The Shimonoseki Kaikyo Marathon 2026 (下関海響マラソン) takes place on Sunday, November 1, 2026. According to the race guidelines on the official website, the marathon has a field of 10,000 runners (first come, first served), starts at 8:30 a.m. with a 6-hour time limit, and both starts and finishes in front of Kaikyo Messe Shimonoseki. HotelBank Editorial Team tracked booking progress for Shimonoseki City accommodation (N=33 properties / 1,620 rooms, based on listings on major booking sites) from 45 days before each stay date. Race eve, Saturday, October 31, had already reached an estimated OCC (based on OTA-listed inventory) of 90.4% at 45 days out and stood at 94.3% at the latest observation (9 days out), filling at a pace nearly 15pt ahead of the previous Saturday, October 24, which was at 75.5% at 45 days out. By contrast, race night, Sunday, November 1, moved only from 76.2% at 45 days out to 83.6% at 10 days out, and two nights before, Friday, October 30, from 64.1% at 45 days out to 77.7% at 8 days out — the demand concentrated on race eve has not fully spilled over to the surrounding nights. This article overlays the booking curves for race eve, race night and two nights before on the same weekdays of the weeks before and after, and sets out how much inventory remains on which night and what can be considered in the roughly one week that is left.

Coverage: accommodation in Shimonoseki City, Yamaguchi Prefecture (all categories), N=33 properties / 1,620 rooms (31–32 properties / 1,599–1,618 rooms on some observation dates). Occupancy in this article is an estimate based on OTA-listed inventory; price metrics are not covered. Definitions are at the end of the article. Data as of October 23, 2026 (latest observation: October 22).

Race eve, Saturday, October 31 — 90.4% at 45 days out, 15pt faster than a regular Saturday, but gaining only +3.9pt since then

We start with race eve, the night on which demand is most concentrated. Estimated OCC (based on OTA-listed inventory) for Saturday, October 31 was 90.4% at the September 16 observation, 45 days before the stay date, and the sold-out property share (the estimated share of properties with no listed inventory visible on OTAs and other channels) was 46.9%. Among Saturdays, the previous week’s October 24 stood at 75.5% at 45 days out (sold-out property share 21.9%) and the following week’s November 7 at 82.2% (30.3%), so race eve started its progress 14.9pt above the previous week and 8.2pt above the following week. With a 10,000-runner field and first-come entry, the data suggest that runners were securing race-eve accommodation as soon as their places were confirmed, well before the 45-day mark.

The gains after 45 days out, however, are small. Race eve was at 92.0% at 30 days out (October 1) and 94.3% at the latest observation, 9 days out (October 22) — only +3.9pt since 45 days out. A regular Saturday, October 24, by contrast, built up +18.9pt, from 75.5% at 45 days out to 85.4% at 30 days out and 94.4% at 9 days out, catching up to almost the same level as race eve at the same 9-day point. In other words, race eve is defined not by “how high it ends” but by “how fast it fills”, and the demand the market as a whole can still win in the last two weeks is very limited. The sold-out property share for race eve also hovered around 60%, at 57.6% at 30 days out, 60.6% at 15 days out and 63.6% at 10 days out, and at 9 days out only 92 of the market’s 1,620 rooms remained. This shape, in which pre-race demand is largely decided by 45 days out, matches what we saw for Kanazawa City (91.1% at 45 days out, +10.6pt vs the previous Saturday) in Kanazawa Marathon 2026 Eve: OCC 91% at LT45; Yokohama, Kobe Compared.

Source: MetroEngines Research; compiled by HotelBank Editorial Team

In the chart, race eve, October 31 (solid blue), runs flat in the 90% range from 45 days out, while the previous Saturday, October 24 (navy), climbs steadily to catch up and the following Saturday, November 7 (light blue), moves in the low-to-mid 80% range. For the market, race eve is already at the stage of “how to reduce lost bookings”, and the question is how to combine the few remaining rooms with other nights when selling them.

Race night, Sunday, November 1, and two nights before, Friday, October 30 — inventory remains on the surrounding nights

The two nights around race eve look different. Estimated OCC for race night, Sunday, November 1, was 76.2% at 45 days out (September 17), 77.0% at 30 days out (October 2) and 83.6% at the latest observation, 10 days out (October 22), with the sold-out property share moving 15.2% → 21.2% → 36.4%. Compared with a regular Sunday — the previous week’s October 25 was at 65.1% at 45 days out, 73.5% at 30 days out and 79.3% at 10 days out — race night was 11.1pt higher at 45 days out and 4.3pt higher at 10 days out. Race night clearly carries some overnight demand, but the gap is smaller than race eve’s +14.9pt, and 266 rooms were still left in the market at 10 days out. Given that the following Sunday, November 8, is at only 72.2% at 17 days out (sold-out property share 24.2%), race night is best described as “stronger than a regular Sunday, but not a night that sells out like race eve”.

The reason lies in the race’s timetable. With an 8:30 a.m. start and a 6-hour time limit, most finishers cross the line by early afternoon and then choose between heading home the same day or staying one more night and travelling the next day. Monday, November 2 is the day before a holiday, as November 3 is Culture Day, yet the market for November 2 is at 69.3% at 11 days out with a sold-out property share of 12.1%, so inventory is still ample. Extended-stay demand from race night into the following day does not yet show fully in the current curves.

Source: MetroEngines Research; compiled by HotelBank Editorial Team

Two nights before, Friday, October 30, shows almost no visible event effect. It moved only +0.3pt in 30 days, from 64.1% at 45 days out (September 15) to 64.4% at 30 days out (September 30), and stands at 77.7% at the latest observation, 8 days out (October 22), with a sold-out property share of 18.2%. A regular Friday, October 23, started from almost the same point at 65.5% at 45 days out but had reached 85.0% by 8 days out, so two nights before is 7.3pt lower at the same point. The following Friday, November 6, had also reached 85.9% at 15 days out, versus 73.1% for two nights before at 15 days out — a gap of 12.8pt. Arrive-the-day-before demand appears to be concentrated on race eve, and guests arriving two days ahead still look like a small minority across the market as a whole. The same pattern — event demand concentrated on a specific night and not readily spilling over to the nights around it — was also observed in Hiroshima Arena Concert Oct 17: 85.1% OCC at T-45, Little Spillover.

Source: MetroEngines Research; compiled by HotelBank Editorial Team

Putting this together, of the three nights only race eve was effectively decided by 45 days out; race night is higher than a regular Sunday but still has 266 of 1,620 rooms in inventory, and two nights before trails a regular Friday. The table below lays out this structure on the three-point frame (45 days out, 30 days out, latest) as of October 23.

Estimated OCC (based on OTA-listed inventory). Sold-out property share in parentheses (the estimated share of properties with no listed inventory visible on OTAs and other channels). N=33 properties / 1,620 rooms (32 properties / 1,618 rooms for Oct 23 and Oct 25; some 45-days-out observations cover 32 properties / 1,601–1,618 rooms). All latest observations are as of October 22.

Source: MetroEngines Research; compiled by HotelBank Editorial Team

Day-of-week patterns in September and October — Shimonoseki is strongest on Saturdays, and in October even weekdays ran in the mid-80% range

As a yardstick for judging the three nights around the race, it is worth checking day-of-week results for the last two months as well. In finalized September data for Shimonoseki City (N=22–35 properties, monthly average 81.5%), Saturdays averaged the highest at 87.3% (sold-out property share 62.2%) and Wednesdays the lowest at 77.5%. September’s Sunday figure of 81.8% includes the holiday period that recorded 100.0% on Sunday, September 20 and 99.9% on Monday, September 21, so a regular Sunday should be read as somewhat lower than this (booking progress in Yamaguchi Prefecture over that five-day holiday is broken down by property type in Yamaguchi Silver Week: Ryokan 99.5% on 9/22, Business 65.5% on 9/23). October, covering the elapsed period from the 1st to the 23rd (N=31–33 properties, average 86.2%), moved into a higher band than September, weekdays included: Saturdays at 92.9% (sold-out property share 74.2%), Thursdays at 88.8% and Fridays at 86.3%. The month’s peaks were 96.6% on Saturday, October 10 and 96.7% on Thursday, October 22, making it a month in which a holiday weekend and large weekday demand overlapped.

Simple average of daily estimated OCC for each day of week. September: N=22–35 properties (cohort varies by date); October: N=31–33 properties (elapsed period, 1st–23rd; at most 32 properties / 1,618 rooms).

Source: MetroEngines Research; compiled by HotelBank Editorial Team

Against this day-of-week pattern, race eve’s 94.3% on October 31 sits in the same band as October’s Saturday average of 92.9%, so the market can expect to land at about the level of “a strong October Saturday”; race night’s 83.6% on November 1 is on a par with October’s Sunday average of 84.0%; and two nights before, at 77.7% on October 30, is more than 8pt below October’s Friday average of 86.3%. Although some late pickup remains for two nights before, it is starting from a lower position than an October Friday.

For revenue managers running hotels and ryokan in Shimonoseki City — implications and action plan

All of the above are an outline of opportunities, not a guarantee of results. Decisions on your own figures should be based on each property’s booking status.

Summary — measure the three nights at three points: 45 days out, 30 days out and latest

Ahead of the Shimonoseki Kaikyo Marathon 2026, the Shimonoseki City accommodation market (N=33 properties / 1,620 rooms) shows a three-layer structure: race eve, Saturday, October 31, plateaued early at 90.4% at 45 days out → 92.0% at 30 days out → 94.3% at 9 days out; race night, Sunday, November 1, at 76.2% → 77.0% → 83.6%, slightly above a regular Sunday but with inventory left; and two nights before, Friday, October 30, at 64.1% → 64.4% → 77.7%, trailing a regular Friday. As reusable yardsticks, we would point to three: (1) race-eve-type nights are largely decided at the 45-days-out level, with little late pickup; (2) on race-night-type nights, the gap over a regular same-weekday night narrows from 45 days out to the final days; and (3) two-nights-before-type nights carry almost no event effect and should be judged on the regular weekday curve. With a little over a week left, what can still be moved is inventory on race night and two nights before, not race eve, and multi-night conditions and post-race stay promotions are likely to be the key.

About the Data

Source: MetroEngines Research; compiled by HotelBank Editorial Team

References and Sources