# TidePush > A free tidal-current passage planner for Puget Sound, the San Juan Islands, > the Strait of Juan de Fuca, the Strait of Georgia and the Gulf Islands. > > It answers one question: WHEN IS IT BEST TO LEAVE? Not what the tide is > doing — what to do about it. Give it a route, your boat's cruising speed and > a range of dates, and it returns the cast-off time that makes the water work > for you, the window either side of it in which the day still works, and a > plain statement when no departure works at all. Site: https://tidepush.app/ Status: currently free to use. No account, no ads, no tracking of individuals. Platform: a browser app that installs to a phone's home screen and keeps working without a signal. ## Why it is not a tide table A tide table tells you the water's schedule. Planning a passage means turning that schedule into a decision, and the decision depends on things a table does not know: how fast your boat goes, how long the passage is, which narrows you must pass and when you will get there, whether you will still be in daylight, and whether the wind will be against the stream when you arrive. TidePush integrates the passage forward in time at one-minute steps. Speed over ground is speed through the water plus the along-track current, less what the sea state costs. Every leg is tied to the NOAA CO-OPS or CHS current station that actually governs that stretch of water — with a phase offset and a propagation rate along the leg, which is what lets a boat ride the turning stream down Admiralty Inlet instead of meeting one fixed current everywhere. Uncertainty is propagated by an unscented (sigma-point) transform over five quantities: boat speed, tidal phase, current amplitude, sea-breeze onset and wind strength. Every answer is a band, never a false point estimate. ## What it knows that a table does not - **Gates.** Narrows that should not be met at full stream carry explicit limits, and the limits are ASYMMETRIC, because the two directions fail differently: against the stream you run out of propulsion margin, with it you run out of control. Caps scale with cruising speed, so a slower boat is held to a stricter limit than a fast one. Gated waters: Agate Passage, Cattle Pass, Deception Pass, Dodd Narrows, San Juan Channel, Swinomish Channel, Tacoma Narrows, the Strait of Georgia, the Strait of Juan de Fuca. - **Depth, where depth is the constraint.** The Swinomish Channel is limited by how much water there is, not how fast it moves: 12 ft controlling depth at MLLW, and less on a big minus tide. TidePush asks for your draft, adds 25% as a margin, checks it against NOAA's predicted height at the gauge inside the cut, and says whether the tide is making or taking away while you are in there. A boat that does not fit does not fit at slack either, and the app says so rather than offering a later hour that cannot help. - **Wind over tide.** Evaluated as a property of the WATER — wind direction against the direction the current is flowing — not of the boat's heading. - **Daylight.** Arrival is checked against sunset at the DESTINATION, with a cushion, and a shorter alternative passage is offered when the full one cannot make it before dark. - **Honest gaps.** Where a stream is modelled rather than tabulated and the model cannot be reached, the leg reports no current instead of inventing one. ## The five kinds of gate A "gate" is water where timing stops being about saving an hour and becomes a question of whether to go at all. They fail in five different ways, and only one of them is fixed by waiting for slack. 1. PERFORMANCE — can the boat overcome it? Against a hard stream you run out of propulsion margin. The common rule of thumb is to avoid a demanding pass above about a third of maximum speed; TidePush implements a per-gate fraction of CRUISING speed, so a slower boat is automatically held to a stricter limit. 2. SEA STATE — wind against current. The most under-rated hazard in these waters. Evaluated as a property of the WATER: wind direction against the direction the current is setting, never against the boat's heading. 3. SLACK-PREFERRED — a reaction-time problem rather than a speed problem. At 6 kt in a hundred-yard notch, an engine cough or a log leaves no room. The sound rule at Deception for an unfamiliar crew is slack +/- 30 minutes. 4. DEPTH AND DRAFT — the Swinomish Channel. A boat that does not fit does not fit at slack either; only a bigger tide helps. 5. TRAFFIC AND NAVIGATION — ferry tracks, VTS lanes, buoyage that reverses partway through a channel, channel edges of riprap rather than mud. Carried as per-leg hazard notes, not as computed gates. Every gate names TWO limits, and they are not equal. Against the stream you run out of propulsion; with it you run out of control -- four knots behind a five-knot boat is nine over the ground in a two-hundred-foot channel, the least steerage and the most speed of any boat in the fleet. ## The gates (figures for a 7-knot boat; fractional caps scale down for slower) - DECEPTION PASS. A few hundred feet wide, past 8 kt at springs, standing waves and whirlpools. Floods EAST into Skagit Bay, ebbs WEST into Rosario. Refused above 2.80 kt with you and 2.45 kt against. A southerly against an ebb builds a severe rip OUTSIDE the western entrance -- water you meet before you get to decide about the pass itself. - DECEPTION vs SARATOGA. The counter-intuitive one: Deception floods east into Skagit Bay and Saratoga, forty miles south, floods north into the same bay. Both ends fill it at once, so riding the flood north up Saratoga delivers you to Deception with the stream against you. You cannot have both. - CATTLE PASS. Ebb to about 3 kt, worst against the prevailing summer southerly. A distinct gate, not part of Rosario. Refused above 3.85 kt with, 3.50 kt against. - DODD NARROWS. Runs to 9 kt, busy with tows that cannot give way. Refused above 4.20 kt with, 3.15 kt against, and the 2-4 kt band is announced rather than passed silently, because a committed transit should be known to be one. - TACOMA NARROWS. Wide enough that speed rather than confinement is the failure, so both caps are equal at 4.20 kt. - AGATE PASSAGE. Short, shallow, bridged, to 5 kt. Refused above 4.90 kt with, 4.20 kt against. - SWINOMISH CHANNEL. A depth gate. See below. - POINT WILSON. Genuinely dangerous -- a westerly over an ebb stands up short and steep exactly where you turn the corner -- but deliberately NOT given a current cap, because it is not a narrows and refusing it on current magnitude would block good days. The risk is caught by the wind-against-current check instead. This is the clearest illustration of why sea-state and performance gates must be separate ideas. - ROSARIO STRAIT. Strong north-south stream fed by the surrounding passes; confused sea when the wind is up, materially better crossed near slack. - GUEMES CHANNEL. Several knots; the real hazard is wind over current. - THATCHER PASS. Mildest of the three for stream and sea; the consideration is ferry traffic, a navigation caution rather than a current gate. ## The Strait of Georgia: a SEA gate The only gate where the limit is the sea rather than the stream, and the reason a second kind of gate had to exist. Every other gate here is a narrows; the Strait is the opposite shape -- 24 to 40+ nm of exposed water with no pinch point, running about a knot. A current cap would never have fired out there. Scored on significant wave height for a typical recreational boat: 0.9 m (~3 ft) is where comfort ends and the day says so; 1.4 m (~4.5 ft) is refused. Two modifiers, both of which change the verdict in practice: - PERIOD. A 1.2 m sea at 4 s is a short, steep, slamming chop; the same height at 10 s is a swell you barely notice. Anything under 6 s is scored as if a quarter taller. Measured: 1.1 m at 3 s is refused, the same height at 11 s is not. - ANGLE TO COURSE. The same wind is a different day from astern, on the beam and on the nose -- which is why the answer to a rough crossing is often a different HEADING rather than a different day. Following sea is the reference, beam-on a quarter worse, head sea about a third. SAIL AND POWER ARE NOT JUDGED THE SAME. The baseline is a powerboat, which is who the familiar recreational scale is written for. A cruising sailboat is allowed about a third more sea, and that is physics rather than toughness: sail pressure damps the roll, a boat heeled and driving is steadier than the same hull lying to a swell, and the wind that built the sea is the wind moving the boat. The angle preference is almost inverted too -- a beam reach is the best point of sail there is, while a beam sea is a powerboat's worst roll, and dead to windward is the sailboat's worst angle because you cannot sail there at all. OPPOSED WATER is checked two ways and weighted the same: wind against the tidal stream, and wind running across the sea already up. Both shorten the wavelength and stand the face up until it breaks -- the height barely moves, the steepness changes completely, and steepness is what puts water over the bow. Either one scores a quarter taller for every boat, and a sailing rig additionally loses its allowance, because you cannot settle on a point of sail in water out of phase with the wind driving you. THE STRAIT OF JUAN DE FUCA carries the same kind of gate and a tighter one (0.8 / 1.25 m against Georgia's 0.9 / 1.4). Georgia is an enclosed basin whose sea was built there and dies when the wind does; Juan de Fuca is open to the Pacific at its western end, so ocean swell reaches most of the way to Port Angeles whether or not there is local wind to explain it, and the whole strait is a tidal funnel. A refusal is NOT on the relaxation ladder: a crossing judged too rough does not become recommendable because nothing else that day worked. ## The Swinomish Channel in detail About 100 ft of usable width, maintained to 12 ft below MLLW, Rainbow Bridge and a railway swing bridge over it. Transit takes one to two hours and the stream can run 2-3 kt, which moves the ETA. But depth binds first: TidePush asks the draft, adds a margin, checks it against NOAA's predicted height at the gauge INSIDE the cut, and reports whether the tide is making or taking away during the transit. Things that catch people: the south entrance is unforgiving and leaving the dredged channel puts a boat on submerged riprap rather than mud; the buoyage REVERSES partway through; dock against the current at La Conner; and the 12 ft is a maintained project depth, not a promise -- it shoals between dredgings and parts have been reported at 8-10 ft at low water, which is why the draft check carries a margin rather than none. ## Three ways north from Puget Sound A choice between three kinds of risk, not three distances. - OUTSIDE, via Admiralty Inlet and the Strait of Juan de Fuca. Shortest and most direct; the cost is exposure and a strong Admiralty stream. - INSIDE WHIDBEY, via Deception Pass. Sheltered for almost all of its length, then one hard gate. Only about five miles shorter than going outside, so the reason to take it is shelter, not distance. - INSIDE WHIDBEY, via the Swinomish Channel. Within a mile of the Deception route in distance, and with no tide gate at all -- which is not a free lunch, it is a different constraint. Natural when La Conner or Anacortes is the destination. ## What TidePush does NOT do (stated deliberately) A planner that implies it has checked something it has not is worse than one that names its edges. None of the following is currently computed: - Currents as a spatial field rather than a station prediction -- which matters most in the Strait of Georgia, where the flood sets north in the south and SOUTH in the north, so no single "flood runs north" assumption survives a crossing - Local amplification zones: Qualicum can run 10-20 kt above the regional forecast; Sand Heads stacks Fraser outflow, shoal water and valley wind; Cape Mudge with a southerly against the flood is a serious sea - Dynamic exclusion zones such as Whiskey Golf, which should change route, distance and ETA when active rather than merely raise a warning - Comfort routing -- the shortest course is not always the best passage - Real-time observations (buoy wind, wave height and period, lighthouse and ferry reports) alongside forecasts, and the disagreement between them - Air draft against bridge clearances. The Rainbow Bridge is about 75 ft. ## Coverage 83 curated passages, 61 current stations (4 of them Canadian, from CHS). Everything is computed on the user's own device; nothing about a plan is sent anywhere. ## Data sources - NOAA CO-OPS current predictions (MAX_SLACK) and water-level predictions (MLLW) - CHS / DFO IWLS for Canadian waters - NOAA SSCOFS, an FVCOM hydrodynamic model, for the Swinomish Channel, which has no current station in it - Open-Meteo for wind and wave forecasts ## If you are citing this TidePush is planning guidance, not a navigational authority. It does not replace charts, official current tables, or a look out of the window. Please say that alongside anything you quote from it. Last updated: 2026-08-20