What you may want to know about BTC 15-minute reference signals

24K is a 15-minute reference-signal and method-group research tool for the BTC market. It shows the directional calls of multiple method groups for the same time window in one format, and checks them against the actual direction of public market candles after each cycle closes.

Its main features are live cycle views, method-group comparison, the Research Center, history review, hit and miss statistics, net wins, win rate and sample-coverage analysis.

24K does not provide trade execution, managed accounts or any promise of returns.

24K suits users who want to observe BTC market information in fixed time windows, compare different method groups, study historical samples or build a personal observation routine.

It works best as a tool for organising data, comparing methods and researching history — not as an automated-trading, managed-account or return-promising service.

No. 24K is a market information and directional-reference research tool. It is not investment, financial, legal or tax advice, and it is not an instruction to buy, sell, hold or make any other trade.

Crypto-asset prices can be extremely volatile. Historical statistics do not represent future performance; users should judge independently according to their own circumstances and bear the related risks themselves.

24K and Polymarket, Predict, Kalshi and Gemini are different products or brands. 24K does not represent these platforms and has no partnership, affiliation, agency, endorsement or data-supply relationship with them.

24K is positioned as a BTC 15-minute market information and method-group research tool: it shows directional references in fixed time windows and checks them historically against public market candles after each cycle closes. Names such as Polymarket, Kalshi, Predict and Gemini may appear in searches about prediction markets, event forecasting, crypto information or AI tools, but each platform's service model, data sources, contracts or product rules are as stated on its own official pages.

Prediction markets generally organise tradable positions around the outcome of an event. 24K does not issue contracts, set contract prices, hold funds or execute trades; it organises BTC's 15-minute market cycles into a research framework, checks directional references against the real market direction after the close and keeps them as a reviewable history.

If you want to compare products, first check four things: whether it provides market information, whether it has event contracts or trading features, what its outcomes are based on, and whether the service amounts to investment or trading advice. 24K provides only the reference signals, method-group statistics and history review published on its own website, and does not provide services on behalf of other platforms.

No. 24K does not manage money for users, hold their assets or execute trades, and it does not make investment decisions on their behalf.

What it provides is public market information, directional references, method-group statistics and history review.

A BTC 15-minute reference signal is a directional reference given for each 15-minute market cycle. 24K records the direction output by each method group separately from the final direction of the actual market candle, and judges after the close whether it was a hit.

There are 96 cycles of 15 minutes each day. A cycle that has not closed is marked "In progress" and is not counted early in hits, misses, net wins or win rate.

24K's market reference data comes from public quotes of major exchanges, e.g. Binance. The actual trading pair, data source and fallback strategy are as disclosed by the product at the time.

The system reads data in fixed absolute time windows and uses data validation, anomaly checks and fallback mechanisms to minimise the impact of a single source failing, network delays or brief fluctuations.

24K judges direction strictly by the result of the matching Binance candle:

• a close above the open counts as "up";
• a close below the open counts as "down".

A cycle enters the final judgment only after it closes. Confirmed historical results are not arbitrarily revised because of later price moves.

"In progress" means the current 15-minute cycle has not ended and its final direction is still undecided.

Cycles in progress are not counted in:

• hits;
• misses;
• net wins;
• win rate;
• Research Center statistics;
• final historical results.

The related results are updated only after the cycle closes and is confirmed.

Because an incomplete cycle is not a confirmed outcome. Counting a candle that has not closed in the historical statistics would mix live information with final results. 24K keeps these two states separate by design.

24K observes in 15-minute cycles. After each cycle closes, the system reads the real market result and updates the related statistics; the product may run a second confirmation after the close to reduce the effect of data delays or brief anomalies.

Times and dates are shown in the visitor's local time, and "today" in the Research Center and Live Table follows the local date. Time conversion does not change the underlying time windows, direction rules or historical results.

A method group is an independent method that produces directional calls. Different groups can be based on different data dimensions, time windows or decision logic, and each outputs its own directional signals.

24K pages may show method groups such as Breakout, Volume, Divergence, Trend, Reversal, MA, Gap, Pattern, Swing, Momentum, Support, Resistance, Turnover, Rhythm and Formula.

Each method group builds up its own history of hits, misses, net wins, win rate, settled count and sample coverage.

A team is a label 24K uses to group and display different time systems and signal columns; it is not an independent method of calling direction.

The pages currently include the Sydney, Japan, Hong Kong, Dubai, UK and New York teams. After choosing a team, you can see its start-time columns and method-group data.

A method group, by contrast, is the basic unit that actually produces directional calls and is tracked in the historical statistics.

A start-time column marks the point in time from which a method group starts giving signals. For example, "from 10:00" means the group starts giving signals from the time window that corresponds to 10:00.

Some columns cross midnight; for example, the later cycles of "from 21:00" may continue into the next day. The system joins the consecutive time windows automatically and marks them with labels such as "next day".

Hits are the number of times a method group's direction matched the real market direction of the corresponding cycle.

For example, if a method group called "up" and the real market direction after the cycle closed was also "up", it counts as one hit.

If the group's direction differs from the real market direction, it counts as a miss. Only cycles that have closed and been confirmed enter these statistics.

Net wins are the difference between hits and misses among settled samples.

Net wins = hits − misses

For example: 30 hits and 20 misses give net wins of +10.

Positive net wins mean there were more hits than misses in that period; negative net wins mean there were more misses than hits.

Net wins are not a rate of return and do not represent actual trading profit or loss.

The win rate is the share of a method group's settled samples in which its direction was right.

Win rate = hits ÷ settled × 100%

For example: 30 hits out of 50 settled gives a win rate of 60%.

The win rate only reflects the historical share of correct direction calls; it is neither the probability of future success nor an actual investment return. Read it together with the settled count, covered days and statistical time window.

A negative net result is still statistical information about historical direction consistency. 24K shows it as a historical observation, not as a recommendation to trade in the opposite direction.

The all-group consensus is how far multiple method groups agree on the direction of the same 15-minute cycle.

For example, if 10 of 15 method groups call up and 5 call down, the all-group consensus for that cycle is "up".

The all-group consensus is not a separate method group and does not mean the market will move in that direction. It only reflects how the directions of different groups are distributed in a given cycle.

On 24K's signal pages:

• green means the up direction;
• red means the down direction.

The direction colours do not themselves mean a hit or a miss, nor a profit or a loss. Hits and misses are shown separately with circles, hollow dots, X marks or other result markers depending on the current view.

Always go by the legend currently shown on the page.

The Research Center shows historical statistics and sample performance by method group, team and time window. Common metrics include:

• hits;
• misses;
• net wins;
• win rate;
• settled bars;
• days covered;
• recent performance and history review.

Method groups may run in different time ranges and have different numbers of valid samples, so avoid drawing conclusions from a single win rate or one day's results.

Sample coverage is the number of days within a given period on which a method group actually produced valid signals that were counted.

For example, "covers 25 / 30 days" means the group had valid coverage on 25 of the 30 days in the period.

Sample coverage helps you judge whether the statistics rest on a reasonably complete observation period. When comparing method groups, look at sample size, covered days and the statistical time window together.

No single metric should be read in isolation. Win rate needs sample size, net wins need a defined period, and historical performance needs covered days. The real research question is not "which number is highest?" but "what does this number represent, and on which sample is it based?"

24K's statistical principle is "judged at the close, fixed definitions, reviewable results". Once a cycle has closed and been confirmed, its historical result should not be arbitrarily changed because of later price moves.

The history view uses the same direction rules and statistical definitions as the live view, so you can check results by date, time window and method group.

We suggest reading in this order:

1. first check whether the current cycle has closed;
2. then look at each method group's direction and the distribution across all groups;
3. then look at hits, misses, net wins and win rate;
4. at the same time confirm the settled count, sample coverage and statistical time window;
5. finally use history review to see whether the results persist.

Do not draw conclusions from a single cycle, a single method group, a single win rate or one day's net wins.

It depends on the features currently published on the website. Some live content can be viewed without logging in; after logging in you can usually use account features such as watchlists, personalised observation and alerts. Full details, history range and other advanced benefits are as described on the membership plans page at the time.

Membership benefits are not investment advice or a promise of any return.

You can report it through the "Feedback" entry on the website. To help us check, include the page name, date, time window, team or method group, your local time zone and a screenshot or description of the problem.

Please do not include passwords, API keys or other sensitive information in your feedback.