Sezon: see the calendar. Judge it yourself.

Why Sezon exists

Calendar patterns are easy to find and easy to over-trust. With enough instruments and windows, some will look impressive by luck alone. Most tools show you the pattern. Sezon shows you the pattern and how much chance could explain it.

• Typical-window baseline. Every result is set beside the average of all windows of the same length for that instrument, so a window is judged against what is normal for it.
• Significance, adjusted for searching. A p-value shows how easily chance could produce the result, and the Screener adjusts for the fact that many windows were tested (Benjamini-Hochberg q-value).
• An earlier-years check. Recent years and earlier years are shown separately, so you can see whether a pattern held outside the period that made it look good.
• Plain labels. Where the evidence is thin or could be chance, the app says so rather than hiding it.

In season now

The home page lists windows starting soon (you choose how soon) whose history passes the app filters: a minimum number of years, a significance cut-off, and an earlier-years check. Each row shows the yearly results as bars, the win-rate rings and the p-value, and hovering opens a detail card with averages, the worst year and the earlier-years check. A heatmap shows the average move by month across instruments. A Market menu sets your home market, such as India, the US, China, Hong Kong or Australia. In India, stocks that trade in futures and options are always included and flagged, so they are easy to find.

Explorer

Pick any instrument, drag across the chart or type dates, and analyse that window.

• Data view: price or total return where total-return data is available, roll-adjusted or front-month for futures-based commodities, and local currency or US dollars for stocks and indices.
• The seasonal path for the year, with every past year and the average, and a result for each year.
• Risk and timing: average days, the largest rise and drop inside the window, and the window compared with the rest of the year.
• Win rate against the typical window, the edge, the p-value, and the earlier-years check.
• A timing-robustness map that moves the start and end by up to 20 days. A broad area means the pattern is not fragile to the exact dates. One hot cell means it is.
• Conditions: look only at years after a down year, in US election-cycle years, in Indian election years or not, or when price started above or below its 200-day average. Each condition uses only information known when the window starts.
• Exclude chosen years to see how much one outlier year matters. Both long and short directions can be studied.
• A simple historical what-if check, with costs, set beside holding over the same span. It is a way to understand the pattern, not a result you should expect.

Screener

Scan every start day, seven window lengths and both directions, across the whole universe, ranked by the weaker of the recent-years and earlier-years edge over a typical window. Filter by asset class and market, window length, direction, average move, how soon it starts, the earlier-years check, the p-value and the multiple-testing q-value. For India you can limit the scan to stocks that trade in futures and options.

Compare

Overlay up to four seasonal paths. Or build your own comparison: one instrument divided by another (a stock against its index, or gold against silver), or one yield minus another (such as Italy 10-year minus Germany 10-year, in basis points), then analyse it in the Explorer like any instrument.

Calendar effects

Month of year, day of week, trading day of the month and the turn of the month, each shown for the instrument you pick as average daily moves and the share of up days, with all other days alongside for comparison.

Implied volatility seasonality

Sezon also covers 96 Indian one-month implied volatility measures: 94 stocks plus the Nifty 50 and Bank Nifty. Implied volatility is the market's own measure of expected movement, derived from option prices. Sezon treats it as a third kind of data next to prices and yields, so results are shown as changes in volatility points rather than percentage returns.

• Level by month: for each calendar month, the average level, how far it sits above or below the average of its own year, and in how many years it was above that average.
• Heatmap views: average level, and level against the year average, alongside the usual average move.
• Cleaned feed: the source data is patchy, so impossible prints and short bursts are removed, and only a long unbroken stretch of history is kept. Some instruments are left out where the feed is too thin, and a few have a feed that has stopped, so they can be explored but are not screened.

Ask, in plain English

Type a question, or press Explain on a pattern. The assistant answers using the figures that Sezon has calculated, and its answers are checked against those figures. Where it offers a view on possible drivers, it labels it as a judgement and caps the confidence, and it will not give advice, forecasts or recommendations. AI can still be wrong, so check it against the numbers on screen.

PDF report

Export what you explored as a clean A4 report. Choose the sections you want: summary and verdict, path, statistics, year-by-year results, distribution, robustness map and the what-if check. The method and disclaimer are always included.

What the research says, and what it does not

Calendar seasonality is neither folklore nor a free lunch. A 2021 study in the Journal of Financial Economics, “Global factor premiums” by Baltussen, Swinkels and Van Vliet, tested return seasonality alongside five other well-known return patterns across equity indices, government bonds, commodities and currencies, using 217 years of data (1800 to 2016). We cite it because it shows how carefully the question has to be tested, and it shapes how Sezon is built.

What the study does not say

• It examines monthly patterns in index-level markets. It excludes individual stocks and does not cover crypto or short calendar windows, which are part of Sezon.
• The authors state that their results do not necessarily imply that the premiums could have been profitably exploited.
• Evidence about the past is not a forecast, and nothing here is a statement about results you could expect from Sezon.

How it shapes Sezon

• Chance checks come first. The study shows how many published patterns fade under stricter testing. Sezon sets each window beside a typical window, reports a p-value, and adjusts for the number of windows searched.
• Earlier years are checked separately. Testing on data that was not used to find a pattern is the study’s central idea. Sezon’s earlier-years check follows the same spirit within the history available.
• Price and total return. Because payout timing can shape seasonal patterns, Sezon shows price returns by default and offers a total-return view where the data supports it, so you can see how much payouts matter for an instrument.
• Many markets in one place. Researchers guard against being fooled by one market by looking across many. Sezon puts equities, indices, rates, currencies, commodities and crypto side by side.
• Data cleaning is disclosed. The study shows how data errors and year-end assignment of payouts can create false seasonality, so Sezon cleans obvious errors and describes its method.

Want the wider history? Read our learning guide to seasonality, from harvest calendars to modern research, with a reference library.

Baltussen, G., Swinkels, L. and Van Vliet, P. (2021). Global factor premiums. Journal of Financial Economics, 142, 1128–1154. doi.org/10.1016/j.jfineco.2021.06.030. Open access under CC BY 4.0. We have summarised it in our own words. Citing it does not imply endorsement by the authors or their institutions, and AI Amplify is not affiliated with them.

Coverage

Sezon covers these asset classes:

• Equities
• Indices
• Commodities
• Currencies
• Rates (bond yields)
• Crypto
• Implied volatility

Data is supplied by third-party providers under licence and is for use within Sezon only. It can contain errors or gaps and is not warranted. See the disclaimer.

Access and plans

Sign in with Google. New accounts are reviewed and approved by our team, and start on the Free plan. Prices are in Indian rupees and are shown in the app before you pay.

• How you pay: by UPI. You then send us the payment details, we check the payment and confirm it. You have Pro straight away while we check, and if we cannot find the payment you return to the Free plan.
• No automatic renewal. Pro does not renew by itself and we never debit your account. We remind you in the app before it ends.
• Payments are final. We do not refund payments, including any unused part of the period. Please read the terms inside the app before you pay.
• Tax: GST is not charged at present while our registration is in process. Once it is complete, GST at 18% will be added, and invoices will be emailed for payments made after that.
• Works on desktop and phone, and can be added to a phone home screen.

Open Sezon