Across 31 August – 6 September 2026 the Malta Traffic Index averaged 21.3 — moderate by the standards of our own monitoring. That figure is the mean of 153 readings taken across 7 days on the island's motorway, primary and secondary network.

That is effectively unchanged from the previous week (21.9). Weeks that sit this close together usually mean nothing unusual happened — no major closure, no weather event, no holiday distorting the pattern.

The single worst reading of the week was 89, recorded on Thursday 3 September 2026 around 09:00.

Thursday was the heaviest day of the week, averaging 28.8 with a peak of 89. Sunday was the calmest at 4.7. The gap between the two — 24.1 index points — is the practical difference between planning your week around the roads and ignoring them.

The weekday–weekend split was pronounced: 26.4 on working days against 7.5 at the weekend. That is the commute signature — Malta's congestion is overwhelmingly generated by people travelling to work and school, not by leisure traffic.

The heaviest hour of the average day was 18:00, at 38. The opposing peak — 09:00 — reached 31.

The evening was clearly the worse end of the day this week. That is the usual Maltese pattern: the morning commute is compressed into a sharp window because everyone is bound by the same start times, while the evening spreads across several hours as work, school pick-ups, errands and leisure journeys overlap.

Measured corridor by corridor, Paola — Ħal Luqa Rd carried the heaviest average delay of the week: journeys through that segment took 44.2% longer than they would on an empty road, across 153 readings. At its worst moment the same segment was running 162% over free-flow.

Behind it came Kappara Junction (32.2%) and Manuel Dimech — Sliema (32.2%). These are averages across the whole week including empty night hours, so the peak-hour experience on each is considerably worse than the headline figure suggests.

We also recorded 66 readings where a monitored segment was reported fully closed. Closures are counted separately from delay because a closed road has no travel-time ratio — it simply removes capacity and pushes the load onto whatever runs parallel.

Three corrections apply to this section. Until 19 September 2026 our Coast Rd — Salina monitoring point was returning the same stretch of road as Ħamrun Bypass, so the two showed identical figures and one road was counted twice. The 153 Coast Rd — Salina readings taken before that date are left out of this report.

Until 18 September 2026 our Fgura — Triq Hompesch monitoring point was returning the same stretch of road as Paola — Ħal Luqa Rd, so the two showed identical figures and one road was counted twice. The 153 Fgura — Triq Hompesch readings taken before that date are left out of this report.

3 readings from Ħal Luqa — Airport showed a travel time more than 5 times the free-flow time and are left out as unknown. That monitoring point sits on a very short stretch of road, so a queue of a few minutes there reads as a delay of well over a thousand percent and does not describe the corridor. The week's highest index reading of 89 was itself driven by one of these readings; the index figures in this report are shown as they were recorded at the time.

The week ahead carries a dated event from Malta's closure calendar: Victory Day (Il-Vitorja) (Grand Harbour, Senglea, Naxxar, Mellieħa, Xagħra (Gozo)). Expect the affected localities to close their cores and the approach roads to load well before the published window.

On the evidence of this week, Thursday is the day worth planning around, and the evening peak is the window to avoid if your journey is flexible.

Every figure above comes from our own continuous measurement of Malta's road network — the same Malta Traffic Index shown live on the map, recorded and kept. The index blends the weighted Travel Time Index across a fixed panel of monitoring points on motorway, primary and secondary roads with the share of those points currently running slow or closed.

This report is compiled automatically from those recorded readings at the close of each week. The numbers are measurements, not estimates: nothing here is modelled, interpolated or predicted, and weeks with insufficient coverage are not published at all. Sampling density follows real site usage, which is why overnight coverage is thinner than daytime coverage and why we state coverage gaps explicitly rather than smoothing over them.