Choose the share count before calculating
TradingView defines free float as issued common shares available for open trading, excluding specified closely held holdings. Float is a subset of total common shares outstanding. Use the float share count for the same security and share class as the volume series; total shares outstanding is a different denominator.
- Save the provider, field name, source URL, reported effective or observation date, and the time you retrieved the value. A retrieval timestamp alone does not establish when the underlying share count changed.
- Check issuer filings and corporate-action announcements for changes after that observation date. Do not assume a provider refreshes float immediately after an offering, conversion, lock-up change, or split.
- If providers disagree, investigate their dates and inclusion rules. Keep the disagreement in the record rather than choosing the smaller number to increase the ratio.
- Use shares on both sides: 8M means 8,000,000 shares. Dollar volume, market capitalization, a dollar-valued public float, and futures contracts cannot be divided by stock float shares to produce this ratio.
Worked example: one denominator, three volume windows
These invented observations use an unchanged, verified float of 8,000,000 shares. Assume the provider reports 3,000,000 pre-market shares, 12,000,000 regular-session shares, and 1,000,000 post-market shares for the same trading date. The sessions do not overlap. M below means million shares.
- At a regular-session checkpoint with 4M shares traded so far, the regular-session ratio is 4M ÷ 8M = 0.5x. Including the earlier 3M pre-market shares gives 7M ÷ 8M = 0.875x. Both are reproducible, but they answer different questions.
- Do not add a daily total to its component intraday bars. Do not sum successive cumulative snapshots: if the counter reads 4M and later 5M, the additional volume is 1M, not 9M.
| Window | Calculation | Rotation |
|---|
| Pre-market only | 3M ÷ 8M | 0.375x = 37.5% |
| Regular session only | 12M ÷ 8M | 1.5x = 150% |
| All three sessions | (3M + 12M + 1M) ÷ 8M | 2x = 200% |
Why one full rotation does not identify new owners
Imagine an 8M-share float in which the same 1M-share parcel is sold eight times during the chosen session. Each sale contributes 1M shares to volume, giving 8M ÷ 8M = 1x. The other 7M shares never trade in this simplified example. If the parcel trades sixteen times, rotation reaches 2x while those other shares still remain untouched.
- One 100-share execution contributes 100 shares of volume, not 200 because it has both a buyer and a seller. A later resale adds another 100.
- A volume total contains no inventory of unique beneficial owners. You cannot infer the fraction of holders who exited, the number of distinct buyers, or how much supply remains available from rotation.
- The ratio also contains no short-position or borrow data. High rotation may accompany a sharp rise, a decline, or repeated trading in a range; it does not establish forced covering or a future squeeze.
Define when cumulative volume starts and resets
Choose regular-session rotation or an explicitly defined extended-session window before collecting data. Record the exchange timezone, trading date, start and end boundaries, and whether the current unfinished bar is included. TradingView documents separate pre-market and post-market screener fields and clears those fields at the start of a new pre-market session; a custom cumulative indicator can use a different reset rule.
- For a regular-session calculation, start at zero at that session boundary and add each eligible bar once. At the next session, discard the prior accumulator. A continuously growing multi-day sum is not daily float rotation.
- For a window that includes pre-market, regular trading, and post-market, retain the counter across those transitions and reset only at the declared next window. Respect holidays, early closes, and the selected exchange calendar.
- If history begins halfway through the session, mark the result partial. Missing bars or missing volume are unknown data, not zero activity.
- Check the feed as well as the chart symbol. TradingView documents differences between its default Cboe US stock feed and primary-exchange data, plus intraday filtering differences. Do not assume every visible volume series covers the whole market.
- When comparing days, use matching session windows and observation times. A 10:00 snapshot and a full-day total measure different amounts of elapsed trading.
Company actions can change the result by a factor of ten
FINRA explains that a reverse split reduces share counts without, by itself, changing proportional ownership. In a hypothetical 1-for-10 reverse split, an 8M-share float becomes 0.8M shares if the float composition otherwise stays unchanged. Post-split volume of 1.2M shares gives 1.2M ÷ 0.8M = 1.5x. Using the stale 8M denominator gives 0.15x, ten times too small.
| Problem | Effect | Action |
|---|
| Forward split with an old, smaller float | Post-split rotation can appear too high. | Match volume and float to the same split basis and effective date. |
| New freely tradable shares after the float snapshot | An understated denominator inflates the ratio. | Verify issuance, eligibility for trading, and provider treatment; an announcement alone is not an updated float. |
| Current float applied to an old chart | Historical rotation may use information unavailable at that time. | Use a dated historical float available at the observation time, or label the result a current-float approximation. |
| Two providers report 8M and 10M float | The same 12M volume produces 1.5x or 1.2x. | Resolve the definitions and dates, or retain both as a sensitivity range. |
Reject invalid inputs before interpreting a threshold
The calculation requires a finite, positive float and finite, nonnegative volume with matching units. A large numerical result is not evidence that its inputs are valid. These boundary cases are independent arithmetic checks, not trading thresholds.
| Inputs or condition | Correct treatment |
|---|
| Verified volume 0; float 8M | 0x for that defined window. |
| Float 0, negative, or missing | Undefined; do not show 0x or infinity as a usable ratio. |
| Volume missing or negative | Invalid or incomplete observation; investigate the feed. |
| Volume 12M; float entered as 8 shares | 1,500,000x is a unit error if the intended float was 8M. |
| Float date unknown or an unresolved split occurred | Arithmetic may run, but the interpretation is unverified. |
Keep a float-rotation research record
Before interpreting a threshold, preserve the inputs and a screenshot with the symbol, interval, session, and observation time visible. Pineify Trading Journal has text-based trading plans, manual trade entry, tags, and closing notes. Use a plan to record the research assumptions and notes to compare any resulting trade with its original rule.
- Suggested record: security and share class; feed; float value, source and effective date; retrieval timestamp; volume window and reset; observed volume; calculated ratio; corporate-action checks; screenshot filename; outcome and deviations.
- Keep the original screenshot with your research files. AI Chart Analysis accepts an uploaded chart screenshot for review of visible chart context. Check its interpretation against the image and the underlying numbers.
- Neither a screenshot review nor this journal workflow verifies or automatically retrieves the float denominator. Supply that evidence yourself. These review steps do not execute trades.
