OFS F6 Short Reversion
The short side of the reversion trade. When a candle closes far ABOVE the volume weighted average price of the session and the order flow shows buying being absorbed, the strategy sells and aims back at that average. Three positions a day, a hard 200 tick stop.
| Current version | 0.6.6 · 2026-08-27 |
| Compatibility | ATAS 8.0.14+ and ATAS X 8.0.14+ |
| Requirements | Windows 10/11, .NET 10, ATAS or ATAS X, a 1000-tick MNQ chart with cluster (footprint) data and a connected trading account. No external data feed or API subscription is needed - everything |
| Updates | All future updates included |
| Delivery | ATAS catalog (automatic) + download in your account |
Almost ready for your charts
OFS F6 Short Reversion is built and measured — it is just not in the shop yet, so there is no price and no Add to cart to look for on this page.
Everything you see here is already true — the rules it follows, the numbers it produced and the warnings that come with them. Have a read now and decide later: the moment it goes on sale, the license options and the free trial appear right here.
Want to hear the day it lands?
Release news only. No trading tips, no spam — unsubscribe in one click.
Description
WHAT IT DOES
F6 is the mirror of F5. It watches the day's volume weighted average price and how far price has
stretched ABOVE it. When a candle CLOSES 2.5 standard deviations above that average, the strategy sells at the
OPEN OF THE NEXT CANDLE and aims back at the average. Nothing is predicted - the trade is a bet
that an unusually large stretch away from the day's centre gets pulled back.
HOW A TRADE IS BUILT
Entry a closed candle 2.5 standard deviations above the day's average price. A market order goes out the
moment that candle closes and fills on the opening price of the next one. What is drawn
on the chart is what the account does - the two cannot drift apart.
Target the day's average price at the moment of entry. Fixed; it does not move.
Stop the same distance as the target, capped at 200 ticks.
Exit stop, target, or a forced close at midnight. There is no time-stop.
Up to three positions a day, opened as one wave. Once they are used up the day is over, which
bounds the worst possible day to three times the stop.
THE FILTERS ARE MIRRORED, NOT COPIED
Two gates decide whether the day is traded, and both are the reverse of what F5 looks for:
Absorption the trigger candle must show heavy BUYING that barely moved price - its positive
delta divided by its range must exceed 4. Somebody is selling into that buying.
Day delta skipped when the day's cumulative delta is above +12000. That is a genuine buying
day, not an overshoot, and price does not come back down.
WHAT THE MEASUREMENT SAYS
Replayed on a 1000-tick MNQ chart, 15 December 2025 to 18 August 2026 - eight months, 176 trading
days, costs included at 2 ticks per trade, one contract:
119 trades, taken on 46 of the 176 days
win rate 52 %
profit factor 2.71
net +19 097 ticks = 109 ticks a trading day
worst drawdown 1 774 ticks
worst single day -606 ticks
eight of nine months profitable
the three best days account for 33 % of the profit, and the median trading day is +461 ticks
HOW IT COMPARES WITH F5
On the same eight months the short side traded on more days than the long side (46 against 28),
concentrated its profit less (33 % in the top three days against 47 %), and returned more per
trading day. That is the opposite of what an earlier measurement on five months of 500-tick data
suggested, and the longer sample is the one to trust.
READ THIS BEFORE YOU EXPECT THOSE NUMBERS
The settings were tuned on this data. On the sealed final third that the tuning never saw, the
strategy stayed profitable but returned 72 ticks a day instead of 127, with the win rate falling
from 56 % to 41 %. Expect roughly half.
There is also a warning in the shape of the result: the short edge was strongest in the first two
thirds of the sample and weakest in the last, which was a rising market. A short strategy depends
on the market being willing to fall.
ONE CONTRACT BY DEFAULT
The measured drawdown of 1 774 ticks is 887 dollars on one contract - under a 1000 dollar account
limit. Two contracts would put it at 1 774 dollars and the account would not survive the same
stretch.
WHAT YOU SEE ON THE CHART
The day's average price and its sigma bands, with the entry threshold marked. An arrow at the open
of the candle where the trade actually started, entry, stop and target lines labelled Entry, SL and
TP, and the reason it ended: TP, STOP, DAYSTOP or MIDNIGHT. The TUNE button opens a tuning window
with live statistics and one-click access to the five configurations from the backtest.
Installation guide
2. Copy OFS.F6ShortReversion.dll into your ATAS indicators folder.
3. Add "OFS F6 Short Reversion" to a 1000-TICK MNQ chart with cluster data.
4. Paste your license key into the License group.
5. Set Bar Time Offset: +2 in summer, +1 in winter. Every time rule depends on it.
6. Leave Autopilot OFF and let the history replay. Read the trades drawn on the chart, and open
TUNE to see the statistics.
7. Only once you have seen the result on your own data, and decided the drawdown fits your account,
connect an account and switch the autopilot on.
8. Optional: Telegram notifications (group 13 - Telegram). In Telegram open @BotFather, send /newbot and paste the token into Bot Token; write any message to your new bot, then click Get Chat ID Now (the Chat ID fills itself) and Send Test Message. Turn on Enable Telegram - real entry fills, exits with their result in ticks and execution errors then arrive on your phone. Everything is off by default and nothing is sent without a valid licence.
F6 can run on the same chart as F5 - they take opposite sides and each manages only its own
orders.
Changelog — version 0.6.6
The centred new-version card - product name, your version and the new one, a Close button - now also shows a short description of what each release brings. Identical in every OFS indicator.
0.6.5 - A FILL CAN NO LONGER BE CREDITED TO THE WRONG ENTRY, AND THE DAILY COUNTER FOLLOWS THE MARKET DAY.
Order comments were matched by prefix, and entry number 11's comment starts with entry number 1's - so from the tenth entry of an instance's life onwards, a stop or target fill could be booked against the wrong entry: wrong filled quantity, protective orders re-placed for the wrong amount at a stale price. Comments are written exactly, so they are now compared exactly. Nothing changes in how orders are sent.
The daily realized counter (the one the daily loss limit reads) used to reset at the computer's local midnight while the midnight safety close measures market time. On a machine in a different time zone the two could disagree by hours. Both now step by the same market clock.
0.6.4 - A NEWLY PASTED LICENSE KEY TAKES EFFECT IMMEDIATELY.
Entering a key used to wait for the next scheduled license check; the check now runs the moment the key changes.
0.6.3 - NEW UPDATE CARD, CENTRED AND WITH DIRECT LINKS.
The card that appears when a newer version is available has been redesigned, identically for every OFS indicator. It now sits exactly in the centre of the chart, shows each product with its current and new version, recommends updating, and carries clearly visible buttons: an Info button that opens the product's page on orderflowsmith.com directly, a Close button at the bottom and the familiar cross in the corner. When several OFS indicators on one chart have updates at once, they all share the single card - one line and one Info button each.
0.6.2 - ALERT SOUND IS A LIST WITH A PLAY BUTTON.
The alert sound used to be a text field with "alert1" typed into it. If you did not already know what the sound files are called, there was no way to find out from inside the indicator, and a typo produced silence rather than an error. It is now the platform's own sound picker: a drop-down of the installed sounds with a small play button beside it, so you can hear a sound before you settle on it. It is the same control ATAS uses in its own indicators.
Because the list comes from the platform rather than from us, it also shows sounds you added yourself - drop a .wav into the platform's Sounds folder and it appears.
A sound you had chosen before is kept.
Two more things were wrong here. Alerts had no sound setting at all, and the name this indicator handed to the platform was not the name of any sound file - so the alert appeared in the alerts window but never actually made a sound. Both are fixed.
0.6.0 - HOW MANY CONTRACTS YOU TRADE IS YOUR DECISION AGAIN.
Switching a setup used to silently reset Trade Quantity back to 1. If you had set 3 contracts and then tried another setup, your size was gone without a word - and the previous release even announced that reset as a feature. It does not happen any more: a setup changes THRESHOLDS, never your position size. The upper limit on the setting is gone too.
What has not changed: the default is still 1 contract. Position sizing is a risk decision and it stays entirely yours.
0.5.9 - MAX ONE CONTRACT OF EXPOSURE ON EVERY SETUP. PROFIT held up to three contracts and WINRATE two; both now take a single entry: PROFIT +1,574 t / PF 3.08 / MDD 252, WINRATE 75 % win / +1,499 t / PF 7.05 (4 sealed trades - a thin sample, said openly). Entries Per Day and Contracts Per Entry remain ordinary settings for a conscious increase; the defaults never scale by themselves.
0.5.8 - EVERY STRATEGY BUTTON NOW SETS CONTRACTS = 1. The setups configure the engine AND the position size: the sealed figures are per one contract, so a button that left an old contract count on the chart would show numbers that matched only by accident. Raise Contracts afterwards if you consciously want more size.
0.5.7 - ONE POSITION BY DEFAULT. The shipped setup now opens a SINGLE entry with a single contract (sealed: 6 trades / 50 % / +1,574 t / PF 3.08 / MDD 252 t on the 58 out-of-sample days). The measured three-entry setup (+4,158 t / PF 2.83 / MDD 756) lives on the PROFIT button - one click away, but a conscious one. The DEFAULT button in the header returns exactly the shipped single-entry setup.
Also: THE UPDATE CARD'S CLOSE BUTTON IS A REAL BUTTON - a framed, highlighted square instead of a faint cross, comfortably clickable.
0.5.6 - ONE VOCABULARY FOR THE TIME WINDOW across every OFS strategy: Session Window, Session From (HH:MM), Session Until (HH:MM) - the same three names in F2, F5 and F6, all in group 05, entered as HH:MM. The session end also moved out of Filters into group 05 where it belongs, and it now honours minutes, not just whole hours. Saved charts keep their values.
0.5.5 - THREE ALIGNMENTS WITH THE OTHER OFS STRATEGIES:
USE SESSION WINDOW - the time window now has a proper OFF switch (group 05). ON by default - the sealed setups were measured with it. Switch it off and the strategy trades the whole day, the ribbon stretches to 24 h, and the drawn backtest follows the switch exactly like the live account.
TELEGRAM (group 13) - the same block F2 has: real entry fills with live SL/TP prices, exits with the result in ticks, execution errors with a 10-minute anti-spam. Everything off by default; setup takes four clicks with @BotFather.
SETTINGS ORDER - panel settings moved to their own group 11 - Panel (they sat inside Display), alerts are group 12, Telegram 13 - the same template every OFS indicator follows. The session ribbon now shows the trading window only.
0.5.4 - The session ribbon got a tidy layout: the hour scale (0/6/12/18/24) has its own line and the sentence with the session times and the current time sits on the next one - they no longer fight for space, every number shows.
0.5.3 - THE SESSION RIBBON READS ITSELF. Hour numbers 0-18 under the strip, the session times and the current time written out next to them (session 15:30-18:00 - now 14:32), and the now-marker is an arrow instead of an anonymous line. One blue for the window; the brighter top half marks the measured core hours.
0.5.2 - A SESSION RIBBON ON THE PANEL. A slim 24-hour strip under the title shows the trading window from YOUR settings (one source of truth - change the session and the ribbon follows), the measured core hours in a lighter shade, and a now-marker whose colour is the state of the strategy: green in position, blue hunting for an entry, amber standing aside with the gate named, grey quiet outside the window with a countdown to the open, red blocked. The TRADING line says the same in words. On by default; Show Session Ribbon turns it off.
0.5.1 - THE STRATEGY BUTTONS SWITCH PROPERLY. The active-setup check compared only the fields the old presets differed in, so with the new setups several buttons could light at once. The comparison now covers every field a setup controls, it lives in one shared place, and a release test walks every pair of setups to prove they are mutually exclusive.
0.5.0 - TUNED THE HONEST WAY, LIKE F2 AND F5: 25,920 combinations tuned on two thirds of the tape, candidates verified on 58 SEALED days the tuning never saw. The short side surprised:
SCALING IN WORKS ON SHORTS. Unlike the long side, later legs of a short wave EARN MORE per trade (141 -> 168 -> 196 ticks a leg) - so the new default keeps THREE entries, one contract each. The worst day stays bounded by three stops.
THE NEW DEFAULT (PROFIT) DROPS THE GATES: plain sigma 2.5 threshold, stop cap 250, session till 18:00, no absorption, ADX or day-delta filter - the sealed days preferred the clean threshold: +4,158 ticks, profit factor 2.83, drawdown 756 ticks. The previous gated setup made a similar profit (+4,426 t) but with PF 2.50 and 2.3x the drawdown (1,774 t) - the new default wins on risk.
WINRATE keeps the gates: absorption 4, stand aside when the day is bought by more than +8,000 delta, three-quarter target, two entries - 75 % win rate, +3,012 t, PF 7.03 on the sealed days (only 8 trades there - a thin sample, said openly).
STRATEGY row in TUNE like F2/F5, DEFAULT in the header, only the active setup lights up. Release tests pin both setups to their exact measured results.
0.3.11 - ONLY THE ACTIVE SETUP LIGHTS UP in the TUNE window. The DEFAULT button in the header used to glow permanently; now it highlights only while the default setup (E) is really applied, exactly like the A-E strategy buttons - what is lit is what is running.
TUNE buttons also size themselves to their text instead of a fixed width, so no label can overrun its button.
0.3.10 - Internal: the panel footer is now built by the same shared code as every other OFS indicator. What it shows does not change.
0.3.9 - EVERY TRADE IS NOW MARKED THE SAME WAY IN EVERY OFS STRATEGY. F2 FiboFade, F5 Long Reversion and F6 Short Reversion draw a trade identically, so if you run two of them on one chart you read both the same way:
the stop line says SL, in the middle
the target line says TP, in the middle
the entry line carries no text at all - only how long the trade lasted, for example 47 min
There is no Entry on the entry line any more: which line is the entry is already said by the arrow and the color, and the word only got in the way. Nor does the exit reason stand in place of SL or TP, so the stop line always reads SL and the target line always reads TP.
Every label sits in the middle of its line and the line BREAKS around it, so nothing is drawn through the text and no label is left half-covered by another trade. Where a trade is too short on the chart to hold a label, the label is left out rather than squeezed in.
It can all be switched off under Show Trade Labels, and the duration on its own under Show Trade Duration.
0.3.8 - EVERY CLOSED TRADE NOW SHOWS HOW LONG IT LASTED, on its entry line, next to the word Entry - for example Entry - 47 min. The time is taken from the candle times on your chart, so it is right on any chart type, and it tells you at a glance whether the strategy is holding for minutes or for hours. Switch it off under Show Trade Duration.
The Entry, SL and TP labels are also easier to read. They were written straight over the chart, so candles showed through the gaps between the letters and the text looked struck through. The line now breaks around each label, exactly as it does in F2 FiboFade.
0.3.7 - The free trial is 7 days, the same as every other OFS indicator. This product card still carried 14 days, so re-uploading the package kept putting the longer trial back.
0.3.6 - A NEW VERSION NOTICE. Once a day the indicator asks orderflowsmith.com whether a newer version has been released and, if so, shows a small card on the chart with what the new version brings you.
It sends nothing at all - no license key, no product name, not a single parameter - so the request cannot tell anyone what you use, and one request covers every OFS indicator on the chart. The answer is signed with the same key as your license, so nothing else can put text on your chart. It never blocks anything and it stays silent when there is no connection.
There is one card per chart, not one per indicator, and it lists every product that has an update. Close it and it will not come back until the next release. It can be switched off in Settings under 01 License.
0.3.5 - LICENCE GATE ON NOTIFICATIONS. Platform alerts and Telegram messages now require a valid key, exactly like the drawing on the chart does. The gate sits in the place where a notification is actually sent, not at the caller, so it cannot be bypassed by a new call site. Any queued notifications are discarded while the key is missing, so they do not pile up and all fire at once the moment a key is entered.
0.3.4 - The red DRAWING ERROR line is gone from the panel. It said nothing a trader could act on and only cluttered the window; a drawing failure now goes to the log instead, and only when the message changes. The license notice and the reason trading is blocked stay on the panel.
Three metrics were added to the TUNE statistics so every OFS strategy reports the same set: Worst Trade, Long and Short (trade count with its own win rate).
0.3.3 - Clicking a button in the TUNE window now recalculates immediately. It used to wait 250 ms after the last change, and that wait restarted with every further change - so with the thirty settings a strategy preset writes, clicking through A, B, C, D and E faster than that never recalculated at all, and the figures only caught up once you stopped clicking. A click is a single deliberate action, so the whole preset is now applied as one change and recalculated once, immediately, before the window is redrawn. The same applies to every other row in the window.
0.3.2 - The TUNE window and the control panel are now exactly the same width, and both are narrower. Each window works out the width it would need on its own and both are then given the larger of the two, so switching between them no longer moves anything. Nothing is fixed in pixels, so the size follows the font.
Metrics are spelled out instead of abbreviated: Win Rate, Profit Factor, Gross Profit, Gross Loss, Max Drawdown, Trading Costs, Payoff Ratio, Expectancy, Trading Days and Total Trades.
Three new figures about losing runs. Worst Day tells you about a single bad day, but a drawdown is built by bad days following one another. Worst Run is the largest combined loss of consecutive losing days, Losing Days is how long that run lasted, and Losing Streak is how many trades in a row lost, counted in the order they closed.
The word leg is gone from the panel and the log - everything says entry, matching the Entries Per Day setting. The EXEC line is shorter: 2 entries - 2/2 protected - day +120 t.
0.3.1 - A filter-by-filter audit found six gates still written for the long side. Five of them watched the day LOW and falling bars, which is the exact opposite of what a short entry needs: max drop from the day open, minimum distance above the day low, maximum bounce off the day low, bars since the day low, and consecutive down bars. They are now max RISE from the open, distance below the day HIGH, pullback off the day high, bars since the day high and consecutive UP bars. The sixth was the optional stop placement: it took the highest of the recent LOWS instead of the recent highs, which would have put a short stop inside the range rather than above it.
All six ship OFF by default, so measured results are unchanged - but any of them switched on would have filtered backwards. Tests now feed the engine a candle near the day high and check each gate rejects and passes where it should.
0.3.0 - The side is now fixed in code, not a setting. F6 trades SHORT only. The Trade Side setting has been removed because the absorption filter looks for buying being absorbed and the day-delta gate stands aside on buying days - both built for the short side - so switching it would have left the filters asking for the wrong thing. The engine refuses an opposite trade even if the field is set wrong, and a release test proves it.
0.2.4 - The panel keeps a FIXED width instead of resizing whenever a number changes, and uses the same width sample as F5 so the two look identical side by side. The distance to the next entry now measures upward, toward the short trigger, rather than reusing the long-side formula which understated it.
0.2.3 - Nothing is drawn outside the panel any more: its width is measured from every row that is actually drawn rather than from a fixed sample. The TARGET line now works out the distance to the next entry - entry in 396 t, or entry ready.
0.2.2 - Terminology corrected: spread means bid versus ask, so the settings that measure the standard deviation around the day average are now Min Sigma, Max Sigma and Show Sigma Bands. The panel line that read AVERAGE now reads TARGET and shows where the target sits, how far price is from it, and where the next entry would trigger.
0.2.1 - Both panels are considerably narrower: the control panel is compact again and the TUNE window uses the standard futures abbreviations (WR, PF, MDD, GP, GL, Exp.) with tighter spacing. Panel names now read OFS F6 SHORT REVERSION everywhere.
0.1.3 - The line of backtest figures under the strategy buttons is gone. Those numbers were fixed in the code and never recomputed, so next to the live statistics they read as if they applied to your chart. Every number in the TUNE window is now computed from your loaded history.
0.1.2 - The TUNE window now says where each set of numbers comes from: the line under the strategy buttons is prefixed BACKTEST 8M 1000t, the block below is headed THIS CHART and is computed live from the loaded history over the selected window.
0.1.1 - LICENCE GATE. Without a valid key the indicator draws NOTHING - no trades, no bands, and the TUNE window will not open. The Legs setting is now Entries Per Day, and the TUNE window keeps a constant width when buttons are switched.
0.1.0 - First build. The short side of the reversion trade: entry 2.5 spreads ABOVE the day's volume weighted average price on a closed candle, filled at the next candle's open, target the average price, stop capped at 200 ticks, three positions a day as a single wave. The absorption and day-delta filters are mirrored from F5 rather than copied - F6 looks for buying being absorbed and stands aside on genuine buying days. Measured on eight months of 1000-tick MNQ data (15 December 2025 to 18 August 2026, 176 trading days): 119 trades on 46 days, 52 % win rate, profit factor 2.71, +19 097 ticks, worst drawdown 1 774 ticks, eight of nine months profitable. On the sealed final third it kept about half of that. Five configurations from the backtest are one click away in the TUNE window and release tests fail if any of them stops producing the number on its button.
Frequently asked questions
How is F6 different from F5?
Which one performs better?
When exactly does a trade open?
Why one contract and not two?
What is the absorption filter measuring?
It sits out for days at a time. Is something wrong?
Is there a time-stop?
Reviews
No reviews yet — be the first verified buyer to share your experience.