Beginner's Guide

One row = one period (hour, day, week, month or quarter) summarised by four prices: Open, High, Low, Close.

How to read: Open is the first price of the period, Close the last, High the top and Low the bottom it reached in between.

Example: Daily candle: Open 60,000 · High 62,500 · Low 59,200 · Close 61,800 means price started at 60,000, touched 62,500 and 59,200, and ended at 61,800.

The price at the very start of the period.

How to read: Almost every percentage on this site is measured from the Open, so it is the "starting line" for that period.

The highest price reached at any moment during the period.

Careful: Price may only have touched it briefly. The Close can be far below the High.

The lowest price reached at any moment during the period.

Careful: Price may only have touched it briefly. The Close can be far above the Low.

The price at the very end of the period. For a period that is still running, it is simply the latest price.

How to read: Close compared with Open decides whether the period counts as a gain (green) or a loss (red).

Coordinated Universal Time: the single worldwide clock all periods on this site start and end by.

How to read: A "day" runs 00:00–23:59 UTC, a "week" starts on Monday 00:00 UTC. Convert to your own time zone before acting on a time-of-day pattern.

Example: If you live in UTC+2, the daily candle opens at 02:00 your time.

The price to buy or sell Bitcoin right now for US dollars, as opposed to futures or options.

The most recent price we have for BTC in US dollars, taken from the latest daily candle.

Careful: The data comes from Binance and is refreshed regularly, not tick by tick. Do not use it for split-second trading.

How far the latest price is from the previous UTC day's close, in dollars and in percent.

How to read: Green means price is above yesterday's close, red means below.

Careful: It compares against the last daily close, so it is not a rolling 24-hour window.

How far the latest price is from the previous hour's close, in dollars.

How to read: Green means price is above the last hourly close, red means below.

How much price has moved since this week, month or quarter opened (so far, if it is still running).

Example: Month opened at 60,000 and price is 63,000 now: change is +3,000 USD and +5.00%.

How much the Close is above (gain, +) or below (loss, −) the Open of the same candle, as a percentage.

How to read: Percentages let you compare 2015 and today fairly: +5% is +5% whether BTC costs 300 or 90,000.

Example: Open 100 → Close 105 = +5.00%. Open 100 → Close 96 = −4.00%.

Careful: This is Close versus that candle's own Open, not versus the previous candle's Close.

Close minus Open in dollars. Positive = the period ended higher than it began.

Up if the period closed at or above its Open, Down if it closed below.

How to read: It only describes that single period. It is not a forecast.

The distance between the highest and lowest price in the period, in dollars.

How to read: A wide range means a volatile, fast-moving period; a narrow range means a calm one.

The range (High − Low) as a percentage of the Open.

How to read: Use it to compare how "wild" different periods were, regardless of the price level.

Example: Open 100, High 108, Low 96 → range 12 → 12.00%.

How much Bitcoin was traded during the period (in BTC). "K" means thousands.

How to read: High volume means many participants were active, so a price move is usually considered more meaningful.

A grid with one coloured cell per calendar month: green if Bitcoin closed the month up, red if down, stronger colour for bigger moves.

How to read: Read across a row to see one year; read down a column to see how the same month behaved in every year. The bottom rows give that month's median change and how often it closed up.

Careful: Each month has only one observation per year, so a column is built from a small number of years.

A bar chart that sorts every day into a bucket by how much it moved (close versus open) and counts how many days fall in each bucket.

How to read: Tall bars near 0% mean most days are small moves. Long, thin tails on either side are the rare big days. The shaded zone holds the middle 68% of days, the same idea as P_68.

Careful: The shape can change a lot between calm and wild periods. Use Stats From / To to compare.

How far Bitcoin moved over the whole calendar year, from the first month's open to the last month's close. The current year is year-to-date.

How many past cases the numbers are based on.

How to read: More samples = more reliable. With fewer than about 10 cases, treat a percentage as a hint, not a rule.

Example: "Up 80%" from 5 samples only means 4 of 5 cases. "Up 55%" from 1,000 samples is much more trustworthy.

A dimmed cell is based on very few past cases (fewer than 5), so it can swing wildly.

The middle value: half of the cases were lower and half were higher.

How to read: Unlike the average, a few huge months or days cannot drag it around, so it describes the typical case better.

Example: Monthly changes of +2%, +3%, +5%, +8% and +300%: the average is about +64%, the median is +5%.

The ordinary mean of all past cases in that column.

Careful: A few extreme days can pull the average away from what usually happens. Compare it with P_68.

A "typical but not extreme" size of move: in about 2 out of 3 past cases the move was no bigger than this.

How to read: Think of it as the edge of the normal zone. Moves beyond P_68 happen, but only roughly 1 time in 3. For loss/low columns the mirror value is used (the move was no worse than this about 2 times in 3).

Example: High %Delta P_68 = +2.4% → on about 68% of past days, price never climbed more than 2.4% above the open.

Careful: It describes the past. It is not a guarantee about the next period.

The single largest move ever recorded in that column (for loss/low columns: the deepest drop).

How to read: Shows the worst/best case history has seen so far. Highlighted in yellow in the tables.

Careful: A new record can still happen. Extremes are rare, so do not plan around them being typical.

The share of periods that closed above their open (Gain) or below it (Loss).

How to read: Around 50% means a coin flip; 60% Gain means green closes were noticeably more common in that slot.

Example: Monday Gain 54% / Loss 46% → out of every 100 Mondays, about 54 closed green.

Careful: Gain + Loss can be just under 100% because the rare exactly-flat periods are counted separately.

How big the closing move was, measured from the Open. The Gain block only looks at green periods, the Loss block only at red ones.

How to read: Answers "when it closes green, how much does it usually gain?" and "when it closes red, how much does it usually lose?"

How far above the Open price climbed at its best moment, in percent.

How to read: Useful as an upside "reach": the P_68 value is a realistic target zone, Max is the stretch.

How far below the Open price fell at its worst moment, in percent (always zero or negative).

How to read: Useful as a downside "dip depth": the P_68 value is a realistic pullback zone, Max is the deepest ever.

The whole High-to-Low swing of the period as a percentage of the Open, i.e. how much room price used.

How to read: Higher numbers = a more volatile slot. Handy for judging how wide to place a stop or a target.

How far the high was above the period's Open, in percent: the best-case upside of that period.

Example: O2H +12% for a month: at some point that month price was 12% above where the month opened.

How far the low was below the period's Open, in percent: the worst dip of that period.

Example: O2L −8% for a month: at some point that month price was 8% below where it opened.

How much price gave back from the period's high to its close, in percent (zero or negative).

How to read: A big negative number means the rally faded before the period ended.

How much price recovered from the period's low to its close, in percent (zero or positive).

How to read: A big positive number means buyers stepped in after the dip.

The row label: a weekday (or an hour of the day) in UTC. Every past Monday, or every past 14:00 hour, is grouped into one row.

The date range the statistics are calculated from. Narrow it to see how a pattern behaved in recent years only.

How to read: Markets change. If a pattern looks strong over all years but weak over the last two, be careful.

Buttons that narrow the table to only the years or periods that match, for example only months that closed Up.

How to read: Combine with the Samples number: the more filters you apply, the fewer cases remain.

A named time block inside a bigger period. The slot columns say in which block the High or Low happened.

How to read: If highs keep landing in the same slot, that time block is historically where rallies peak.

The UTC hour of the day (H00–H23) in which the day's High or Low was made.

Example: Highs @ H14 = the day's top was made between 14:00 and 14:59 UTC.

Careful: One day proves nothing; look at the hour that repeats most often.

Which 4-hour block of the week (numbered from Monday 00:00 UTC) contained the week's High or Low. #1 = Mon 00–04, #42 = Sun 20–24.

How to read: Use the Slots menu to switch to 6-hour blocks (28 per week).

The calendar day of the month (D01–D31, UTC) on which the month's High or Low was made.

Example: Hi 1d Slot D03 = the month's peak came on the 3rd.

Which week counted from the quarter's start (W1–W14) contained the quarter's High or Low.

Which 5-minute block of the hour (S1–S12) contained the hour's High or Low. S1 = minutes 0–4, S12 = minutes 55–59.

The weekday (UTC) on which the high or low of that period was made.

The position inside the month (Day of M = date, 1–31) or week (Day of W) on which the high or low happened.

How to read: Early-month or early-week highs mean the rally peaked early and faded; late ones mean strength into the close.

How many weeks after the quarter opened the high or low was made (1 = first week of the quarter).

Which week of the calendar month: W1 = days 1–7, W2 = days 8–14, and so on. W5 is the short tail of the month.

The ISO week number, counted from Monday. Comparing the same week number across years shows yearly seasonality.

The ISO year and week number of the row, for example 2026-W14.

Whether the whole calendar year that row belongs to ended higher (Up) or lower (Down) than it started.

How to read: It lets you ask "in years that ended Up, how did this month behave?" — a way to compare bull-market and bear-market years.

Careful: It is known only in hindsight (for the current year it is year-to-date), so it is context, not a trading trigger.

Up if that one week, month or quarter closed above its Open, otherwise Down.

Where the year sits in Bitcoin's roughly 4-year halving cycle: H = a halving year (2012, 2016, 2020, 2024), H+1 = the year after, M = the middle year (two years after), H−1 = the year before the next halving.

How to read: Bitcoin's new supply is cut in half about every four years (the "halving"). Historically the price behaved differently in each phase, so these buttons let you compare like with like.

Careful: Only a handful of cycles exist, so each cycle group has few samples.

A scheduled event, about every four years, that cuts the reward miners receive for new blocks in half, slowing the creation of new bitcoin.

Whether that month, week or quarter closed above its Open (Positive) or below (Negative).

Bull when the daily close has been above its 200-day average, bear when below. A change only counts after 10 days in a row on the new side, so short dips do not flip it.

How to read: The self-tuning price targets can learn only from periods since the last trend change. They use that option when it has recently been the more accurate one.

Careful: The 200-day average is slow by design: it confirms a new trend weeks after it started.

A number of same-coloured candles in a row: several green (up) days or several red (down) days.

Example: Up 3 = the last three daily candles all closed above their open.

How many consecutive up (or down) daily candles have just happened, counting the latest one.

Which way history leans for the next day after a streak of this exact length, and by how much.

Careful: A lean of 52–55% is a small edge, not a prediction. Check the sample size.

Out of all streaks of this colour, the share that lasted exactly this many days.

How to read: Long streaks are rare, so the numbers shrink quickly to the right.

Out of all days of this colour, the share that were part of a streak of this exact length.

After a streak this long, the chance the following day closed up (Positive), down (Negative) or unchanged (Flat).

How to read: Compare with roughly 50% (a coin flip). Only large gaps from 50%, backed by many samples, are meaningful.

Percentages from very few cases are pulled toward the long-run average so that one lucky case never shows as 0% or 100%.

How to read: The fewer samples a cell has, the more it leans on the general average.

The chance the current streak gets extended all the way to this length, by multiplying the odds of each further step.

Example: Down 4 now; 5th day 35%, 6th day 15%, 7th day 5.5% means a 7-day down streak is a long shot.

The average percentage change of the day after a streak of this length.

A step-by-step projection of the next five days, blending the streak odds with how that weekday usually behaves.

Careful: Each extra step is less certain than the one before. Treat days 4–5 as a rough guide only.

Chance the day closes up according to the streak history only (smoothed).

Share of past days with that same weekday that closed up.

A weighted blend of the streak odds and the weekday odds (streak counts twice as much). Up and Down add up to 100%.

The streak length the projection assumes going into that day.

The future date (UTC) the row makes a projection for. D+1 is tomorrow.

Levels that price reached before the end of the period in most past periods, measured from the same point in time as now and applied to the current price.

How to read: Up target = reached in about 7 of 10 past periods; stretch = about 1 in 3. "Most likely" is when the remaining high or low was usually made. Both the up and the down target can be hit in the same period.

Careful: History, not a forecast of direction and not advice. If the "tested" rate is much lower than the chance, trust the target less.

How often the price targets actually came true: every prediction is stored and checked against what really happened.

How to read: Every 30 minutes the site re-runs the same method on the recent past using only data available at each moment, and also logs a live prediction. It then switches to the setting whose chances came true most accurately. "Time window right" is how often the high or low really happened inside the predicted time window.

Careful: A good track record in the recent past does not guarantee the next period. Live results are the most honest number once enough have accumulated.

For each day, week, month or quarter in the past we noted WHEN its high and low happened. The chart shows how often each time slot held the high (green) or the low (red).

How to read: Tall bars are the most likely times. The dashed line is what an even spread would look like. "Still ahead" is the chance the high or low of the current period has not happened yet, given the time now.

Careful: Highs and lows bunch near the start and end of a period for almost any wandering price, so a tall bar at midnight is a tendency, not a trading edge on its own.

Where price has usually travelled during this day, week, month or quarter, measured from the period's open.

How to read: We take the recent history of the same kind of period and apply its typical highs and lows to the current open.

Careful: It is a statistical range, not a forecast or advice. Price can and will leave it.

The "normal" zone: in about 2 out of 3 past periods the high stayed below and the low stayed above these levels.

How to read: The amber band on the meter. "reached" means this period has already traded beyond it, i.e. a stronger-than-usual move; otherwise the card shows the distance from the current price.

The most extreme high and low seen in the same history window, applied to the current open.

How to read: Marks both ends of the meter. Reaching a stretch level is rare.

The bar shows the whole historical span from stretch low to stretch high. The amber part is the likely band, the thin white tick is the period's open, and the coloured dot is the current price.

The share of comparable past periods that closed above their open.

How to read: Near 50% = no edge. Even 55–60% is only a slight tilt.

Careful: It is a long-run base rate and does not react to today's news or trend.

When we replayed the method on past periods, how often both the high and the low stayed inside the likely band.

How to read: Because two limits must both hold, roughly 50–60% is what a well-behaved band delivers. It is a health check for the method.

When we replayed the method on past periods, how often its up/down lean turned out right.

How to read: Around 50% means no better than a coin flip. The card only reports a direction edge when the result beats 50% by more than chance would explain for that number of tested periods.

A fairness test: for each past period we rebuild the forecast using only the data that existed before it, then compare with what really happened.

How to read: It shows how the method would have done historically. "N days tested" is how many past periods were scored.

Careful: Past accuracy does not guarantee future accuracy.

A tendency for a calendar period (a month, a weekday, a week of the year) to behave similarly year after year.

How to read: Compare "this month so far" with the average of the same month in earlier years, to see if it is running hotter or colder than usual.

Careful: With only 10–15 years of data per month, a seasonal pattern can easily be a coincidence.

The average open-to-close change of this calendar month across all years on record.

How many daily candles are in the database. More history makes the statistics steadier.

A tiny line chart of the last ten daily closes, coloured green if price ended higher than it started, red if lower.