The rule says the tide moves one twelfth of its range in the first hour after high water, then two, three, three, two and one. It is a hand approximation to a sine curve, and people tend to assume it is rough. It is not. Compared against a true sinusoid of the same range and duration, the worst error anywhere in the tide is 2.6% of the range — and at the hour marks it is very nearly perfect.
At 2, 3 and 4 hours after high water, the rule of twelfths is exactly right — 25%, 50% and 75% of the range are the sinusoid's own values, to the digit. The only real disagreement is in the first and last hours, where the rule says 8.3% and the sinusoid says 6.7%: an error of 1.6% of the range. On a 12 m spring tide in the Bristol Channel that is about 20 cm. On a 2 m neap it is 3 cm.
So if the rule lets you down, it is almost never the twelfths that did it.
The rule quietly assumes the tide takes six equal hours to fall. Plenty of places do not oblige. A semidiurnal tide is really about 6 hours 12 minutes from high water to low — half of the 12h 25m lunar day — and in many estuaries the flood and the ebb are markedly unequal, because the shape of the seabed distorts the wave as it runs in.
Assume six hours where the real fall takes five, and three hours after high water the rule tells you 50% of the range has gone when the truth is 65%. That is a 15% error — six times the worst error the twelfths themselves ever make. On a 6 m range that is nearly a metre of water you thought you had.
The ones whose curve is not a sine wave at all. Southampton and the Solent have a famous double high water and a long stand; Poole similar. Parts of the Bristol Channel have a markedly faster flood than ebb. In those places a published tidal curve for the port is not a nicety — it is the only thing that works, and the twelfths will lie to you in a way that no amount of care with the arithmetic can fix.
Because the tide follows the Moon, not the Sun. The Moon returns to the same meridian every 24 hours 50 minutes, so a semidiurnal cycle is 12h 25m and each half is 6h 12m. Over a day that difference accumulates to nearly an hour, which is why tides run later each day by about 50 minutes.
Chart datum, which is approximately the lowest astronomical tide — the level the sea almost never goes below. That is why charted depths look alarmingly shallow: they are the worst case, and the height of tide is what you add to them. It also means a negative height of tide is possible, rare, and exactly when you would least like it.
That the tide is semidiurnal and symmetrical: one high and one low, falling smoothly, with the rule of twelfths as the approximation and a sinusoid of the same range and duration as the reference. Both are approximations to a real curve that is the sum of dozens of harmonic constituents.
That the numbers you type are right. This page does no astronomy and has no tide table in it: it takes the high and low water you give it and draws the shape between them. Garbage in is garbage out, and at sea that has consequences.
That the weather is doing nothing. Barometric pressure and wind move real tides by amounts that dwarf the rule's own error — roughly 1 cm per millibar away from 1013, so a deep low can raise the sea 30 cm above prediction, and a strong onshore blow more still. Surge is not in this page and is not in your tide table either.
Port ranges shown are nominal typical figures for scale, not tidal predictions. Verified in
tide_model.py.
No change to this page at all since the earliest archived release (v5.66). The full history for the site is in the changelog.