Gann Time Cycles: Seasons and Counts
Time cycles turn the calendar into a count, and a count is only useful as a window that either agrees with price or gets discarded. That single rule separates the working version of this tool from the mystical version. Anniversaries and seasons do not make markets move. It gives the trader a schedule of moments worth watching, and watching is all the count ever promises.
What a Time Cycle Actually Counts
A time cycle is a number of trading days or calendar days counted forward from a significant high or low. Pick the pivot, count the days, mark the date. The count produces a window, never an appointment.
The distinction matters more than any other idea in this lesson. An appointment says the market must turn on that date. A window says the market may turn somewhere inside a small stretch of days, and only if price has traveled to a place where turning makes sense. The first claim cannot be tested honestly because it always finds an excuse. The second can be logged, scored, and thrown out when it fails.
Here is why the count earns any attention at all. Price tools and time tools fail in different ways. A support level fails when price slices through it. A count fails when the window opens and nothing happens. When a time window opens while price sits on a level that already matters, a square root, a half retracement, an old pivot, the two tools stack. That stacking is the method. Each tool covers the other's weakness for a few days.
A window with price nowhere is left closed. That sentence deserves to stand alone, because most bad cycle work comes from ignoring it. The trader counts 90 days, sees the date arrive, and forces a trade in the middle of nowhere on the chart. The count said when. It never said where. Without the where, the when is decoration.
A count schedules attention. It never replaces judgment. The window says when to look; the chart says what is actually happening. The trader who lets the date decide the trade has handed the calendar a job it was never meant to do.
One more discipline. Counts come in two flavors: trading days, which skip weekends and holidays, and calendar days, which count everything. Both get used. The honest habit is to pick one convention per count, write it down at day zero, and never switch mid-count to make a date fit.
The Anniversary and the Year's Divisions
The oldest count in this tradition is the anniversary: one calendar year from a major pivot. The claim is that markets remember their own turning dates. Tops and bottoms cluster near the dates of earlier tops and bottoms, often enough that the date deserves a mark on the calendar.
A skeptical reader should ask why that would ever work. The plausible answer is boring, which is a point in its favor. Earnings cycles, tax dates, contract expiries, planting and delivery schedules, and reporting seasons all run on the year. Human decision-making runs on the year too. Whatever the cause, the anniversary is treated as a window, not a prophecy, and it gets the same test as every other count: is price at a meaningful level when the date arrives?
Inside the year sit the natural divisions. Ninety days is a quarter. One hundred eighty is a half. Two hundred seventy is three quarters. Thirty days is the natural month. These are not magic numbers. They are the fractions of the year that business, weather, and reporting already run on, which is the entire argument for watching them.
Every count is treated as a window a few days wide, roughly plus or minus five around the target date. A day-90 count means days 85 through 95 get watched, with the center of the window carrying the most weight. Precision past that is false comfort. Markets do not keep appointments to the day.
Counts also stack with each other. A day-90 window measured from the major low that happens to fall 30 days after a secondary pivot reads stronger than either count alone. Two independent counts pointing at the same stretch of days is the time version of confluence. It still needs price to show up at a level, but it earns a closer look.
Keep the count list short. A chart with fifteen counts running becomes a chart where some window is always open, and a tool that is always open predicts nothing. Three to five active counts from clearly significant pivots is a working set. More than that is noise dressed as rigor.
Seasonal Windows
Seasonal analysis is the same idea worn by the calendar itself. Commodities carry documented calendar tendencies because production and consumption follow the year. Grain markets react to weather scares in the growing season. Energy demand builds into winter. These patterns exist because the physical world has a schedule.
Equity indexes carry softer tendencies tied to flows and the calendar: tax periods, bonus seasons, quarter-end rebalancing, holiday stretches with thin participation. Softer is the honest word. A commodity tendency rests on weather and biology. An index tendency rests on habits, and habits change without notice.
A tendency is a wind, never a promise. A sailor plans around the prevailing wind and still checks the sky every morning. The seasonal read says which way the breeze usually blows during this stretch of the year. It says nothing about today.
This lesson claims no statistics on purpose. Seasonal studies exist, and they disagree with each other depending on the sample, the market, and the years included. The working trader does not need a win rate for the tendency. The trader needs a calendar of stretches worth watching, and the discipline to demand price confirmation inside them.
Seasonal windows get used exactly like count windows. The window opens. The trader checks whether price sits at a level that matters, a division of a range, a prior pivot, a square. If yes, the setup is live and gets traded with normal risk rules. If no, the window closes and nothing happens. A seasonal window that passes quietly is not a failure of the tool. It is the tool doing its job, which is to say no most of the time.
A Worked Example: One Low, Five Counts
Everything below is invented. Hypothetical index, round numbers, no real market. The point is the procedure, not the outcome.
A major low prints at 4,100 on March 4. That is day zero. The index advances to 4,360 by day 41, and the trader starts counting forward from the low using calendar days, with windows of plus or minus five.
| Count | Window | What price did | Verdict |
|---|---|---|---|
| Day 30 | Days 25 to 35 | Minor three-day pause near 4,310 | Weak reaction, noted, no trade |
| Day 60 | Days 55 to 65 | Nothing, trend continued | Miss, logged |
| Day 90 | Days 85 to 95 | Pullback low at 4,232 on day 92 | Window and level agree, marked reaction |
| Day 180 | Days 175 to 185 | Top at 4,540 on day 178 | Window catches the high |
| Anniversary | Following March, plus or minus five | Window open, outcome pending | Scheduled, not decided |
Walk the rows. The day-30 count caught a pause, not a turn, and the honest read is that a pause inside a window is weak evidence. Noted, no action. The day-60 count produced nothing at all. The lesson says so plainly because misses are data. A method that only records its hits is a story, not a tool.
The day-90 window is where the method shows its shape. Price pulled back and printed a low at 4,232 on day 92, inside the window. Now check the level: 4,232 is one half of the 4,100 to 4,360 advance. Half of 260 is 130, subtracted from 4,360 gives 4,230, close enough at this scale. The time window and the price level agree. That agreement is the signal. The trader treats 4,232 as a marked reaction low and manages the position accordingly, with the same stops and sizing as any other setup.
The day-180 window catches the top at 4,540 on day 178. Whether the trader sold there is a separate question, because the window only said watch. What price did inside the window, stalling after a long advance, is what made the call.
The anniversary window opens the following March, and the example ends there on purpose. The window is open. The method schedules the decision. The market makes it.
Time Cycle Questions
Do cycles work on every market?
They work as windows on every liquid market, because the method only asks for a clean pivot and a count. Commodities with strong production calendars tend to respect seasonal windows more often, while indexes lean more on anniversary and division counts. Thin, illiquid markets produce unreliable pivots, which poisons every count built on them.
How wide should a window be?
About plus or minus five days around the target date is the standard working width. Shorter counts can run tighter, three or four days. Longer counts like the anniversary can run a touch wider. Anything past a week either way stops being a window and becomes a permanent state of alert, which is useless.
Do I need exact calendar dates?
No. The window exists precisely because exact dates are a fiction. Count carefully, mark the center date, then treat the surrounding days as the live zone. A turn on day 91 of a day-90 count is a hit. A turn on day 104 is a miss, and it gets logged as one.
What invalidates a cycle read?
Two things. First, the window closes with price nowhere near a meaningful level, which voids the count entirely. Second, price reaches a level inside the window and then breaks cleanly through it, which means the level failed and the count cannot rescue it. Either way, the read is dead and the trader moves to the next window.
The calendar has one more layer in this tradition, and it is the strangest tool in the entire framework. The next lesson takes up the planetary overlay, where the same counting discipline gets applied to astronomical cycles. It will get the same treatment as everything else: what it claims, how it is used, and exactly where the honesty runs out.