Liquidity: Sweep, Rejection and Level Reaction

Lessons in this module

1. 12.1

   A liquidity sweep as an event, not a zone

   Tell a zone where stops cluster apart from the moment it gets swept: a move beyond the boundary with a return inside on higher volume is an event with an exact candle, not an area on the chart

   13 minAccess in Telegram

2. 12.2

   Anatomy of a level rejection: wick, close depth, range

   Break a rejection candle into parts and understand why a long wick proves nothing by itself until the close has gone deep enough back into the zone

   14 minAccess in Telegram

3. 12.3

   Sweep → rejection: distribution and absorption

   Combine two events into one model: first the market takes liquidity, then in the same zone it refuses to go further; at the top this is distribution, at the bottom absorption; in history this combination gave no direction either with the higher-timeframe trend or against it

   14 minAccess in Telegram

4. 12.4

   Binding events to one zone: overlap and tolerance

   Check that the sweep and the rejection really happened at the same level: the share of zone overlap and the midpoint tolerance in ATR are two ways not to glue a random pair together

   12 minAccess in Telegram

5. 12.5

   Reaction decay at a level

   Compare the strength of each new reaction with the previous one: a level does not die suddenly but through a series of ever weaker reactions

   12 minAccess in Telegram

6. 12.6

   Swing freshness: the age of structure as a filter

   Look not only at whether a level was swept but at how long it stood: a fresh extreme and one that has hung there for two hundred bars collect different stops, and the reaction to them differs

   13 minAccess in Telegram

7. 12.7

   Market regime: squeeze, expansion, trend, range

   Separate two different questions, where price is going and how widely it moves: ATR relative to its baseline answers the second, and it is what tells a squeeze from an ordinary range; and know what history showed: right after a squeeze begins, the next day is wider than usual only 46% of the time, and calmer in percent of price; “the move after a squeeze is wider” only appears when measured in the squeezed ATR

   13 minAccess in Telegram

8. 12.8

   Event score and confidence bands

   Combine different kinds of features into one number from 0 to 100 and understand why the scale is cut into bands instead of showing an exact figure

   13 minAccess in Telegram

9. 12.9

   A separate threshold for each event type

   Do not compare different events on one scale: a breakout and a trap collect a different number of features, so a common threshold of 60 is strict for one and a formality for the other

   12 minAccess in Telegram

10. 12.10

    Expected outcome: continuation or reversal

    Turn an event plus context into a forecast of behavior: the same absorption with the trend and against it promises different things

    13 minAccess in Telegram

11. 12.11

    Event heat map and zone clustering

    Show history as density rather than a list of labels: nearby zones merge into one tile, brightness shows strength, old events fade, and the chart does not turn into a mess

    12 minAccess in Telegram

12. 12.12

    Event density as a measure of noise

    Read how often signals fire as a signal of its own: when there are too many events per bar, it is not an abundance of opportunities but a sign that the filters are too loose

    12 minAccess in Telegram

13. 12.13

    Signal statistics in R: a win rate that does not lie

    Measure a series in units of risk and count a win only when the target is hit before the stop, otherwise the win rate looks great while the trading loses money; split the result by entry conditions, “with the condition” versus “without”, and do not trust the difference until it has held up over a hundred trades

    13 minAccess in Telegram

14. 12.14

    Real time versus history: the bar close as a condition

    Understand why logic that is flawless in history may never work live: on an unclosed bar the high and the close are not yet what they will be

    13 minAccess in Telegram

15. 12.15

    The sweep level as a magnet: odds and time until it is taken

    Read the high of the sweep candle as a target that price will most likely return to: know how often and how fast such a level is taken on its timeframe, why the time of day and a calm market speed it up, which of the two nearest levels will be taken first and why that is pure geometry of distances, and why this is a map of targets rather than an entry, since a stop at the same distance wins no more than half of the trades

    13 minAccess in Telegram

16. 12.16

    A sweep on one exchange or on all of them: venue agreement

    Check whether liquidity was taken on other exchanges, not only on yours: crypto has no single order book, and a sweep that turned out to be a breakout on another exchange, or did not happen there at all, reverses noticeably worse; but confirmation everywhere gives no edge by itself either

    13 minAccess in Telegram

17. 12.17

    Climax on a sweep: the reversal myth and a slow magnet

    Test the popular model “extreme volume and range on a sweep means a reversal” against history: it gives no reversal, but price takes the level of such a candle more slowly, and the label figures for it are different

    12 minAccess in Telegram