40 70 %51.5–84.1 %5.5 min
=7🇬🇧 United Kingdom 112 / 114 64 %55.1–72.6 %3.1 min
=7: Statistical tie: it shares the rank because it cannot be told apart from the countries above with the same number (the 95 % interval of the difference includes 0).🇺🇸 United States 344 / 412 63 %57.9–68 %6.0 min

The rank only counts countries that are significantly better (95 % interval of the difference excluding 0); “=” marks a shared rank, i.e. a statistical tie. Wilson intervals; trains on the same day are not independent, so the real uncertainty is larger.

No league: Canada. We do not know what a delay of 0 means in its source.

Not enough data · 16 countries outside the ranking

With fewer than 30 % of trains reporting data, punctuality describes a minority that need not resemble the rest (e.g. one operator only). They are not ranked and their percentage is not shown as if it were the country's. Their 365 trains do appear on the map.

CountryTrains with data / totalCoverageReason
🇮🇹 Italy12 / 12 Small sample (12 < 20)
🇳🇴 Norway12 / 12 Small sample (12 < 20)
🇧🇬 Bulgaria7 / 7 Small sample (7 < 20)
🇮🇪 Ireland1 / 1 Small sample (1 < 20)
🇫🇮 Finland17 / 18 Small sample (17 < 20)
🇫🇷 France5 / 6 Small sample (5 < 20)

No realtime delay feed (10):

Trend · 30 days

Daily punctuality, weighted by services (source: daily_stats).

65 %70 %75 %80 %85 %23/1030/1006/1113/1119/1128-day mean: 73.6 %

2,975,388 services with data over 28 days. 14–16 countries per day.

Delay Distribution · now

Over the 1,522 trains with RT data (not the 2,096 on the map).

  • Early or on time 56.2 %
  • < 3 min 22.1 %
  • 3-5 min 5.8 %
  • 5-10 min 6.8 %
  • 10-20 min 4.1 %
  • > 20 min 5.1 %
Punctual (< 3 min) 3–10 min 10-20 min > 20 min

Historical Punctuality

Weighted by services: a day with 4,000 trains weighs more than one with 400. Source: daily_stats (last delay seen for each train, over the trains seen, not the scheduled ones), days from 23 Oct to 19 Nov.

RankCountryServicesDays with dataPunctualityAverage delay
1🇨🇭 Switzerland 483,04928 / 28 93.3 %93.2–93.4 %0.6 min
2🇫🇮 Finland 32,21528 / 28 92.6 %92.3–92.9 %0.7 min
3🇯🇵 Japan 95,33528 / 28 90.7 %90.5–90.9 %1.0 min
4🇳🇱 Netherlands 142,14625 / 28 88.6 %88.4–88.8 %1.3 min
5🇦🇺 Australia 170,06028 / 28 84.8 %84.6–85 %1.7 min
=6🇳🇴 Norway 27,76128 / 28 79.7 %79.2–80.2 %2.3 min
=6: Statistical tie: it shares the rank because it cannot be told apart from the countries above with the same number (the 95 % interval of the difference includes 0).🇸🇪 Sweden 28,21028 / 28 79.4 %78.9–79.9 %2.2 min
=8🇵🇱 Poland 161,32725 / 28 72.2 %72–72.4 %3.2 min
=8: Statistical tie: it shares the rank because it cannot be told apart from the countries above with the same number (the 95 % interval of the difference includes 0).🇧🇬 Bulgaria 14,38627 / 28 71.6 %70.8–72.3 %4.4 min
10🇩🇪 Germany 1,132,55428 / 28 70.1 %70–70.2 %3.5 min
11🇺🇸 United States 86,38728 / 28 69.7 %69.4–70 %3.9 min
12🇬🇧 United Kingdom 603,59028 / 28 67.3 %67.2–67.4 %3.1 min
13🇮🇹 Italy 35,44828 / 28 52.0 %51.5–52.5 %4.6 min
14🇮🇪 Ireland 18,78228 / 28 40.5 %39.8–41.2 %5.7 min
15🇪🇸 Spain 157,31828 / 28 36.7 %36.4–37 %11.1 min

Period 23 Oct – 19 Nov. 28 days with data in the system. ⚑ = the country has data on fewer than 50 % of those days (14); none in this period. Average delay = mean over all trains with data, weighted by services. The rank only counts countries that are significantly better (95 % interval of the difference excluding 0); “=” marks a shared rank, i.e. a statistical tie. Wilson intervals; trains on the same day are not independent, so the real uncertainty is larger.

Not ranked at 3 min: France (its source only reports delays of 5 min or more, and the daily history only stores 3-minute punctuality).

Stations · now

Next stop of running trains, from worst to best punctuality. Only stations with 10 or more trains with data; n = trains with data.

StationnPunctualityAverage delay
Jamaica12 75 %1.2 min

Only 1 of the 1,523 stations with trains approaching meet the minimum right now. With n = 10, a single train moves punctuality by 10 points: that is why n is always shown.

Punctuality by hour

Last 7 days, each country's local time (source: line_delay_hourly). The units are position readings, not trains: a train counts once per reading.

60 %70 %80 %90 %0h6h12h18h23h
70–90 % fair < 70 % poor Axis from 60 to 90 %

Filters

Punctuality by Type

Click on a bar to filter by that type

1 / 2

Worst Lines

  1. 1 Gold RunnerUS ic 0.0 % 25.8 min
  2. 2 VIA RailCA ic 28.6 % 30.1 min
  3. 3 ICNIT ic 50.0 % 17.5 min
  4. 4 Northeast RegionalUS Regional Express 50.0 % 8.1 min
  5. 5 БВBG ic 57.1 % 10.6 min
  6. 6 Avanti West CoastGB ic 60.0 % 19.0 min
  7. 7 Pacific SurflinerUS Regional 62.5 % 13.0 min

More data by country

Active Services by Country

CountryServicesDelayedAverage delayRT Data
🇺🇸 United States4121276.0 minYes
🇩🇪 Germany405732.6 minYes
🇦🇺 Australia374451.2 minYes
🇯🇵 Japan22000.2 minYes
🇲🇾 Malaysia163——No
🇬🇧 United Kingdom114403.1 minYes
🇨🇭 Switzerland7140.5 minYes
🇷🇴 Romania51——No
🇨🇦 Canada472030.1 minYes
🇩🇰 Denmark42——No
🇵🇱 Poland4085.5 minYes
🇳🇱 Netherlands2500.1 minYes
🇸🇪 Sweden2320.2 minYes
🇫🇮 Finland18——No
🇨🇿 Czechia14——No
🇵🇹 Portugal13——No
🇮🇹 Italy12——No
🇳🇴 Norway12——No
🇪🇸 Spain9——No
🇸🇰 Slovakia8——No
🇧🇬 Bulgaria7——No
🇫🇷 France6——No
🇭🇷 Croatia5——No
🇱🇺 Luxembourg2——No
🇧🇪 Belgium2——No
🇮🇪 Ireland1——No

How late are the late trains (only trains with any delay, ≤2h; countries with ≥30% coverage)

CountryMedianAverageP90Maximum>5minTotal
🇨🇦 Canada31.0 min36.7 min75.6 min87.0 min2023
🇵🇱 Poland14.0 min17.1 min32.0 min65.0 min611
🇺🇸 United States4.0 min11.0 min33.0 min108.0 min87211
🇩🇪 Germany4.0 min7.3 min16.0 min66.0 min52131
🇬🇧 United Kingdom4.0 min7.7 min16.0 min91.0 min2563
🇳🇱 Netherlands2.9 min2.9 min2.9 min2.9 min01
🇦🇺 Australia1.5 min2.7 min5.5 min42.4 min16146
🇸🇪 Sweden1.3 min1.9 min4.3 min6.0 min17
🇯🇵 Japan1.0 min1.2 min2.0 min2.0 min031
🇨🇭 Switzerland0.5 min1.5 min3.4 min15.4 min229

Data Quality by Country

CountryMethodInterpolatedRecords
🇦🇺 AustraliagpsNo327
🇦🇺 AustraliascheduledYes47
🇧🇪 BelgiumscheduledYes2
🇧🇬 BulgariascheduledYes7
🇨🇦 CanadascheduledYes47
🇨🇭 SwitzerlandscheduledYes71
🇨🇿 CzechiascheduledYes14
🇩🇪 GermanyscheduledYes405
🇩🇰 DenmarkscheduledYes42
🇪🇸 SpaingpsNo9
🇫🇮 FinlandgpsNo18
🇫🇷 FrancescheduledYes6
🇬🇧 United Kingdomstation_eventNo114
🇭🇷 CroatiascheduledYes5
🇮🇪 Irelandstation_eventNo1
🇮🇹 ItalyscheduledYes12
🇯🇵 Japanstation_eventYes191
🇯🇵 JapangpsNo29
🇱🇺 LuxembourgscheduledYes2
🇲🇾 MalaysiascheduledYes158
🇲🇾 MalaysiagpsNo5
🇳🇱 NetherlandsgpsNo25
🇳🇴 NorwaygpsNo12
🇵🇱 PolandscheduledYes40
🇵🇹 PortugalscheduledYes13
🇷🇴 RomaniascheduledYes51
🇸🇪 SwedengpsNo23
🇸🇰 SlovakiascheduledYes8
🇺🇸 United StatesgpsNo412

Historical Data

88 records
DateServicesPositionsPunctuality Average delay
Thu 19 Nov159,48117,750,550 70.0 % 3.6 min
Wed 18 Nov158,49717,670,651 70.7 % 3.7 min
Tue 17 Nov154,68016,516,107 70.4 % 3.7 min
Mon 16 Nov158,61817,292,237 69.8 % 3.6 min
Sun 15 Nov119,37612,794,616 78.6 % 2.4 min
Sat 14 Nov135,06214,154,142 76.4 % 2.6 min
Fri 13 Nov159,03317,529,630 71.8 % 3.4 min
Thu 12 Nov159,04617,574,205 71.2 % 3.5 min
Wed 11 Nov148,08015,604,360 75.1 % 2.8 min
Tue 10 Nov157,42116,632,176 72.7 % 3.1 min
Mon 09 Nov159,75617,587,162 73.7 % 3.0 min
Sun 08 Nov118,86712,866,322 81.1 % 2.3 min
Sat 07 Nov135,57514,209,531 77.9 % 2.6 min
Fri 06 Nov159,54917,597,675 71.7 % 3.3 min
Thu 05 Nov156,76817,260,560 72.0 % 3.1 min
Wed 04 Nov159,05817,797,352 75.0 % 2.9 min
Tue 03 Nov158,44617,482,937 73.0 % 3.2 min
Mon 02 Nov159,31317,582,369 74.2 % 2.8 min
Sun 01 Nov119,30913,101,832 77.9 % 2.3 min
Sat 31 Oct134,09914,406,063 79.0 % 2.3 min
Fri 30 Oct158,50917,805,762 73.9 % 3.0 min
Thu 29 Oct155,94417,029,795 73.7 % 3.1 min
Wed 28 Oct155,83117,624,232 73.1 % 3.0 min
Tue 27 Oct156,59517,553,530 73.7 % 3.0 min
Mon 26 Oct156,66417,688,657 70.4 % 3.7 min
Sun 25 Oct116,90112,993,466 75.4 % 3.6 min
Sat 24 Oct132,98114,525,898 74.7 % 3.2 min
Fri 23 Oct156,80717,665,516 69.8 % 4.0 min
Thu 22 Oct157,99017,845,278 73.8 % 2.9 min
Wed 21 Oct158,13217,778,143 73.2 % 3.1 min
Tue 20 Oct160,57317,355,891 75.6 % 2.7 min
Mon 19 Oct158,13517,428,643 75.4 % 2.8 min
Sun 18 Oct120,16112,716,675 82.0 % 2.0 min
Sat 17 Oct134,93914,302,601 79.2 % 2.3 min
Fri 16 Oct158,86817,844,016 74.7 % 2.8 min
Thu 15 Oct157,94617,714,724 73.8 % 2.9 min
Wed 14 Oct157,67017,111,478 74.5 % 2.7 min
Tue 13 Oct157,89117,638,453 73.9 % 2.9 min
Mon 12 Oct155,46617,259,513 75.3 % 2.8 min
Sun 11 Oct118,70612,829,739 82.0 % 1.9 min
Sat 10 Oct133,83714,131,381 78.9 % 2.3 min
Fri 09 Oct159,00317,565,641 74.7 % 2.9 min
Thu 08 Oct158,39017,315,208 74.3 % 2.9 min
Wed 07 Oct158,73917,639,653 76.0 % 2.7 min
Tue 06 Oct158,67017,556,347 77.0 % 2.5 min
Mon 05 Oct153,99916,056,796 77.7 % 2.4 min
Sun 04 Oct115,98512,375,758 81.8 % 1.9 min
Sat 03 Oct127,76415,375,336 77.1 % 2.2 min
Fri 02 Oct136,71717,124,526 75.6 % 2.4 min
Thu 01 Oct137,47416,838,843 75.5 % 2.4 min
Wed 30 Sept136,73917,336,837 76.5 % 2.4 min
Tue 29 Sept137,88317,545,397 77.1 % 2.3 min
Mon 28 Sept134,95117,015,261 77.9 % 2.2 min
Sun 27 Sept111,43514,957,353 78.0 % 2.2 min
Sat 26 Sept124,90516,279,191 75.8 % 2.2 min
Fri 25 Sept136,82917,073,752 74.9 % 2.5 min
Thu 24 Sept136,76617,500,820 75.3 % 2.4 min
Wed 23 Sept136,83617,537,411 74.0 % 2.6 min
Tue 22 Sept137,14617,566,133 74.4 % 2.5 min
Mon 21 Sept137,92817,480,242 74.8 % 2.5 min
Sun 20 Sept105,69313,514,794 78.6 % 2.1 min
Sat 19 Sept119,87214,927,587 74.6 % 2.7 min
Fri 18 Sept134,24116,324,696 72.7 % 2.7 min
Thu 17 Sept132,51816,465,081 73.2 % 2.8 min
Wed 16 Sept133,75216,902,488 72.2 % 2.9 min
Tue 15 Sept133,08116,737,777 75.4 % 2.4 min
Mon 14 Sept132,29716,422,246 75.3 % 2.5 min
Sun 13 Sept105,76013,505,724 78.5 % 2.1 min
Sat 12 Sept118,74315,021,367 75.1 % 2.5 min
Fri 11 Sept132,30716,251,637 73.2 % 2.8 min
Thu 10 Sept132,61716,585,010 74.4 % 2.5 min
Wed 09 Sept129,21515,886,085 73.6 % 2.8 min
Tue 08 Sept132,25216,704,662 74.5 % 2.5 min
Mon 07 Sept132,52316,420,645 75.7 % 2.4 min
Sun 06 Sept106,40513,572,621 78.9 % 2.1 min
Sat 05 Sept118,21014,926,226 74.2 % 2.8 min
Fri 04 Sept127,62515,244,162 72.7 % 2.9 min
Thu 03 Sept126,87315,202,375 75.1 % 2.5 min
Wed 02 Sept127,21915,278,935 76.8 % 5.8 min
Tue 01 Sept126,81315,165,243 75.5 % 6.7 min
Mon 31 Aug126,26615,158,283 77.6 % 6.6 min
Sun 30 Aug102,03811,976,226 81.2 % 6.4 min
Sat 29 Aug112,77413,424,281 81.4 % 4.9 min
Fri 28 Aug130,13415,291,735 74.2 % 6.3 min
Thu 27 Aug129,98215,529,526 74.8 % 6.4 min
Wed 26 Aug130,27715,512,515 75.7 % 5.9 min
Tue 25 Aug135,86215,812,163 76.7 % 6.5 min
Mon 24 Aug135,34715,900,811 74.6 % 7.2 min

Definitions and sources

Punctuality at destination

Share of runs whose arrival at the last stop was observed with a delay below the threshold. Early arrivals count as on time. Cancellations are not in this percentage (only in GB's PPM and CaSL and in the cancelled card). Delay is measured against the day's scheduled timetable.

Unobserved arrival

The run appears in the feed but stops being reported before the destination. We do not know whether it arrived on time, so it is left out of the percentage. It does count in the coverage. Above 20 %, the country is not ranked and pessimistic and optimistic bounds are given.

Coverage = with data / scheduled

Scheduled: the day's runs in the open timetable (GTFS or equivalent). With data: those matched one to one with a run with a reliable delay in the real-time feed. Below 30 % the country is not published. If coverage changes by more than 10 points from one day to the next, the figure is shown dimmed.

95 % confidence interval

Plausible range of the punctuality (Wilson). To order two countries we look at the interval of their difference (Newcombe), not at whether their intervals overlap. Wilson treats each train as independent; trains on the same day and line are not, so the real uncertainty is larger. The interval does not capture the bias of unobserved trains: that is what coverage is for.

Source of each country and its licence
CountrySourceLicenceDelay resolution
NO 🇳🇴 NorwayEnturNLOD 2.0seconds
PL 🇵🇱 PolandPKP PLK S.A.provider's terms of usewhole minutes
GB 🇬🇧 United KingdomNetwork RailOGL v3.0whole minutes
BG 🇧🇬 Bulgarialivetransport.eu (BDŽ timetables and realtime)CC BY 4.0whole minutes
CH 🇨🇭 Switzerlandopentransportdata.swissprovider's terms of useseconds · arrival sometimes missing
FI 🇫🇮 FinlandFintraffic / digitraffic.fiCC BY 4.0seconds · arrival sometimes missing
AU 🇦🇺 AustraliaTransLink (Queensland Government)
Transport for NSW
Transport Victoria
CC BY 4.0
CC BY 4.0
CC BY 4.0
seconds · arrival sometimes missing
US 🇺🇸 United StatesAmtrak
MTA (LIRR, Metro-North, Staten Island Railway)
MBTA
provider's terms of use
provider's terms of use
provider's terms of use
seconds · arrival sometimes missing
DE 🇩🇪 Germanygtfs.de / DELFI e.V. (realtime)CC BY-SA 4.0seconds · arrival sometimes missing
ES 🇪🇸 SpainRenfe
FGC
CC BY 4.0
CC BY 4.0
seconds · arrival sometimes missing
FR 🇫🇷 FranceSNCF (transport.data.gouv.fr)
Eurostar
ODbL 1.0
Licence Ouverte 2.0
only incidents ≥ 5 min
CZ 🇨🇿 CzechiaCIS JŘ – Ministerstvo dopravy (timetables)
PID – ROPID (via Golemio)
provider's terms of use
CC BY 4.0
seconds
BE 🇧🇪 BelgiumSNCB/NMBS (via iRail)provider's terms of useunclassified
CA 🇨🇦 CanadaVIA Rail (timetable, open.canada.ca)OGL – Canadameaning of 0 unknown
DK 🇩🇰 DenmarkRejseplanenprovider's terms of useunclassified
EE 🇪🇪 EstoniaPeatus.eeprovider's terms of useunclassified
HR 🇭🇷 CroatiaHŽPPprovider's terms of useunclassified
IE 🇮🇪 IrelandNTA – Transport for IrelandCC BY 4.0whole minutes
IL 🇮🇱 IsraelIsrael Ministry of Transportprovider's terms of useunclassified
IT 🇮🇹 ItalyViaggiaTreno (RFI / Trenitalia)
Regione Lombardia / Trenord
provider's terms of use
CC BY 4.0
whole minutes
JP 🇯🇵 JapanODPT – Public Transportation Open Data Centerprovider's terms of usewhole minutes
LU 🇱🇺 Luxembourgdata.public.lu / ATPCC BY 4.0unclassified
LV 🇱🇻 LatviaViviprovider's terms of useunclassified
MY 🇲🇾 Malaysiadata.gov.my / KTMBCC BY 4.0unclassified
NL 🇳🇱 NetherlandsNS APIprovider's terms of useseconds
PT 🇵🇹 PortugalCP – Comboios de Portugalprovider's terms of useunclassified
RO 🇷🇴 Romaniadata.gov.roprovider's terms of useunclassified
SE 🇸🇪 SwedenTrafiklab (Samtrafiken)provider's terms of useseconds
SI 🇸🇮 SloveniaIJPPprovider's terms of useunclassified
SK 🇸🇰 SlovakiaŽSRprovider's terms of useunclassified

Analysis with open data; we are not a control centre. Figures may differ from each operator's official ones because of threshold, coverage and definition.

Sources: current_positions, daily_stats and line_delay_hourly · 21 Nov 2026, 02:13 CET