The Metro Tunnel has been in full service for nearly 6 weeks. Handily for data wonks, the 'Big Switch' happened on 1 February, giving us a full month's of performance data in the Track Record that has just come out.
That data describes the delivery and punctuality of the 18000 train services scheduled to run that month. It affects payments to the operator, and, if performance dips below a certain level, compensation to passengers.
That data describes the delivery and punctuality of the 18000 train services scheduled to run that month. It affects payments to the operator, and, if performance dips below a certain level, compensation to passengers.
Expectations
Contrary to expectations built up for the better part of a decade (most recently the premier's new timetable in place everywhere comment), the non Metro Tunnel lines didn't get much out of the February 1 'Big Switch' timetable.
The Craigieburn and Upfield line timetables were particularly underwhelming given that the Metro Tunnel's business case proposed an upgrade from every 20 to every 10 min off-peak, similar to what the Frankston line has had for over a decade. Craigieburn got a handful of extra trips while Upfield's was a 'rob Peter to pay Paul' job with trips shuffled and early evening frequency cut.
Transport Victoria presented these benefits here: https://www.facebook.com/reel/2274082129746422
Especially note this comments: "Craigieburn and Upfield lines will be getting a reliability boost with more space and less crowding in the city". We'll test that in a moment.
What about other lines?
Werribee, Williamstown and Frankston got new timetables with through-running removed and Frankston re-entering the City Loop. That may improve reliability numbers (as cross-city operation was always fairly fragile) though the bogey of loop bypasses is back for Frankston line passengers.
Sunbury, Pakenham and Cranbourne lines of course use the Metro Tunnel instead of the City Loop. The Sunbury line is no longer subject to delays in the City Loop from late Craigieburn and Upfield line trains. That's significant as the Craigieburn line is crowded while Upfield suffers from a single track on parts. On the other hand the through-running may make it susceptible to issues on the other side of town, with this equally so for Pakenham and Cranbourne.
Werribee, Williamstown and Frankston got new timetables with through-running removed and Frankston re-entering the City Loop. That may improve reliability numbers (as cross-city operation was always fairly fragile) though the bogey of loop bypasses is back for Frankston line passengers.
Sunbury, Pakenham and Cranbourne lines of course use the Metro Tunnel instead of the City Loop. The Sunbury line is no longer subject to delays in the City Loop from late Craigieburn and Upfield line trains. That's significant as the Craigieburn line is crowded while Upfield suffers from a single track on parts. On the other hand the through-running may make it susceptible to issues on the other side of town, with this equally so for Pakenham and Cranbourne.
Notwithstanding this, the higher frequency and upgraded signalling now in effect on the central part of the Metro Tunnel line should allow faster recovery from delays. The effectiveness of this should be discernible from performance data that we'll look at in a moment.
Those analysing punctuality should also consider dwell times at the new stations. Too short dwell times mean poor recorded punctuality while excessive dwell times is both annoying and inflates recorded performance. New systems often have a settling-in period as passengers get used to new travel patterns, with running and dwell time reviews highly desirable later.
What happened? (the big picture)
Your choice of baseline is critical when making comparisons.
Especially as February is not a typical month. It's busier than January but, with universities not fully started, it's quieter than March. Lower loadings is typically good for train punctuality. But hot weather, which February has a lot of, is not.
To account for these variations I compared Metro train service delivery and punctuality for February 2026 against the 12 month average for 2025. I also looked at the last five years' worth of February performance numbers to better compare like months. Numbers are reported here for those interested (use the PowerBI app).
First of all the big picture.
Network wide service delivery in February 2026 was the same as the 2025 calendar year average at 98.9%. But on time performance was slightly better at 94.3% versus 94.1% network average.
Both service delivery and on time performance was at least as good as (and more often better) than that recorded for all Februarys going back to 2022 if not further.
Especially as February is not a typical month. It's busier than January but, with universities not fully started, it's quieter than March. Lower loadings is typically good for train punctuality. But hot weather, which February has a lot of, is not.
To account for these variations I compared Metro train service delivery and punctuality for February 2026 against the 12 month average for 2025. I also looked at the last five years' worth of February performance numbers to better compare like months. Numbers are reported here for those interested (use the PowerBI app).
First of all the big picture.
Network wide service delivery in February 2026 was the same as the 2025 calendar year average at 98.9%. But on time performance was slightly better at 94.3% versus 94.1% network average.
Both service delivery and on time performance was at least as good as (and more often better) than that recorded for all Februarys going back to 2022 if not further.
Big new train timetables don't always go to plan but this one has done pretty well on a macro scale. Metro managers won't be unhappy with these numbers. That's important right now given that MTR is one of the three shortlisted franchise groups for MR5.
What happened? (performance by line)
Overall network results can mask big variations in performance trend by line. In this case actual results deviated from expectations, including some raised by government. I'll present more detail later, but for now below is a snapshot of where performance improved and where it didn't.
What happened? (performance by line)
Overall network results can mask big variations in performance trend by line. In this case actual results deviated from expectations, including some raised by government. I'll present more detail later, but for now below is a snapshot of where performance improved and where it didn't.
To summarise the Clifton Hill and Burnley groups (not listed as they didn't get a timetable change) maintained relatively good performance. And delivery on the old cross-city group (ie Frankston/Werribee/Williamstown) improved.
Both contributed to the strong network average cited before, offsetting poor delivery results on the Craigieburn and Upfield lines. Performance of lines now serving the Metro Tunnel was close to their past averages except for Sunbury's punctuality whose strong rebound contributed to a better than average network result while delivery was not so good.
The main movers are summarised below (bar lengths relative only).
Craigieburn and UpfieldBoth contributed to the strong network average cited before, offsetting poor delivery results on the Craigieburn and Upfield lines. Performance of lines now serving the Metro Tunnel was close to their past averages except for Sunbury's punctuality whose strong rebound contributed to a better than average network result while delivery was not so good.
The main movers are summarised below (bar lengths relative only).
These lines should have got a performance lift. All Sunbury line trains were removed from a shared loop portal and hardly any new services were added in their place.
The state government said reliability would improve with these lines specifically mentioned.
Only thing is that it didn't.
Instead service delivery fell relative to both the previous year's average and the four previous Februarys on both lines. Craigieburn's delivery was so poor that if the whole Metro network behaved like that then compensation would be paid due to it falling below the 98.5% threshold.
People naively expecting Upfield line trains to continue much past Coburg were especially short-changed last month, in several senses of the word. More than 3% of trains did not go the full distance, with the rate of trains terminating short between 3 and 6 times previous February averages. While short trips get reported there appears to be no separate passenger compensation trigger for them, unlike for delivery and on-time running.
The state government said reliability would improve with these lines specifically mentioned.
Only thing is that it didn't.
Instead service delivery fell relative to both the previous year's average and the four previous Februarys on both lines. Craigieburn's delivery was so poor that if the whole Metro network behaved like that then compensation would be paid due to it falling below the 98.5% threshold.
February's punctuality on both lines (see below) also dropped below the 2025 average, although it was still better than February of that year. Again Craigieburn's was substandard, with punctuality consistently below the 92% network compensation trigger. On the numbers Craigieburn line passengers should be getting compensation for late running but do not because the government drew up the franchise agreement to only pay when the whole network average is poor.
February 2026 was especially poor for Craigieburn relative to other lines. Its 88.7% made it Metro's most delayed line by a long shot. The next tardiest were Stony Point (91.3% on time) followed by Belgrave and Frankston (92%). Thus Craigieburn passengers are about twice as likely to experience delayed trains in February 2026 given the network average was 94.3% on time.
Unlike Craigieburn, the Upfield line's punctuality last month (93.9%) was only slightly below the network average. But there's a catch as you'll see later.
February 2026 was especially poor for Craigieburn relative to other lines. Its 88.7% made it Metro's most delayed line by a long shot. The next tardiest were Stony Point (91.3% on time) followed by Belgrave and Frankston (92%). Thus Craigieburn passengers are about twice as likely to experience delayed trains in February 2026 given the network average was 94.3% on time.
Unlike Craigieburn, the Upfield line's punctuality last month (93.9%) was only slightly below the network average. But there's a catch as you'll see later.
People naively expecting Upfield line trains to continue much past Coburg were especially short-changed last month, in several senses of the word. More than 3% of trains did not go the full distance, with the rate of trains terminating short between 3 and 6 times previous February averages. While short trips get reported there appears to be no separate passenger compensation trigger for them, unlike for delivery and on-time running.
You can read more about last month's Upfield line performance on this item from Climate Action Merri-bek.
To summarise, February 2026 was a terrible month for the Craigieburn and Upfield lines. Not only did they miss out on significantly upgraded timetables but promised reliability improvements failed to materialise with service delivery getting worse, not better.
If I was in government I'd be worriedly monitoring future months' performance closely given growth pressures in the outer north and marginal seats like Pascoe Vale closer in. Unless it can turn things around quickly, this timetable risks being a net negative for Craigieburn and Upfield passengers due to reduced delivery (both lines), chronic delays (Craigieburn) and early terminations (Upfield).
Sunbury/Pakenham/Cranbourne
It was hard to forecast how these lines would go, with these now running through the new Metro Tunnel.
Sunbury line's numbers were the most interesting. February's service delivery deteriorated to 98.5% versus a 99% average for 2025. Expressed in another way that's a 50% increase in the proportion of trains cancelled, or an avoidance of compensation by the narrowest margin if replicated network wide. On the plus side the improved Sunbury line frequency means that except for major disruptions the waiting time to the next train is likely to be halved if services do get cancelled.
To summarise, February 2026 was a terrible month for the Craigieburn and Upfield lines. Not only did they miss out on significantly upgraded timetables but promised reliability improvements failed to materialise with service delivery getting worse, not better.
If I was in government I'd be worriedly monitoring future months' performance closely given growth pressures in the outer north and marginal seats like Pascoe Vale closer in. Unless it can turn things around quickly, this timetable risks being a net negative for Craigieburn and Upfield passengers due to reduced delivery (both lines), chronic delays (Craigieburn) and early terminations (Upfield).
Sunbury/Pakenham/Cranbourne
It was hard to forecast how these lines would go, with these now running through the new Metro Tunnel.
Sunbury line's numbers were the most interesting. February's service delivery deteriorated to 98.5% versus a 99% average for 2025. Expressed in another way that's a 50% increase in the proportion of trains cancelled, or an avoidance of compensation by the narrowest margin if replicated network wide. On the plus side the improved Sunbury line frequency means that except for major disruptions the waiting time to the next train is likely to be halved if services do get cancelled.
Out of all the numbers surveyed, the largest statistical gain was undoubtedly Sunbury's on-time performance. This was much higher under the new timetable, exceeding anything in recent previous years and the 92% compensation trigger (if replicated network wide). While the February 2026 result appears creditable the real test is what happens after usage patterns have stabilised and (allegedly excessive) dwell times have been reviewed.
Pakenham and Cranbourne line statistics were less interesting than Sunbury's. Both delivery and punctuality in February 2026 were slightly below the 2025 annual average. Delivery on both is similar to previous Februarys while punctuality is better (Pakenham's numbers below, Cranbourne's not a lot different).
Frankston/Werribee/Williamstown
I've grouped these due to their past association in the cross-city group up to January 31, 2026. All lines have had works related altered operating patterns and/or significant periods with replacement buses.
All three lines had a good story to tell regarding service delivery with this higher in February 2026 than either the yearly average or recent Februarys. Maybe splitting operations at Flinders Street helped, although it is also true that frequent and reliable through-city running lines make trains useful to a wider variety of trips as Metro Tunnel users have now experienced.
As for punctuality, Werribee's February 2026 performance was similar to the 2025 average while Frankston and Williamtown's is about 0.5 to 1% worse. On a percentage basis Williamstown was similar to the network average (94.3% punctual) while Frankston scored near the bottom of the pack with 92%. Werribee's numbers are below.
Other Metro lines
Lines that did not get a timetable change performed relatively well in February 2026. This helped offset problems with the northern and Metro Tunnel groups, pushing the average up.
V/Line
1 February included changes V/Line's timetables, including extra trips on the Seymour line. 96.7% of scheduled V/Line services were delivered in February. This is inferior to the 2025 annual average (97.2%) but better than the very poor January 2026 result (91.1%). Punctuality had a similar pattern.
Seymour, which had extra services scheduled, fared a bit worse last month with 95.5% of trains running in February 2026 versus 97.8% for 2025 and 93.2% in January 2026. Seymour's punctuality was 85.5% last month, the lowest number since April 2024.
The publicly-operated V/Line's operational performance has been sub-par for years, with the franchised electrified Metro network typically cancelling or delaying a far lower proportion of services. This government's record of avoiding electrification on lines that could benefit, including Wyndham Vale, Melton and Wallan, has increased the number of people using less reliable V/Line frequently as suburbs continue to expand beyond the electrified network. With the last major V/Line to Metro electrification being way back in 2012 (when Melbourne had one million fewer people), expect outer suburban rail electrification to feature prominently in this year's state election campaign.
Conclusions
All three lines had a good story to tell regarding service delivery with this higher in February 2026 than either the yearly average or recent Februarys. Maybe splitting operations at Flinders Street helped, although it is also true that frequent and reliable through-city running lines make trains useful to a wider variety of trips as Metro Tunnel users have now experienced.
As for punctuality, Werribee's February 2026 performance was similar to the 2025 average while Frankston and Williamtown's is about 0.5 to 1% worse. On a percentage basis Williamstown was similar to the network average (94.3% punctual) while Frankston scored near the bottom of the pack with 92%. Werribee's numbers are below.
Lines that did not get a timetable change performed relatively well in February 2026. This helped offset problems with the northern and Metro Tunnel groups, pushing the average up.
V/Line
1 February included changes V/Line's timetables, including extra trips on the Seymour line. 96.7% of scheduled V/Line services were delivered in February. This is inferior to the 2025 annual average (97.2%) but better than the very poor January 2026 result (91.1%). Punctuality had a similar pattern.
Seymour, which had extra services scheduled, fared a bit worse last month with 95.5% of trains running in February 2026 versus 97.8% for 2025 and 93.2% in January 2026. Seymour's punctuality was 85.5% last month, the lowest number since April 2024.
The publicly-operated V/Line's operational performance has been sub-par for years, with the franchised electrified Metro network typically cancelling or delaying a far lower proportion of services. This government's record of avoiding electrification on lines that could benefit, including Wyndham Vale, Melton and Wallan, has increased the number of people using less reliable V/Line frequently as suburbs continue to expand beyond the electrified network. With the last major V/Line to Metro electrification being way back in 2012 (when Melbourne had one million fewer people), expect outer suburban rail electrification to feature prominently in this year's state election campaign.
Conclusions
February 2026's stable delivery and punctuality is mostly attributable to good performance on the lines that did not get a timetable upgrade in that month, ie the Burnley and Clifton Hill group. The Werribee, Williamstown and Frankston lines also played their part though the cross-city group being broken.
Did the Metro Tunnel and its timetable contribute to trains running reliably and on time in February 2026? The short answer is no. Except possibly for Sunbury's improved punctuality, the Metro Tunnel and the accompanying timetables had less benefit than hoped. This can be demonstrated by the sharply reduced performance of Craigieburn and Upfield and the generally middling performance of the Metro Tunnel lines.
March's numbers, available around April 10, will also be of interest to see whether February's trends are sustained or reversed. The weather is cooler but March sees more passengers with universities resuming, a packed major events calendar and concerns over access to the MCG from the south-east.
Key matters arising from these numbers for rail operations and service include (i) restoring service delivery and reliability on the Craigieburn and Upfield lines, (ii) examining run and dwell times on the Metro Tunnel lines and (iii) creating a reliable cross-city group when Werribee, Williamstown and Sandringham are linked later in the year.
Key matters arising from these numbers for rail operations and service include (i) restoring service delivery and reliability on the Craigieburn and Upfield lines, (ii) examining run and dwell times on the Metro Tunnel lines and (iii) creating a reliable cross-city group when Werribee, Williamstown and Sandringham are linked later in the year.
