Deep Water research

Rail Punctuality in Britain, Germany and Switzerland

Compare passenger-rail punctuality in Great Britain, Germany and Switzerland in 2024 using public primary sources such as ORR, Deutsche Bahn and SBB. What do the headline percentages tell us, and how do delay thresholds, cancellations and network coverage change the comparison? Distinguish whole-network from long-distance measures, preserve units and denominator conventions, and explain important uncertainty. Make the full report easy to digest: lead passages with their main finding, use clearly labelled source-supported charts where numerical comparison is valid, and place caveats beside the figures they qualify. Do not present unlike punctuality definitions as a comparable league table.

Sep 24, 2026120 sources reviewed

Key Takeaways

Taken together, [1][6] and [17] indicate opportunities to make direct comparisons with the statistics in this release are limited [1].

  • Great Britain [1]: proportion of trains arriving at final destination early or less than five minutes after scheduled time for London and South East, Regional and Scotland operators, or less than ten minutes for Long Distance operators.

  • Germany [6]: share of punctual stops among all intermediate and final stops in Germany where scheduled arrival time in passenger transport is exceeded by less than six minutes.

  • Switzerland [17]: a train is considered on time if it arrives within three minutes of its scheduled arrival.

  • In 2024, Swiss Federal Railways (SBB) achieved an all-time high in punctuality, with 93.2% of trains arriving on time [17]; in Switzerland, a train is considered on time if it arrives within three minutes of its scheduled arrival [17]; 84.8% of trains arrived within three minutes (early or less than three minutes after the scheduled time) of their scheduled arrival time [10]; Figure 20: Proportion of trains arriving at recorded stations stops within three minutes, Great Britain, annual data, FYE March 2015 to 2024-25 [10]; Deutsche Bahn reports 62.5% for DB Fernverkehr [6] in figures for Stand 2024 bzw. per 31. Dezember 2024 with 2024 and 2023 columns [6], defined as the share of punctual stops among all intermediate and final stops in Germany, counting a passenger stop as punctual if scheduled arrival is exceeded by less than 6 minutes [6].

  • In Switzerland, a train is considered on time if it arrives within three minutes of its scheduled arrival [17]; this contrasts with other countries, such as Germany (six minutes) and France and Italy (five minutes) [17]; Punctuality / % 2024 2023 2022 DB Group (rail) in Germany 89.4 90.1 90.9 DB rail passenger transport in Germany 89.5 90.3 91.0 DB Long-Distance 62.5 64.0 65.2 [5]; Punctuality (whole journey) (DB Long-Distance) 67.4 68.9 69.3 [5]; DB Long-Distance 62.5 64.0 65.2 DB Regional 90.7 91.4 92.2 DB Cargo (Germany) 68.0 70.5 66.1 DB Regional (bus) 85.9 85.2 86.0 DB Cargo 68.2 69.7 66.3 Punctuality (whole journey) (DB Long-Distance) 67.4 68.9 69.3 shows Deutsche Bahn distinguishes standard DB Long-Distance punctuality from a separate whole-journey punctuality measure for DB Long-Distance [5].

  • Cancelled trains do not enter the punctuality statistics [35]; planned stops where the train fails to stop at the location are deliberately excluded as they are captured in the reliability measures [23]; the cancellations score in Great Britain in 2024-25 was 4.1% [10]; the cancellations measure is a weighted score which counts full cancellations as one and part cancellations as half [10]; it only includes trains cancelled after 22:00 the previous evening and trains removed from the days’ timetable before then may not be included [10].

  • Coverage is for the financial year (FY) 2024-25 and includes comparisons with the previous FY 2023-24 and 2019-20 [10]; due to differences in how passenger rail performance is measured in other countries, opportunities to make direct comparisons with statistics in this release are limited [1].

[!WARNING] Taken together, [1][17] and [6] indicate ranking the three headline percentages as a league table misleads.

Abstract

Taken together, SBB [17], ORR [10] and Deutsche Bahn [5] indicate 2024 passenger-rail punctuality needs reading inside each network's timing tolerance and counting base, not as a cross-country ranking [1][24].

Taken together, [5] and [6] indicate that scope changes the figure: Deutsche Bahn reports 62.5% punctuality for DB Long-Distance in 2024 [5][6] and 89.5% punctuality for DB rail passenger transport in Germany in 2024 [5][6]. Deutsche Bahn describes the Belgian figure of around 92% punctuality as covering all train services, with mostly local trains running on a network one-fifth the size of the German rail network, as “Kirschen mit Kürbissen verglichen” [21].

SBB reports 93.2% of SBB trains arriving on time in 2024 [17][20] using arrival within three minutes of scheduled arrival as on time [17][27], while Deutsche Bahn reports 62.5% long-distance operational punctuality in 2024 [5][6] as the share of punctual stops among all intermediate and final stops in Germany with scheduled arrival in passenger transport exceeded by less than 6 minutes [6] for every performed train/bus run [5], with cancelled trains excluded from Deutsche Bahn's punctuality statistics [35] and whole-journey punctuality for DB Long-Distance reported as 67.4% in 2024 [5]. ORR records up to around 96% of all station stops with no estimates for the c.4% not recorded [4][24]. ORR reports 84.8% in Great Britain within three minutes in financial year 2024-25 [10] for recorded station stops arrived at early or less than three minutes after the scheduled time [24], while its cancellations score in Great Britain in 2024-25 was 4.1% [10], counting full cancellations as one and part cancellations as half [10] and only including trains cancelled after 22:00 the previous evening [10], and Deutsche Bahn includes train cancellations in traveller punctuality [3][21]. Taken together, [1] and [24] indicate opportunities to make direct comparisons with other countries' passenger-rail performance statistics are limited [1][24].

Table of Contents

  • Key Takeaways
  • Abstract
  • 1. Introduction
  • 2. Background
  • 3. Findings
    • 3.1 2024 Headline Punctuality Figures by Network
    • 3.2 Delay Thresholds and Punctuality Definitions Compared
    • 3.3 Cancellations and Reliability Denominator Effects
    • 3.4 Uncertainties Limiting Direct Cross-Country Comparisons
  • 4. Discussion
  • 5. Conclusion
  • References

1. Introduction

One release suggests opportunities to make direct comparisons with statistics in that release are limited due to differences in how passenger rail performance is measured in other countries [1].

Taken together, [1][3][4][5] indicate that direct percentage comparisons conflate different clocks, denominators and cancellation boundaries; see Background for detailed definitions and Findings for tabulated comparisons.

Most of the data contained within this statistical release is collected automatically from Network Rail's TRUST System (Train Running System on TOPs (Total Operation Processing System)) and the latest data should be treated as provisional, as train operators provide Network Rail with information e.g. on cancellations, which can be updated over time [31].

Deutsche Bahn figures refer to the 2024 reporting and calendar year in one integrated report, which suggests percent units for that year [6]. Headline percentages, coverage periods and scope boundaries are presented in Findings and are not repeated here.

Background defines each operator’s clock, denominator and cancellation boundary in detail. Findings presents 2024 headline percentages only with their original units, thresholds and coverage intact. Discussion then tests which comparisons survive like-for-like scrutiny and explains remaining uncertainty. Conclusion answers the research question. Keep the sequence.

2. Background

Due to differences in how passenger rail performance is measured in other countries, opportunities to make direct comparisons with statistics in this release are limited [1][24][31]. Taken together, [4][6] and [17] indicate thresholds differ, with On Time defined as early or less than one minute after the scheduled time and Time to 3 and Time to 15 defined as early or less than three and 15 minutes respectively after the scheduled arrival time [4], a passenger stop counted as punctual if the scheduled arrival time is exceeded by less than six minutes [6], and a train in Switzerland considered on time if it arrives within three minutes of its scheduled arrival [17]. Taken together, [4][6] and [24] indicate denominators differ, with On Time and Time to 3 measured as percentages of recorded station stops [4][24], Public Performance Measure measured as the proportion of trains arriving at their final destination early or less than five minutes after the scheduled time for London and South East, Regional and Scotland operators, or less than ten minutes for Long Distance operators [4][24], and the punctuality share measured in relation to all intermediate and final stops in Germany [6].

  • Quarterly and periodic (4-weekly) statistics cover punctuality, reliability and causes of delay for passenger trains on the mainline network in Great Britain [9][11].

  • Network Rail supplies periodic operational data on a 4-weekly basis [9][11].

  • Industry introduced detailed station-to-station, minute-to-minute measures on 1 April 2019 [12][23].

  • Early trains count toward on-time totals [23][24].

Taken together, [4][23] and [24] indicate that On Time, Time to 3 and Time to 15 are measured over recorded station stops, while the Public Performance Measure is measured at the terminus or final destination [4][23][24]. On Time means arrival at a recorded station stop early or less than one minute after the scheduled time (to 59 seconds) against the public timetable, with early trains counted toward the On Time total [4][23][24]. Time to 3 and Time to 15 measure the percentage of recorded station stops arrived at early or less than three and 15 minutes respectively after the scheduled time, and the percentages are cumulative [4][24][26]. The Public Performance Measure is the proportion of trains arriving at the final destination or terminus early or less than five minutes after the scheduled time for London and South East, Regional and Scotland operators, or less than ten minutes for Long Distance operators, having called at all of its planned station stops [4][23][24]. For the Public Performance Measure, the count of PPM passes is divided by the count of trains planned, and any total or other cancellation is regarded as not arriving punctually [23].

  • Up to around 96% of station stops are recorded. No estimates for punctuality at c.4% not recorded [4][31].

  • Timetable is finalised at 22:00 previous evening; trains removed before then are not included [2][4][24].

  • P-coded pre-cancellations are trains removed before finalisation at 22:00 previous evening [31].

  • Network Rail's TRUST System is Train Running System on TOPs (Total Operation Processing System) [31].

  • Deutsche Bahn compares scheduled arrival with actual arrival for every performed train or bus run and aggregates on-schedule plus up-to-maximum-delay arrivals into punctuality rate [5][6][19].

  • Operational punctuality for passenger stops is less than six minutes over scheduled arrival [6][21] (at most 5:59) [3]. Published percentages are punctual stops as share of all intermediate and terminal stops [3][21].

  • Monthly statistics cover more than 800,000 passenger-train runs, with more than 20,000 long-distance and around 780,000 local runs including all S-Bahn services, and include all traffic stops [3][21].

  • Traveller and whole-journey measures and cancellation handling are in Findings.

  • Regional figure uses third-weighted representation across alternating-current S-Bahn, direct-current S-Bahn and regional rail without S-Bahn [5][6][7].

  • Taken together, [5] and [6] indicate coverage differs by series: series labelled DB Group (rail) in Germany and DB rail passenger transport in Germany, a DB Regional series, and a series labelled Pünktlichkeit Schiene in Deutschland covering internal and external operators [5][6].

Taken together, [17] and [22] indicate Switzerland uses a shorter on-time window than Deutsche Bahn. In Switzerland, a train is considered on time if it arrives within three minutes of its scheduled arrival [17]. The Deutsche Bahn threshold is defined above. The Swiss railway network is one of the busiest in the world [15]. Every day, well over a million passengers and more than two hundred thousand tonnes of freight travel on board more than ten thousand trains along the SBB rail network [15]. SBB reports punctuality for yesterday and over the last month, across Switzerland and per region [15]. Further punctuality figures can be found on the SBB statistics portal [15]. Even small, everyday disruptions can have far-reaching effects [15].

Evidence suggests baselines shifted after 2020 [2][10]. Evidence suggests punctuality in Great Britain improved from April 2020 on fewer planned trains and passengers during the pandemic and has since slipped toward pre-pandemic levels [2][10]. One report suggests pre-pandemic punctuality sat at around 84-85% within three minutes and notes 92.9% in the financial year ending March 2021 on reduced services [10]. Punctuality in German rail transport kept declining despite intensive operational management [5][7][19]. Deutsche Bahn attributes the drag to aged and overloaded infrastructure with many restricted-speed sections, very high construction volumes, high traffic density, staff shortages and lack of system robustness, plus floods in southern Germany in early June 2024, third-party interventions and GDL strikes [5][6][19]. One report suggests strikes hit early 2024 [13].

This section places caveats beside the figures they qualify. Coverage, provisional status, timetable-boundary and strike-timetable caveats are defined earlier in this section and are not repeated here. One report suggests Deutsche Bahn tracks 90-second delays internally as Lost Units [6]. General guidance on administrative-data completeness and on terms such as ‘around’ is covered where uncertainty is discussed and is not repeated here.

3. Findings

3.1 2024 Headline Punctuality Figures by Network

  • SBB 2024: 93.2% arriving on time, an all-time high [17].

  • DB Fernverkehr: Pünktlichkeit DB Fernverkehr in % is 62,5 [6] in a table headed 2024, 2023 [6]. Development is Punctuality / % for 2024, 2023, 2022 [5].

  • Great Britain FY 2024-25: 84.8% arrived within three minutes (early or less than three minutes after scheduled time), down 0.4pp on previous year and up 0.8pp on 2019-20 (table 3133) [10]. Coverage is FY 2024-25 with comparisons to FY 2023-24 and 2019-20 [10].

Figure 20 is the proportion of trains arriving at recorded stations stops within three minutes, Great Britain, annual data, FYE March 2015 to 2024-25 [10]. Statistics are for surface rail only, and do not include underground, light rail, and tram systems [10]. The Rail Factsheet reports 67.6% of trains arrived within 59 seconds of their scheduled arrival time, a decrease of 0.2 pp on the previous year and an increase of 2.6 pp on the FYE March 2020 (ORR Passenger Rail Performance table 3133) [2]. In the latest quarter (1 April to 30 June 2024), PPM for Great Britain was 87.4% [4]. In the latest quarter (1 April to 30 June 2024), 70.1% of recorded station stops in Great Britain (14.4 million out of 20.5 million) were arrived at On Time [4]. For the 12 months up to June 2024 (1 July 2023 to 30 June 2024), 67.5% of recorded station stops in Great Britain (53.6 million out of 79.4 million) were arrived at On Time [4]. PPM for the 12 months up to June 2024 (1 July 2023 to 30 June 2024), was 85.6% [4]. In the FYE March 2024, there was a 6.9% increase in planned train services and a less than 1 pp decrease in punctuality compared with the previous year [2].

Deutsche Bahn reports long-distance operational punctuality was 62.5% in 2024 [6]. Reisendenpünktlichkeit for DB Fernverkehr was 67.4% in 2024 [7]. Operational punctuality fell from 64.0% in 2023 and 65.2% in 2022, Deutsche Bahn reports [5]. Passenger whole-journey punctuality reached 67.4% in 2024 against 68.9% in 2023 and 69.3% in 2022 on a whole-journey basis, Deutsche Bahn reports [5]. Taken together, [19] indicates that whole-journey 67.4% exceeds the operational 62.5% for the same year. Deutsche Bahn places headline passenger figures on rail in Germany in percent [5]. For broader networks, Deutsche Bahn reports rail passenger transport in Germany at 89.5%, down from 90.3% in 2023 and 91.0% in 2022 [5]. Deutsche Bahn reports Group rail in Germany at 89.4%, down from 90.1% in 2023 and 90.9% in 2022 [5]. Regional punctuality stood at 90.7% in 2024 against 91.4% in 2023 and 92.2% in 2022 on an adjusted third-weighted basis, Deutsche Bahn reports [5]. Rail punctuality in Germany for external and internal railway undertakings was 88.1%, down from 88.9% [6]. Monthly statistics exceed 800,000 passenger-train runs, Deutsche Bahn reports [3]. More than 20,000 are long-distance and around 780,000 local including all S-Bahn services, Deutsche Bahn reports [3]. Traveller punctuality is evaluated from the journeys of on average more than 10 million monthly long-distance passengers [3].

In SBB passenger transport, punctuality increased versus the prior year in nearly all regions [20]. German-speaking Switzerland had the highest punctuality, with Region Mitte at 93.71% and Region Ost at 93.65% [20]. RailwayPro reports Ticino punctuality reached a record high of 92.6%, 2.1 percentage points above the previous year [17]. In French-speaking Switzerland, punctuality rose by 2.7 percentage points to 91.9%, slightly below its target of 92.0% [17]. Western Switzerland and Ticino increased their punctuality values and contributed significantly to the improvement [20]. Connection punctuality was 98.7%, as in the previous year [20]. Each day, well over a million passengers and more than 200,000 tonnes of freight travel on more than 10,000 trains on the SBB rail network [15]. Heavy snowfall on 22 November 2024 led to a punctuality negative record [20]. Good average weather over the year positively affected punctuality [20].

Comparison of 2024 headline punctuality by network, period and scope.

Network–period Headline figure Unit and scope
Great Britain, FY 2024-25 [10] 84.8% within three minutes [10] mainline network in Great Britain [9]; surface rail only [10]
DB Long-Distance, calendar 2024 [5][6] 62.5% [5] percent [5]
Punctuality (whole journey) (DB Long-Distance), calendar 2024 [5][6] 67.4% [5][6] percent [6]
DB rail passenger transport in Germany, calendar 2024 [5][6] 89.5% [5] percent [5]; in Germany [5]
SBB, in 2024 [17] 93.2% [17] SBB trains [17]

Taken together, [8][19][18] and [10] indicate the following sub-annual and historic figures cover different periods than the 2024 headlines. Deutsche Bahn reports 62.7% DB Long-Distance punctuality in H1 2024 [8]. Deutsche Bahn reports 90.8% DB Regional punctuality in H1 2024 [8]. DB Long-Distance punctuality was 62.5% in full-year 2024, compared with 62.7% in H1 2024 and 63.4% in H1 2025, Deutsche Bahn reports [19]. IamExpat reports 64% of Deutsche Bahn long-distance trains arrived on time in 2023 and 81.8% in 2020 [18]. Tagesschau reports Fernverkehr (long-distance) punctuality of 62.5% in the past year was a historic low [13]. Blick reports 94.1% of SBB trains reached their destination on time in the past year, 0.9% better than 2024 [16]. Rail Trends 2025 reports 92.9% of trains arrived within three minutes in FYE March 2021, a temporary improvement driven by reduced services and passenger numbers during COVID-19 [10]. In the reported August, DB Fernverkehr operational punctuality was 54.2% while long-distance passenger punctuality was 56.6%, Deutsche Bahn reports [3].

The UK Rail Factsheet defines On Time as early or within 59 seconds of scheduled time [2]. The Office of Rail and Road defines Time to 3 as the percentage of recorded station stops arrived at within three minutes [9]. For every performed train/bus run, Deutsche Bahn compares target arrival time to actual arrival time [5]. It summarises arrivals on schedule or up to a defined maximum delay in the punctuality rate [5]. Performance is judged against the plan of the day confirmed at 22:00 on the previous evening, the Office of Rail and Road reports [1]. The share of fully or partially cancelled trains further increased, the Bundesnetzagentur reports [14]. In the past year, 4% of trips in long-distance transport were cancelled without replacement, Deutsche Bahn reports [21]. The latest data should be treated as provisional, the Office of Rail and Road reports [1]. Periodic operational data from Network Rail are typically made available 10 to 12 working days after the period ends once loaded and validated, the Office of Rail and Road reports [11]. Further punctuality figures can be found on the SBB statistics portal [15]. More detailed data, including historic punctuality figures, can be found on the Office of Rail and Road website [12].

3.2 Delay Thresholds and Punctuality Definitions Compared

RailwayPro reports that in Switzerland a train is considered on time if it arrives within three minutes of its scheduled arrival, contrasting with Germany (six minutes) [17].

Comparison of delay thresholds and timing points across Great Britain, Germany and Switzerland.

Measure Threshold Where measured
Great Britain On Time lateness less than 1 minute (to 59 seconds) against GBTT in TRUST, early trains counted [23]; early or less than one minute after scheduled time, early classified as on time [24] lateness at a particular Recorded Station Stop against GBTT in TRUST [23]; assessed at the origin, at arrival at each intermediate station and at termination [23]
Great Britain Time to 3 early or less than three minutes after scheduled time [24] percentage of recorded station stops arrived at within three minutes [9]
Great Britain Time to 15 lateness less than 15 minutes (to 14 minutes 59 seconds) against GBTT in TRUST, early trains counted [23] percentage of recorded station stops arrived at early or less than 15 minutes after scheduled time, cumulative with On Time and Time to 3 [24]
Great Britain PPM / Right Time PPM: early or less than five minutes after scheduled time for London and South East, Regional and Scotland operators, or less than ten minutes for Long Distance operators [24]; Right Time: arrive at destination within 59 seconds of the scheduled public arrival time at the destination as shown in the Applicable Timetable [23] PPM: arriving at terminus having called at all planned stops [23]; Right Time: punctual arrival at terminus only, not at each station on journey [23]
Germany operational passenger: scheduled arrival time exceeded by less than 6 minutes, maximum 5:59 [3]; freight: scheduled arrival time exceeded by less than 16 minutes [6] share of punctual stops out of all intermediate and terminal stops [3]; share of punctual stops among all intermediate and final stops in Germany [6]
Germany traveller maximum 14:59 minutes versus planned arrival [3] arrival of passengers at the booked destination station taking into account working or missed connections, cancellations, replacement trains, timetable changes and alternative connections [3]
Switzerland considered on time if it arrives within three minutes of its scheduled arrival [17] considered on time if it arrives within three minutes of its scheduled arrival [17]; SBB considers a train delayed if it arrives three minutes after schedule [22]

See the comparison above for the specific On Time, Time to 3 and Time to 15 thresholds. Generic Time to X needs lateness of that Recorded Station Stop of less than X minutes, for example within 2 minutes 59 seconds for Time to 3, as measured against the Public Timetable time in TRUST, with early trains counted [23]. Punctuality is measured against the public-facing passenger timetable (GBTT) [23]. From 1 April 2019, the UK rail industry introduced a new, more detailed and precise set of measures to better understand how timely trains are from station to station and from minute to minute [12]. Origins are included because departing on time from origin is important, just as arriving punctually at all advertised station stops is [23].

For the PPM terminus thresholds and calling-point scope, see Background [23][24]. For three of the open access operators (Hull Trains, Grand Central and Lumo), it is less than ten minutes, while for Heathrow Express services it is less than five minutes [26]. To count as Right Time, the train must arrive at destination within 59 seconds of the scheduled public arrival time at the destination as shown in the Applicable Timetable, which is a train measure accounting for punctual arrival at terminus only – and not at each station on its journey [23]. Where a train fails to stop at one or more booked calling points on the journey, the train is considered to have failed PPM [1].

Request stops are only included when they are activated and an actual arrival time is recorded [23]. Stations which are pick-up only or drop-down only are included within the measure because there are planned and actual date times within TRUST for them [23]. Only events with passenger activity should be included in the punctuality measures (e.g. pick up, drop off etc.) [23]. TRUST does not currently capture information at all platforms at all stations, and it is anticipated that over time more platforms and stations will be added as recording points in TRUST [23]. The official statistics only include a range from 'On Time' to 'Time-to-30', and figures can be quoted for 'Early' and '30 minutes or greater' to capture station stops outside of this range [23]. The figures quantify how punctually trains are reaching stations, from ‘early’ through to ‘30 minutes after’ the timetable [12]. The Metro measures are not measured against the Applicable Timetable [23]. They are only appropriate in parts of the network where the service frequency is at least 10 trains per hour [23]. Network Rail labels train punctuality at station stops as the new industry measure [12].

Average Passenger Lateness (APL) measures the average lateness of a passenger as they alight from their train [24]. It is estimated for each train by multiplying the number of passengers expected to alight at main stations by the punctuality to the nearest minute at those stops [24]. PM is initially calculated on arrival at each Monitoring Point [23]. The definition refers to the number of minutes (rounded down to the nearest whole minute) by which the time at which the train stops at the Monitoring Point is later than the time at which that train is scheduled [23]. PM at the Monitoring Point is the sum of lateness plus cancelled trains multiplied by cancellation minutes, divided by the number of planned arrivals, where cancellation minutes are based on 1.5 times the service frequency [23].

The Federal Network Agency calculates punctuality for passenger services from punctually or delayed served stops, counting every stop rather than only the final stop [25]. Under the EU norm a passenger train is punctual if it runs no more than five minutes behind schedule while the limit for freight trains is 15 minutes, although DB Netz publishes using 5:59 minutes and 15:59 minutes and the market investigation uses the EU norms [25]. National rail freight transport remained markedly more punctual than cross-border freight trains, with international freight trains recording only 42 percent punctual arrivals compared with over 62 percent punctually reaching their destination for freight trains operating within Germany [25].

Traveller punctuality records the arrival punctuality of all travellers across the entire journey chain in long-distance transport [6], as detailed for threshold and destination in the comparison table above. DB Fernverkehr determines this traveller measure [3]. That threshold is the same as for long-distance bus and air travel [3]. It accounts for kept or missed connections, cancellations and replacement trains, timetable changes and alternative connections [21]. Deutsche Bahn distinguishes standard DB Long-Distance punctuality from a separate whole-journey punctuality measure for DB Long-Distance [5].

Deutsche Bahn defines Beginnpünktlichkeit as the share of departures at starting stations with a deviation of less than 6 minutes [6]. Deutsche Bahn reports that scheduled train starts and the neutralization of delay effects from the hubs are extremely important for relieving the burden on the hubs, with the focus on the provisioning process and increasing on-schedule departures from the starting station as key drivers for overall punctuality [19]. Trains with a flex departure (FlexAbfahrt) can leave the platform earlier if required, without changing the official departure time for passengers [5]. Bus stops are included in punctuality measurement only when real-time data are available, and bus final stops are not included [6]. At the Berlin S-Bahn, an arrival or departure deviating by more than two minutes from the timetable counts as delayed [27].

Taken together, the Swiss within-three-minutes on-time definition [17] and the Germany (six minutes) figure [17] indicate Switzerland allows half Germany's lateness. A train is considered on time if it arrives within three minutes of its scheduled arrival [17]. Trains already count as delayed from three minutes in Switzerland [27], with SBB considering a train delayed if it arrives three minutes after schedule [22]. In Switzerland long-distance and regional services are practically equally punctual, in contrast to Deutsche Bahn where a gap exists [28]. The source reports connection punctuality ‘hier’ at 98.9%, above the Swiss average [16].

Arrival delay and departure delay are different events. A late arrival need not cause a late departure and a late departure need not cause a late arrival elsewhere because delays can be recovered [27]. Early arrival generally does not produce early departure because passenger trains wait until scheduled departure [27]. Intermediate points matter. A journey perceived as punctual between two stations may still have early or late running at block posts and overtaking stations [27]. Unpunctuality can be measured during the journey or only at journey end [27]. Rail operations define punctuality as agreement of planned or published departure and arrival times with actually achieved times [27]. Problems can arise at timetable-listed operating points, not only passenger-visible stations and halts [27]. Mere delays exclude early running, while unpunctuality includes it [27]. Common usage does not call early arrival unpunctual because most passengers do not find it disruptive [27].

Delay minutes are defined as the time lost between consecutive timing points on the rail network, with delay incidents producing three or more minutes of delay on Britain's railways attributed to either Network Rail or a train operator [4]. Network Rail's chart shows who was responsible for passenger train delays of 3 minutes or more [12]. The incentive-system dataset records delay events leading to an additional delay of at least two minutes for the train journey/train run, with events from 90 seconds recorded in the incentive system but rounded to whole minutes [25]. One Lost Unit is a punctuality-relevant delay of a train journey with a delay from 90 seconds, with each case forming one Lost Unit so that multiple Lost Units can arise during one trip [6]. For regulatory reasons, Network Rail is assigned responsibility for delays caused by external factors such as weather, trespass, vandalism, cable theft and fatalities [12]. Full details of the delay attribution rules can be found on the Delay Attribution Board's page [12]. Trains arriving 120 minutes or more late at destination do not count towards the cancellation score [23]. SBB turns trains at the border when necessary and uses replacement trains within Switzerland [16]. Only about 2 percent of long-distance trains are affected by delays over 60 minutes; in individual cases it may then be operationally sensible to end a train journey early to enable passengers a fast change to the following interval long-distance train [21]. She said when trains were delayed and trying to make up time to reach bigger stops like nearby Gatwick, stops at smaller stations like hers were cancelled [29].

3.3 Cancellations and Reliability Denominator Effects

  • Part-cancellations cost half a full cancellation in Great Britain [1].

  • Full cancellation: running less than 50% of planned mileage, including not running at all, weight 1 [23].

  • Part cancellation: running at least 50% but less than 100%, or completing mileage but missing one or more planned stops, weight 0.5 [23]. Any train missing one or more planned stops counts as cancelled; origin and terminus always count [23].

  • ORR defines score on trains planned, full as one and part as half [1]. Summed score becomes percentage of trains planned in Applicable Timetable [23], normally to one decimal place [23]. Rate covers planned trains not running full journey or not calling at all planned stops [23]. ORR notes includes missing stations or not reaching destination [1].

Cancelled trains automatically fail Britain's train-based success measures [23]. No cancellation passes [23]. The Rail Delivery Group counts any full or part cancellation as a Right Time failure against trains planned at the terminus [23]. The same rulebook counts any total or other cancellation as not punctual in PPM on the same trains-planned denominator [23]. For CaSL, any part or full cancellation enters as a failure alongside arrivals 30 minutes or more late at the final destination, according to the Rail Delivery Group [23]. CaSL itself is cancelled or 30-or-more minutes late at the final scheduled destination, according to the Rail Delivery Group [23]. ORR states Average Passenger Lateness incorporates lateness caused by cancelled trains [1]. Threshold detail is in Delay Thresholds and Punctuality Definitions Compared.

Planned stops where the train fails to stop at the location are deliberately excluded as they are captured in the reliability measures [23]. For the PPM rule for missed stops, see Delay Thresholds and Punctuality Definitions Compared. A Recorded Station Stop is defined as a location in the applicable timetable with both a planned date time (GBTT) and an actual recorded date time (according to TRUST) [23]. A recorded station stop is defined as a location with both a planned timetable time and an actual recorded time where a train has stopped [1]. Up to around 96% of all station stops are currently recorded, with no estimates made for punctuality at the c.4% of station stops not recorded [31]. Up to around 90% of all station stops are currently recorded, with no estimates made for punctuality at the c.10% of station stops not recorded [1]. Taken together, [31] and [1] indicate different reported coverage levels of up to around 96% versus up to around 90% [31][1]. Definition detail is in Delay Thresholds and Punctuality Definitions Compared.

This industry measure is an indicator of disruption against the timetable operating on the day. The timetable is finalised at 22:00 the previous evening, and trains removed from the timetable before then will not be included [31][32]. This industry measure is an indicator of disruption against the timetable operating on the day, and a lower cancellations score indicates better reliability [1]. Strike action by the railway unions took place on seven days in the latest quarter, with a reduced timetable put in place on the strike days and on some of the days after, and the Cancellations score only takes account of trains cancelled from the planned reduced service [1]. P*-coded pre-cancelled trains are not included, and on days with strike action the Cancellations score only reflects trains cancelled from the reduced timetable [4]. A severely disrupted day at a national (GB) level is defined when the cancellations score is 5% or more, and at a sub-operator level when the cancellations score for any sub-operator is 20% or more [1]. Nationally, there were nine severely disrupted days in the latest quarter (1 April to 30 June 2024) [4].

Operators who use “P*-coding” may therefore have a lower Cancellations score reported in this release than that which a passenger may experience, according to ORR [1]. Pre-cancelled trains are removed from the timetable before it is finalised and therefore may not appear in operators’ Cancellations percentages, according to ORR [24]. Those are changes to train services caused by non-availability of staff or rolling stock that are included in a revised timetable, according to ORR [11]. The ‘adjusted Cancellations score’ was calculated by combining the official Cancellations score data with the resource availability shortage pre-cancellations data [1]. The adjusted cancellations score includes pre-cancellations (p-codes) due to resource availability, that may not be included in the official cancellations score [32]. In the latest quarter, 13 operators reported the use of “P*- coding” for resource availability shortage pre-cancellations [1]. Of these, Grand Central had the highest adjusted Cancellations score at 19.7% [1]. In the latest quarter (1 April to 30 June 2024), fourteen operators reported the use of “P*-coding” for resource availability shortage pre-cancellations [4]. Based on the data we collected, the adjusted Cancellations score for the latest quarter was 4.1%. This was 0.6pp higher than the official Cancellations score [1]. Based on the data we collected, the adjusted Cancellations score for the latest quarter (1 April to 30 June 2024) was 4.0%. This was 0.5pp higher than the official Cancellations score [4]. Based on the data we collected, the adjusted Cancellations score to include the trains removed from the timetable due resource availability shortages in Great Britain was 3.8%. This was 0.5pp higher compared to the official Cancellations Score [26]. Fifteen operators had reported the use of P-coded pre-cancellations, of which TransPennine Express had the largest adjusted Cancellations score of 20.5% [26]. The ad hoc release and associated data table presents data for both on the day cancellations and resource availability shortage “P*-coded” pre-cancellations by cause for each operator [11]. The dashboard, data table and historic factsheets present data for both on the day cancellations and resource availability shortage “P*-coded” pre-cancellations by cause for each operator [32].

3.8% of trains were classified as cancellations in Great Britain in the FYE March 2024, which is unchanged from the previous year, but remains the joint highest annual cancellations score since records began [2]. The cancellations score in Great Britain in 2024-25 was 4.1%, a 0.3pp increase compared with the previous year [10]. The cancellations measure is a weighted score which counts full cancellations as one and part cancellations as half [10]. Cancellations fell to their lowest level since 2014-15, reaching 2.1% in 2020-21 due to the COVID-19 pandemic, when fewer services were scheduled [10]. As the number of trains planned continue to increase as rail travel recovers from the pandemic, cancellations have also increased steadily to a record high in the latest year [10]. In the latest quarter (1 April to 30 June 2024), of the 1.8 million trains planned, 45,900 were full cancellations and 36,500 were part cancellations [4]. The Cancellations score for the 12 months up to September 2023 (1 October 2022 to 30 September 2023) was 3.6% [1]. This was down 0.1pp (i.e. better) compared with the previous 12 months ending September 2022 [1].

See Delay Thresholds and Punctuality Definitions Compared for stop-based denominator detail and traveller-punctuality definition detail. Punctuality in rail transport for 2019–2023 is shown without train cancellations, while the share of runs cancelled on partial sections or fully is recorded additionally, with data collected directly from the infrastructure managers [25]. The share of Nahverkehr trains cancelled on partial sections increased to just under eight percent, so that additionally almost every twelfth run did not reach its planned destination or was only started at a planned intermediate stop [25]. It is described as completely incomprehensible from the travellers' perspective when cancelled trains do not enter the punctuality statistics [35].

Comparison of denominator treatments for cancellations and skipped stops.

System Denominator How skipped/cancelled services enter
GB Cancellations score trains planned in Applicable Timetable [23] full as 1, part as 0.5 [1]; includes missing stations or not reaching destination [1]
GB On Time stop punctuality recorded stops with planned and actual times where train stopped [1] failed stops excluded, kept in reliability [23]; up to ~96% stops recorded [31]
GB train reliability (PPM, Right Time, CaSL) trains planned [23] cancellations count as failures [23]; CaSL adds 30+ min late at destination [23]
German operational punctuality stop-based denominator described above [3] cancellations recorded separately as noted above [25]
BBC stop analysis scheduled stops from 13 ORR periods [29] cancelled stops over scheduled stops [29]; about 3m of 89m cancelled [29]

A heavily delayed train burdens the overcrowded network, so dropping part of its route can be sensible if passengers have alternatives, according to Deutsche Bahn [35]. Cancelling heavily delayed trains was described as operationally sensible when it limits impacts on other trains and complies with labour-law requirements for train crew [35]. Most passengers could use alternative connections and continue their journey, including when trains did not end at the destination station, according to Deutsche Bahn [21]. DB states that night-time transfer and stabling trips cause no cancellations for passengers [21]. Without strike effects, the share of unplanned train cancellations in regional transport would be around 3% and thus below those of 2023 [21]. A decline in vehicle faults and higher vehicle availability had a positive effect [6]. In long-distance services, lower punctuality led to fewer passengers [13]. A similar justification was reportedly given for a train cancellation on 11 September [35].

BBC analysis calculated the percentage of cancelled stops out of those scheduled for the last 13 periods of data published by the Office of Rail and Road [29]. The cancellation rate for the past year was still slightly higher than the previous one - 3.3%, up from 3.2% [29]. Out of all 2,549 stations, 78 were excluded from the analysis due to missing or unreliable data [29]. City Thameslink in London has been named the worst of Britain's busiest railway stations for cancellations in the year to August [29]. About one in 13 of almost 150,000 scheduled stops there were cancelled during that time, according to Office of Rail and Road (ORR) figures analysed by BBC [29]. The data shows that, in the past year, one in 24 planned stops ended up being cancelled in Macclesfield [29]. Earlswood had the sixth worst cancellation rate of any train station in Britain in the past year, excluding a small number of stations with unreliable or incomplete data [29]. The number of cancelled train stops has gone down in the most recent four-week period, when compared to the period just before or the same period last year [29]. The DfT said it was already starting to see 'positive signs of progress with overall cancellations starting to fall' [29]. 'At times we have to make the very difficult decision to cancel some trains to get the timetable back on track' [29]. Train operators 'spare no effort to maintain as many services as possible' but extreme weather, infrastructure faults and trespass could have an impact [29]. This differs from the cancellation analysis by the Rail Delivery Group, which represents Network Rail and train operators [29].

Latest Network Rail TRUST data should be treated as provisional because operator-provided information e.g. on cancellations can be updated over time, according to ORR [31]. Due to the way this data is rounded the total delay responsibility figure can sometimes total 99.9%, a system generated figure that cannot be altered, according to Network Rail [12]. The data.gov.uk dataset is alternatively titled Train punctuality and reliability [30]. Administrative data may be incomplete when collected for a different operational purpose or because of delays in updating the administrative system, according to statistics-regulator guidance [34]. The Swedish RB found that a cancelled stop, where a train follows its planned route but does not stop at a station in accordance with the train plan, is a deviation from the train plan to be included in the charging scheme, which it previously was not [33]. A landslide at St. Gallen St. Fiden impaired the robustness of the system, according to 20 Minuten [20]. No Swiss denominator rule is described here. Comparison uncertainties are in Uncertainties Limiting Direct Cross-Country Comparisons.

3.4 Uncertainties Limiting Direct Cross-Country Comparisons

ORR directs users to its Revisions log for information on any revisions [36]. ORR states that details on how these statistics are compiled are in its quality and methodology report and that information on any revisions is in its Revisions Log [11]. When producers report on statistics that are subject to revision, the details are sometimes included only in the methodology reports and it is not always clear in the bulletins that figures are ‘reported’ incidents or events and therefore may be subject to a degree of regular revision [34]. ‘Rough and ready’ estimates produced at speed usually with a higher degree of uncertainty stand against statistics that are less timely but perhaps more certain, and communicating this distinction and the scale of any revisions is also important, according to OSR [34].

ORR states that its statistical practice is regulated by the Office for Statistics Regulation (OSR) [9][36]. The PPM dataset page lists its designation as National Statistics [30]. OSR produces guidance that draws out relevant areas from the Code of Practice in relation to Changes in statistical methods and also one that looks at where a change in data quality could mean a change in the National Statistics Status [34].

According to the European Commission's RMMS pages, since 2007, the European Commission has collected data on the evolution of rail markets in the Member States through RMMS (Rail Market Monitoring Scheme) questionnaires [38]. This data collection was initially based on a 'gentlemen's agreement', which entailed certain limitations, including possible data gaps and methodological inconsistencies [38]. The recast of the First Railway Package (Directive 2012/34/EU) eventually established a formal legal basis for RMMS reporting and for the harmonisation of data [38]. The Directive also allows the Commission to adopt measures to ensure consistency in the reporting obligations of the Member States [38]. The Commission adopted Implementing Regulation (EU) 2015/1100 in July 2015, setting out detailed reporting obligations for the Member States under the rail market monitoring framework [38]. Since 2016, Member States and Norway have submitted data in accordance with the structure and content prescribed by this Regulation [38].

The Act puts an obligation on all market players to participate in the Bundesnetzagentur's market survey, which eventually might be enforced by imposing administrative fines [37]. In July 2025, the European Commission presented its ninth report on developments in the rail market [38]. While the report gives a high-level overview of the main developments, the full analysis is presented in the accompanying Commission Staff Working Document [38]. Data and figures are also available in an Excel format [38]. The source lists “Reference metadata on methodology and quality of RMMS data” and “RMMS data quality assessment and management” [38].

The average value in this graph differs from the one in the Main Report since the sample is different [33]. This one includes all available data for 2021 (29 countries) while in the Main Report it includes only the countries which provided data for the 2017-2021 period (26 countries) [33]. The IRG-Rail Working Document notes that the total TAC is a proxy of the sum of TAC from railway undertakings and TAC from public subsidies [33]. Other subsidisation schemes such as direct State aids to the infrastructure managers in compensation of discounts applied to the TAC may be excluded from this indicator [33]. There might be other countries whose governments funded parts of TAC via direct or indirect COVID-19 financial support measures, but data were not available at this level [33].

Heavier trains can especially be found in some countries where the most common goods transported are iron ore or coal, while lighter trains prevail in countries with a high share of intermodal transport, the IRG-Rail Working Document records [33]. The typology of goods transported by freight services per country may consequently affect the average indicator of revenue per net tonne-km, the IRG-Rail Working Document notes [33]. All financial indicators are presented in Euro for all countries, but without inflation adjustment, which could partially explain the changes observed in countries, according to the IRG-Rail Working Document [33]. Railway undertakings may use different approaches to calculate the transport costs they charge their clients (giving more emphasis to the weight of goods or to the transport distance), which alters the calculation of freight revenue crucially, according to the IRG-Rail Working Document [33].

OSR found cases where comparisons had been made over time or between countries where the data were not comparable [34]. WIMD can’t be used for comparing with other UK countries – each country measures deprivation slightly differently [34]. OSR describes producers using infographics or introductory bullet points to set out clearly what the data can and can’t be used for, including highlighting where uncertainty associated with different breakdowns may affect the ability to use the data in a certain way especially where comparisons are being made [34].

Quantitative measures such as confidence intervals are sometimes published, and are sometimes presented in isolation without any supporting guidance or context, which can lead to the reported statistic being mistakenly interpreted by some users as the ‘true’ and only value, rather than an estimate that is likely to fall within a range of possibilities [34]. It can be helpful to use terms such as ‘around’, ‘nearly’ or ‘about’ when presenting figures: these terms can help to inform the reader that the figures are not exact and therefore carry a level of uncertainty [34]. In general, the presentation of uncertainty in data tables was found to be a particular weakness, even where uncertainty in the corresponding statistical bulletin had been presented well, with many data tables not clarifying that the statistics are estimates or referencing uncertainty at all [34]. There is wide use of rather loose statements concerning uncertainty such as ‘figures should be treated with caution’; while these caveats can play an important role in highlighting that limitations exist, in some cases these statements are not specific enough to be helpful and can be overused to the extent that they become lost [34].

The Winton Centre also found that any graphical representation (error bars, fan charts, diffusion plots) of uncertainty around time series maintained trust in both the numbers and the producer of the statistics [34]. Fan charts, such as those used by the Bank of England, are a helpful way to communicate inherent uncertainty, and presenting variant projections is also helpful for demonstrating the range of possible values of a projection [34].

Survey uncertainty can arise from survey data and can be driven by survey sample size and its bias, including differential non-response or poor survey design, according to OSR [34]. If linkage rates for a certain age group or ethnic group are higher than in other groups, then that group could be more represented in the analysis and could skew the results [34]. Bias might occur through changes to the variables collected in the administrative dataset or decisions around the inclusion of records with missing data, where a decision to omit these records from the analysis could introduce a bias in the statistics [34]. Where producers had published information on the quality assurance processes and the risks associated with using administrative data, such as misreporting, users were more aware of how to use and correctly interpret the statistics, OSR found [34]. Sometimes, the underlying assumptions in the modelling have not been made clear enough or described in an accessible way to allow users to fully understand the implications on the statistics [34].

4. Discussion

For the definitions applied in this discussion, see Background; for the headlines on their own rulers, see the overview below.

The GB cancellations and timetable handling and the Deutsche Bahn operational and traveller measurement rules used in this comparison are set out in full in Background and are not repeated here.

  • For headline figures on own rulers with thresholds/denominators, see Background and Findings and overview (§3.1).

  • [25] reports Gesamt 1.152 Mio. Trkm, SPFV 164 Mio. Trkm, SPNV 722 Mio. Trkm and gesamte Betriebsleistung 2023 ca. 1.151 Mio. Trassenkilometern [25].

The overview below aligns rulers without ranking them.

A press summary defines a Deutsche Bahn long-distance train as “on time” when it arrives at its destination no more than six minutes after the scheduled arrival time [18]. Regarding the heavily delayed ICE 616 from Munich to Hamburg, which ended early in Cologne, the wording used was: “Zug fällt zur Verbesserung der Statistik ab Köln aus” [35], while Deutsche Bahn denied manipulating figures, calling the employee’s internal wording false [35]. Definitions and cancellation handling are covered in background and §3.4.

Due to differences in how passenger rail performance is measured in other countries, opportunities to make direct comparisons with statistics in this release are limited [1][24][31]. Taken together, [17] and [10] indicate the reporting periods differ as In 2024 versus coverage for the financial year (FY) 2024-25 [17][10]. The timetable is finalised at 22:00 the previous evening and trains removed from the timetable before then will not be included [4], while to measure punctuality the target arrival time is compared to the actual arrival time for every performed train/bus run [5]. Taken together, [4] and [5] indicate timetable handling differs between exclusion before 22:00 and inclusion of every performed run [4][5].

Up to around 90% of all station stops are currently recorded, and no estimates have been made for punctuality at the c.10% of station stops not recorded [1]. Up to around 96% of all station stops are currently recorded, and no estimates have been made for punctuality at the c.4% of station stops not recorded [4]. Taken together, [1] and [4] indicate the recorded share differs between the two sources.

5. Conclusion

Taken together, SBB's 93.2% of trains arriving on time in 2024 [17], where in Switzerland a train is considered on time if it arrives within three minutes of its scheduled arrival [17], ORR's 84.8% of trains arrived within three minutes (early or less than three minutes after the scheduled time) in 2024-25 [10], where Time to 3 is the percentage of recorded station stops that were arrived at early or less than three minutes after the scheduled time [24], and DB Long-Distance punctuality of 62.5% in 2024 [5], measured by comparing target arrival time to actual arrival time for every performed train/bus run [5] with a passenger stop counted as operationally punctual when scheduled arrival is exceeded by less than 6 minutes among all intermediate and final stops in Germany [6], indicate that opportunities to make direct comparisons with statistics in this release are limited due to differences in how passenger rail performance is measured in other countries [1].

Taken together, [17][6][24] indicate punctuality thresholds differ. DB rail passenger transport in Germany punctuality was 89.5% in 2024 [5], while Public Performance Measure (PPM) is the proportion of trains arriving at their final destination early or less than five minutes after the scheduled time for London and South East, Regional and Scotland operators, or less than ten minutes for Long Distance operators [4].

Punctuality is presented without train cancellations, while the share of partially or fully cancelled runs is recorded additionally [25]. The cancellations measure is a weighted score which counts full cancellations as one and part cancellations as half [10], and it only includes trains cancelled after 22:00 the previous evening and trains removed from the days’ timetable before then may not be included [10]. In 1 April to 30 June 2024 the adjusted Cancellations score was 4.0%, 0.5pp higher than the official Cancellations score, with fourteen operators reporting the use of “P*-coding” for resource availability shortage pre-cancellations [4]. For Reisenden-Pünktlichkeit train cancellations are also taken into account [21]. Up to around 96% of all station stops are currently recorded, with no estimates made for punctuality at the c.4% of station stops not recorded [4]. The latest data should be treated as provisional, as train operators provide Network Rail with information e.g. on cancellations, which can be updated over time [31].

The table below follows.

reader scenario recommended choice deciding factor
Use SBB, Great Britain and DB reported punctuality figures [17][10][5] Taken together, [17][24] and [6] indicate keep each figure in its own definition DB counts a stop as operationally punctual when scheduled arrival in freight transport is exceeded by less than 16 minutes [6]
Compare DB long-distance with British Long Distance PPM [5][4] Taken together, [5][6] and [4] indicate analyse DB long-distance apart from British Long Distance PPM Great Britain PPM is the proportion of trains arriving at final destination early or less than ten minutes after the scheduled time for Long Distance operators [4]
Judge reliability with cancellations [25][21][10][4] Taken together, [25][21][10] and [4] indicate pair punctuality rates that exclude cancellations with scores that include cancellations DB traveller punctuality includes train cancellations [21]

Taken together, [17][24][4] and [6] indicate punctuality thresholds differ: in Switzerland a train is considered on time if it arrives within three minutes of its scheduled arrival [17], Great Britain's Time to 3 is the percentage of recorded station stops that were arrived at early or less than three minutes after the scheduled time [24] while Public Performance Measure (PPM) is the proportion of trains arriving at their final destination early or less than five minutes after the scheduled time for London and South East, Regional and Scotland operators, or less than ten minutes for Long Distance operators [4], and a passenger stop in Germany is counted as operationally punctual when scheduled arrival is exceeded by less than 6 minutes among all intermediate and final stops in Germany [6]. Taken together, [25] and [10] indicate the stated cancellation handling: the figure labelled 'Pünktlichkeit im Eisenbahnverkehr (2019-2023; ohne Zugausfälle; Quoten in Prozent)' excludes train cancellations, with the share of partially or fully cancelled runs recorded additionally [25], while the British cancellations measure only includes trains cancelled after 22:00 the previous evening and trains removed from the days' timetable before then may not be included [10]. Taken together, [17][10][5][25] and [1] indicate direct comparison is limited: each publisher reports arrival punctuality in percent [17][10][5], a passenger train running no more than five minutes behind schedule is considered punctual under the EU norm [25], yet opportunities to make direct comparisons with statistics in this release are limited due to differences in how passenger rail performance is measured in other countries [1]. Directive 2012/34/EU established a formal legal basis for RMMS reporting and for the harmonisation of data [38].

Deutsche Bahn plans to lift long-distance punctuality to 75-80% by 2027 [13].

References

[1] https://dataportal.orr.gov.uk/media/1kdfjw4u/performance_stats_release_jul-sep-2023.pdf — https://dataportal.orr.gov.uk/media/1kdfjw4u/performance_stats_release_jul-sep-2023.pdf · government [2] Rail factsheet: 2024 — https://www.gov.uk/government/statistics/rail-factsheet-2024/rail-factsheet-2024 · government [3] Erläuterung Pünktlichkeitswerte für den August 2026 — https://www.deutschebahn.com/de/konzern/konzernprofil/zahlen_fakten/puenktlichkeitswerte-6878476 · general [4] https://dataportal.orr.gov.uk/media/ocib4lie/performance_stats_release_2024-25_q1.pdf — https://dataportal.orr.gov.uk/media/ocib4lie/performance_stats_release_2024-25_q1.pdf · government [5] Deutsche Bahn Annual Report 2024 — https://ibir.deutschebahn.com/2024/en/combined-management-report/product-quality-and-digitalization/the-customer-is-at-the-center-of-our-actions/punctuality/ · general [6] https://ir.deutschebahn.com/fileadmin/Deutsch/2025/DB_IB24_d_web_01.pdf — https://ir.deutschebahn.com/fileadmin/Deutsch/2025/DB_IB24_d_web_01.pdf · general [7] Pünktlichkeit Deutsche Bahn 2024 | Deutsche Bahn Integrierter Bericht 2024 — https://ibir.deutschebahn.com/2024/de/zusammengefasster-lagebericht/produktqualitaet-und-digitalisierung/kunde-im-mittelpunkt-unseres-handelns/puenktlichkeit/ · general [8] Deutsche Bahn Interim Report 2024 — https://zbir.deutschebahn.com/2024/en/interim-group-management-report-unaudited/product-quality-and-digitalization/punctuality/ · general [9] Passenger rail performance | ORR Data Portal — https://dataportal.orr.gov.uk/statistics/performance/passenger-rail-performance/ · government [10] Rail Trends 2025 — https://www.gov.uk/government/statistics/rail-trends-2025/rail-trends-2025 · government [11] Passenger rail performance | Office of Rail and Road — https://www.orr.gov.uk/node/2935 · government [12] Railway Performance - Network Rail — https://www.networkrail.co.uk/who-we-are/how-we-work/performance/railway-performance/ · government [13] Deutsche Bahn mit Milliardenverlust und so unpünktlich wie noch nie — https://www.tagesschau.de/wirtschaft/unternehmen/deutsche-bahn-milliarden-verlust-unpuenktlich-100.html · professional [14] Bundesnetzagentur - Marktuntersuchungen — https://www.bundesnetzagentur.de/DE/Fachthemen/Eisenbahnen/Veroeffentlichungen/Marktuntersuchungen/start.html · government [15] How punctual is SBB? | SBB — https://company.sbb.ch/en/company/responsibility/customers/punctuality.html · general [16] Neuer Rekord: SBB waren 2025 so pünktlich wie noch nie — https://www.blick.ch/wirtschaft/nur-5-prozent-der-zuege-zu-spaet-sbb-waren-letztes-jahr-so-puenktlich-wie-noch-nie-id21636638.html · professional [17] SBB achieves record-breaking punctuality in 2024 — https://www.railwaypro.com/wp/sbb-achieves-record-breaking-punctuality-in-2024/ · professional [18] Over a third of Deutsche Bahn long-distance services delayed in 2024 — https://www.iamexpat.de/expat-info/germany-news/over-third-deutsche-bahn-long-distance-services-delayed-2024 · professional [19] Deutsche Bahn Interim Report 2025 — https://zbir.deutschebahn.com/2025/en/interim-group-management-report-unaudited/quality-and-security/punctuality/ · general [20] SBB: Noch nie so pünktlich wie 2024 — https://www.20min.ch/story/932-prozent-rekordwert-die-sbb-war-noch-nie-so-puenktlich-wie-2024-103268575 · professional [21] Newsblog — https://www.deutschebahn.com/de/presse/Newsblog-12829716 · general [22] Deutsche Bahn punctuality improves slightly — https://www.iamexpat.de/expat-info/germany-news/deutsche-bahn-punctuality-improves-slightly · professional [23] https://www.raildeliverygroup.com/publications/acop/287-rdgntfdorpm-final/file.html — https://www.raildeliverygroup.com/publications/acop/287-rdgntfdorpm-final/file.html · professional [24] https://dataportal.orr.gov.uk/media/4pmhex4x/performance-stats-release-jan-mar-2026.pdf — https://dataportal.orr.gov.uk/media/4pmhex4x/performance-stats-release-jan-mar-2026.pdf · government [25] https://data.bundesnetzagentur.de/Bundesnetzagentur/SharedDocs/Downloads/DE/Sachgebiete/Eisenbahn/Unternehmen_Institutionen/Veroeffentlichungen/Marktuntersuchungen/MarktuntersuchungEisenbahnen/MarktuntersuchungEisenbahnen2024.pdf — https://data.bundesnetzagentur.de/Bundesnetzagentur/SharedDocs/Downloads/DE/Sachgebiete/Eisenbahn/Unternehmen_Institutionen/Veroeffentlichungen/Marktuntersuchungen/MarktuntersuchungEisenbahnen/MarktuntersuchungEisenbahnen2024.pdf · government [26] https://dataportal.orr.gov.uk/media/vlqfv4tv/passenger-performance-apr-jun-2023.pdf — https://dataportal.orr.gov.uk/media/vlqfv4tv/passenger-performance-apr-jun-2023.pdf · government [27] Pünktlichkeit (Bahn) — https://de.wikipedia.org/wiki/P%C3%BCnktlichkeit_(Bahn) · general [28] Trotz Baustellen: Die Bahn war 2025 so pünktlich wie nie. In der Schweiz — https://www.spiegel.de/mobilitaet/bus-bahn/db-vs-sbb-bahnen-so-puenktlich-wie-nie-in-der-schweiz-a-76ebd374-5dfa-4429-9a8a-82991533e37c · professional [29] Train cancellations: Britain's worst major stations named — https://www.bbc.com/news/articles/cgl1ze9ppdpo · professional [30] Performance - National Data Library — https://www.data.gov.uk/dataset/4748abbd-e497-498e-a3ee-f3db5ed819f7/public_performance_measure_ppm · government [31] https://dataportal.orr.gov.uk/media/ebmnxxih/performance-stats-release-jul-sep-2025.pdf — https://dataportal.orr.gov.uk/media/ebmnxxih/performance-stats-release-jul-sep-2025.pdf · government [32] coded cancellations | ORR Data Portal — https://dataportal.orr.gov.uk/statistics/performance/p-coded-cancellations/ · government [33] https://autorite-transports.fr/wp-content/uploads/2022/04/irg-rail_-_11th_mm_report_-_working_document.pdf — https://autorite-transports.fr/wp-content/uploads/2022/04/irg-rail_-_11th_mm_report_-_working_document.pdf · government [34] Approaches to presenting uncertainty in the statistical system - Office for Statistics Regulation — https://osr.statisticsauthority.gov.uk/publications/approaches-to-presenting-uncertainty-in-the-statistical-system/pages/3/ · government [35] Deutsche Bahn: Wurde die Pünktlichkeitsstatistik manipuliert? — https://www.zdfheute.de/wirtschaft/unternehmen/deutsche-bahn-verspaetung-puenklichkeit-statistik-zugausfall-manipulation-100.html · professional [36] Passenger rail usage | ORR Data Portal — https://dataportal.orr.gov.uk/statistics/usage/passenger-rail-usage/ · government [37] Bundesnetzagentur - Publications — https://www.bundesnetzagentur.de/EN/Areas/Rail/Companies/Publications/start.html · government [38] Rail Market Monitoring (RMMS) — https://transport.ec.europa.eu/transport-modes/rail/market/rail-market-monitoring-rmms_en · government

Source quality: 19 government, 10 professional, 9 general.