A candle compresses many transactions into a few prices. Order Flow adds detail, but every feed covers a particular venue, instrument and interval. Check those boundaries first. An empty or stale feed cannot support the same conclusion as current executed volume.
01
The order book is intent, not a trade
Resting orders can disappear. A wall is evidence of visible liquidity, not a guarantee.
Watch whether liquidity remains as price approaches.
Compare resting size with actual trades.
Account for spoofing and cancellations.
02
CVD measures aggressive imbalance
Cumulative Volume Delta compares aggressive buying and selling.
Price up with weak CVD can be fragile.
Improving delta is not an automatic bottom.
Use a relevant execution venue.
03
Absorption: flow without proportional price movement
Large aggression with limited price progress can indicate the other side is absorbing.
Look near meaningful structure.
Confirm with subsequent price reaction.
Absorption is not instant reversal.
04
Flow is one layer
HOSTuvo combines flow with price, microstructure, derivatives, context and cost.
Flow can confirm or contradict structure.
Missing data is not zero.
The execution gate remains separate.
05
Reconstruct the book before interpreting imbalance
A live book begins with a snapshot and then applies updates. A reconnect, missing update or new snapshot can invalidate its previous reconstruction. Compare quantities at consistent distances from the mid-price rather than arbitrary row counts. A large bid that vanishes as price approaches differs from a bid that remains while trades execute against it.
Check continuity and observation age.
Separate cancellations from executed volume.
Compare depth with price impact for your size.
06
Use an explicit CVD example
In a hypothetical interval with 120 buyer-initiated and 100 seller-initiated units, interval delta is +20. CVD adds such deltas from a chosen starting point. Another reset, classification or venue can produce a different value without either feed being dishonest. Compare the same unit and interval; ask whether positive delta actually moved price or met persistent liquidity.
State the reset point and unit.
Keep spot and perpetual data distinct.
Divergence does not imply inevitable reversal.
07
Assess absorption as a sequence
Repeated aggressive buying at resistance with little price progress is consistent with selling liquidity absorbing demand. It can also precede a breakout if supply is consumed. Observe replenishment, subsequent trades and whether price leaves or accepts the area. Define your observation window beforehand so the explanation is not selected only after the result.
Order flow is a set of observations, not a shortcut to knowing the future. Trade feeds describe executed activity; order-book feeds describe displayed liquidity that can change or disappear. Read them with a source, timestamp, calculation unit and meaningful price level. The guides below connect those observations to spread, slippage and a testable scenario.
Know the feed
Verify which venue and instrument supply the trades and whether the connection has gaps.
Keep the calculation explicit
For CVD, document taker-side classification, volume unit and reset point before comparing charts.
Verify the execution market
Check the available bid and ask for the intended venue and size; a reference-market chart is not a guaranteed fill.
Every executed trade has both a buyer and a seller. Aggressor-side analysis asks which side crossed available liquidity, rather than counting buyers and sellers as if one trade could contain only one. Resting bids and asks describe displayed intentions; trade…
First, define the context and horizon of the analysis.
Confirm your applications with independent data layers.
Take into account the quality of feed, spread and real cost of execution.
CVD accumulates signed executed volume from a defined starting point. Changing that anchor changes the displayed level, so the level itself is not comparable across arbitrary chart resets. Base-asset units and quote-currency notional answer different…
CVD is a cumulative calculation from classified trade volume.
Divergence is context to investigate, not an automatic reversal signal.
Keep the venue, unit, reset point and feed quality explicit.
Absorption is an interpretation of substantial aggressive trading meeting limited price progress. It may involve replenishing resting orders, multiple participants or liquidity that is not fully visible in one snapshot. A static large order is insufficient…
First, define the context and horizon of the analysis.
Confirm your applications with independent data layers.
Take into account the quality of feed, spread and real cost of execution.
A footprint groups executed trades by price and time or another bar construction rule. The displayed bid-side and ask-side volumes depend on the feed's trade classification and the chart's aggregation settings. A large imbalance at one price is not…
First, define the context and horizon of the analysis.
Confirm your applications with independent data layers.
Take into account the quality of feed, spread and real cost of execution.
A large notional order far from the current market may have little relevance to a short-horizon decision. Compare the wall with normal depth at a similar distance, and state whether size is measured in the base asset or quote currency. The same nominal size…
First, define the context and horizon of the analysis.
Confirm your applications with independent data layers.
Take into account the quality of feed, spread and real cost of execution.
Trades per second, notional volume per second and the frequency of large matches measure different aspects of activity. One order may be split into many executions, so a rising trade count need not mean more independent participants. Compare the rate with a…
First, define the context and horizon of the analysis.
Confirm your applications with independent data layers.
Take into account the quality of feed, spread and real cost of execution.
Level 2 generally aggregates resting quantity by price rather than identifying each individual order. It can reveal how displayed depth is distributed, but cannot fully reconstruct queue position or trader identity. A usable live book combines a starting…
First, define the context and horizon of the analysis.
Confirm your applications with independent data layers.
Take into account the quality of feed, spread and real cost of execution.
A displayed order can disappear for many reasons. Spoofing involves intent, which a public snapshot generally cannot establish. Analyse observable behaviour first: repeated placement, withdrawal as price approaches, activity on the opposite side and the…
First, define the context and horizon of the analysis.
Confirm your applications with independent data layers.
Take into account the quality of feed, spread and real cost of execution.
The spread is the difference between the best ask and bid; slippage compares an actual or estimated fill with a specified reference. State that reference before adding costs. If entry and exit are already simulated by walking the ask and bid sides of the…
Use the actual venue, quote currency and order size.
Show fee assumptions and avoid double-counting the spread.
Distinguish a chart target from a documented executable fill.
A leveraged position may be liquidated when the venue's margin conditions are breached; the relevant reference can be mark price rather than the last trade shown on a chart. Rules differ by product and margin mode. Forced closing can add directional pressure,…
First, define the context and horizon of the analysis.
Confirm your applications with independent data layers.
Take into account the quality of feed, spread and real cost of execution.
A sweep interpretation usually starts with price moving beyond an obvious prior extreme and then examining what happens next. Stops are generally not fully visible in a public book, so the chart alone cannot prove how much hidden stop liquidity existed or who…
First, define the context and horizon of the analysis.
Confirm your applications with independent data layers.
Take into account the quality of feed, spread and real cost of execution.
A common imbalance measure compares displayed bid and ask quantities, but its meaning depends on which levels are included. Ten levels can span very different price distances across pairs with different tick sizes. A fixed basis-point band around a reference…
First, define the context and horizon of the analysis.
Confirm your applications with independent data layers.
Take into account the quality of feed, spread and real cost of execution.
