When does KAS rise and fall? Hourly and weekly returns

Understanding KAS hourly performance

No statistically confirmed hourly pattern for KAS in this window. Snapshot: 2026-10-08.

What intraday seasonality describes

Intraday seasonality asks whether observations collected at the same hour or weekday have behaved differently over a historical sample. For KAS, the heatmap groups closed spot-market candles by UTC opening time. It is a description of past observations, not a schedule of future price moves. A coloured square summarizes many separate hours, including quiet periods, news shocks and different market regimes. The original observations still vary considerably even when their summary looks consistent. Read the value together with sample size and uncertainty rather than treating colour as a signal.

Reading returns and activity

The default value is the median logarithmic close-to-open return. The median emphasizes the middle observation, while the mean is more sensitive to unusually large moves. Win rate records how often the close exceeded the open; a flat candle remains in the denominator. Volatility measures dispersion of hourly returns without annualization. Daily volume share compares an hour with its complete UTC day on the selected source. These measures answer different questions. Frequent small gains can coexist with a negative mean, and high activity says nothing by itself about price direction.

Why trading sessions matter

Crypto trades throughout the week, but participants do not all follow the same working day. Tokyo, London and New York provide useful reference windows for changing regional activity. Their overlaps can coincide with more participants being active, without causing a reliable direction of return. The session cards use complete eight-hour observations and historical timezone rules. Europe and the United States switch daylight-saving time on different dates. Consequently, a fixed UTC hour may belong to a different regional session depending on the date. The methodology makes this distinction explicit.

Uncertainty is part of the result

Looking across many hours makes an attractive maximum almost inevitable, even when there is no repeatable effect. The page therefore distinguishes descriptive values from confirmed historical differences. Its approximate weekly cluster multiplier bootstrap respects grouping within weeks, and simultaneous intervals account for comparisons within the displayed metric and view. Median inference uses a quantile influence approximation. Small samples and degenerate distributions are deliberately left unconfirmed. A 90-day hour-by-weekday cell usually contains only about thirteen observations, so its bright colour alone is weak evidence for a persistent pattern.

History and source coverage

The requested window is not always the available history. Young listings, exchange outages and restricted historical endpoints can leave fewer observations for KAS. The page exposes the valid-hour count, window coverage, source exchange and snapshot date. Missing candles are not replaced with flat returns. Volume shares require complete days, and session returns require complete sessions. The market comparison includes only coins with sufficient full-window coverage. These rules reduce hidden inconsistencies, but do not eliminate source-specific differences or make an exchange price a universal market price.

Comparing the current market

The market modes represent the current non-stable universe ranked by capitalization. Coin median describes broad behaviour; equal-weight and capitalization-weighted indices use average hourly returns with their respective weights. Current capitalization weights are disclosed because historical weights are not reconstructed here. This means the retrospective excludes assets absent from the current selection and carries survivorship bias. Top-ten and top-hundred comparisons may differ substantially. Neither is a fund return, and neither includes portfolio rebalancing costs, fees or a guarantee that every constituent could have been traded in practice.

Patterns can change

The 90-day, one-year and three-year views reveal whether the observed profile changes with the sample. These windows overlap, so agreement between them is not independent confirmation. News, liquidity migration, listing changes and broader market conditions can shift an apparent strongest hour or weekday. Weekend comparisons use complete daily returns rather than adding hourly medians, which would be a different statistic. A visible gap between weekdays and weekends must still be read with uncertainty. The saved profile is evidence about a stated period, not a claim that the same ranking will persist.

Using the page responsibly

Use this tool to explore historical structure, formulate questions and compare observations. A statistically detectable return can be too small to survive spread, fees, slippage, funding or taxes. The heatmap does not model a trading strategy, assess your financial position or choose an entry price. Shared cards preserve the selected window and snapshot so readers can inspect the underlying evidence. When confirmation is absent, that absence is itself a useful result. Past performance is not a reliable forecast, and no hour shown here is a promise that buying or selling will succeed.