Module 2 / 2
Recording and analysing sessions - without lying to yourself
Memory lies. After a session you remember a "nearly perfect entry" and "I got out a bit too late". The recording shows 3 seconds of hesitation, a click 2 bars too late, and a stop-loss moved by 5 ticks. This lesson turns review from a story into evidence: how to record, what to tag, and how to extract...
- Author
- Zordon Intelligence Sp. z o.o.
- Updated
- 9 min read
- Structure sources
- 2
Memory lies. After a session you remember a “nearly perfect entry” and “I got out a bit too late”. The recording shows 3 seconds of hesitation, a click 2 bars too late, and a stop-loss moved by 5 ticks. This lesson turns review from a story into evidence: how to record, what to tag, and how to extract one change that adds 0.2R to expectancy.
Lesson goal
- Build a recording system: screen, audio of your thoughts, platform data — light on CPU.
- Define 3 types of review: a 5-minute EOD, a 30-minute deep dive, and a weekend one.
- Measure Execution Score, Hesitation Time and Deviation from the plan.
- Link the video to your journal (TraderSync/Edgewonk/Notion) so one click replays the minute of the decision.
- Implement a progress protocol: 1 change per week, max.
Why there is no feedback without video
| Without a recording | With a recording |
|---|---|
| “I entered according to plan” | You can see a 12-second freeze and a market entry instead of a limit |
| “The stop was technical” | The stop was moved by hand twice, MAE -1.8R instead of -1R |
| “The market was weird” | Spread 4x, news in the background, you are trading outside the killzone |
| 0 measurable mistakes | 3 mistakes with a timestamp to fix |
The rule: if you have no recording, review is fiction. The goal is not perfect video, only enough video to count your mistakes.
Recording stack – light and reliable
| Layer | Tool | Pro settings | CPU cost | Why |
|---|---|---|---|---|
| Screen capture | OBS Studio (free) | Canvas 1080p, 30fps, bitrate 4000, encoder NVENC/x264 fast, record only the trading monitors | 5-8% CPU with NVENC | Evidence of execution |
| Thought audio (optional but gold) | Microphone + OBS audio track 2 | Think out loud: “Price reached the POC, delta divergence, I am waiting for the H1 close” | 1% | Catches the FOMO narrative |
| Platform – auto screenshot | TraderSync/ATAS auto screenshot on fill | A chart + DOM screenshot on every fill | 0% | Links the trade to the journal |
| Time & Sales log | ATAS/Sierra log to file | Save T&S to CSV once per minute | <1% | Measuring absorption later |
| Time measurement | On-screen timer (e.g. TradingView bar timer) | Always visible | 0% | Hesitation Time |
OBS setup for trading – step by step:
1. Scene: Trading Only – add Display Capture for monitors 1 and 2 only (not the whole desktop)
2. Settings -> Output -> Recording: MKV, Encoder NVENC H.264 (if you have NVIDIA) or x264 veryfast
3. Bitrate: 4000 Kbps, FPS 30 – enough to read price, 60FPS is not needed
4. Audio: Track 1 – desktop audio (alerts), Track 2 – microphone (your thoughts)
5. Hotkeys: F9 Start Recording, F10 Stop, F11 Save Replay 60s (buffer)
6. File name: %CCYY-%MM-%DD_%hh-%mm-%ss_KZ_% (automatic)
7. Storage: 1h ~1.2GB with these settings. 20h a month = 24GB – a 500GB drive is enough.
8. Startup: OBS autostart at system start, always recording in the background.
Backup and naming:
Folder: /TradingRecordings/2025/11/2025-11-20_NY_AM/
Files: 2025-11-20_1530-1700_ES_M5_ORB.mkv
A txt note with the same name: 2025-11-20_1530-1700_notes.txt with Mistake Tags
Types of review
1. EOD – 5 minutes (mandatory)
[ ] Does the recording exist? If not – tomorrow you do not trade until you fix OBS.
[ ] How many trades? How many A-Quality (100% according to plan)?
[ ] One worst trade – rewind to the minute of the decision, write down the Hesitation Time.
[ ] Daily Tilt Score from the journal + does the video confirm it?
[ ] Tomorrow: 1 reminder rule (e.g. "Do not trade outside the A-List")
Time: 5 min, not 30. 30 min is already a deep dive.
2. Deep Dive – 30 minutes – twice a week
You pick 3 trades: the best A-Quality, the worst C, and one BE.
| Timestamp | What you watch | What you measure | Formula |
|---|---|---|---|
| T-60s to entry | Did the setup appear according to the playbook? Did you wait for the close? | Setup Validity 0/1 | 1 if yes, 0 if no |
| T-0 entry | Order type, slippage, time from signal to click | Hesitation Time, Slippage_R | Hesitation = Time_click - Time_signal (s) |
| T+ life of the position | Did you move the stop? Did you add? | Deviation Count | Number of deviations from the plan |
| Exit | Was the exit according to plan or emotional | Exit Quality | 1=according to plan, 0=panic/FOMO |
| Audio | What did you say? “Because maybe…” or “The plan says…” | Narrative Tag | FOMO, Fear, Patience, Mechanical |
Execution Score per trade:
Execution_Score = (Setup_Validity + (Hesitation<5s?1:0) + (Deviation==0?1:0) + Exit_Quality)/4 *100%
Target: >85% avg. If <70% – the problem is not the strategy, only the execution.
3. Weekend – 60 minutes
You watch 2 ten-minute clips with the biggest intraday drawdown. Not all trades, only the 10 minutes where you lost 2R.
Questions:
- What triggered the first loss? (setup or mistake)
- How many minutes later was the second loss? If <15 min – revenge pattern
- What does the audio say? Silence vs explaining things away
- Were spread and latency fine?
- Decision: 1 physical block for the next week (e.g. auto-lock trading after -2R, tape over the key)
Measurable indicators from the video
| Metric | Definition | Target | Tool |
|---|---|---|---|
| Hesitation Time avg | Average time from signal → click | <5s for A-Quality | Timer + OBS |
| Slippage_R avg | (Fill - Intended)/Risk | <0.15R | Journal + T&S |
| Deviation Rate | Trades with Deviation>0 / all | <15% | Video |
| Tilt Audio Events | Count of “f*ck / I have to make it back / one more” per day | 0 | Audio track |
| No-Trade Discipline | How many times you did not click despite the temptation (count from the video) | >2 per day | Video |
Example – an extract from a review:
Day: 2025-11-18, NY AM, 3 trades
Trade1: ORB long, Hesitation 2s, Slippage 0.1R, Execution 100%
Trade2: FOMO short with no setup, Hesitation 12s, Deviation 1 (no technical stop), Audio: "I'll try it", Execution 25%
Trade3: Revenge after Trade2, 8 min after the loss, Hesitation 1s (impulse), Execution 0%
Conclusion: Trigger = Trade2 FOMO, cost of the day = Trade2+3 = -2.1R, of which -1.1R is deviation.
Block: After 1 trade with Deviation>0, a 30 min break + a checkbox in the calculator "Is this an A-List setup? Yes/No"
Linking the journal and the video
In Edgewonk/TraderSync add a Video_Link field – a link to the local file or to Loom (if you upload a clip).
Notion template:
Page per day: Date | Tilt Score | P&L_R | Link to the video folder | 3 thumbnail screenshots | 1 Loom 60s summary
Rollup: weekly Execution_Score avg, Hesitation avg, Deviation Rate
Automation: Make.com – when a new file lands in /Recordings, create a Notion entry with the link.
Loom for C-quality trades: you record the screen and talk for 60s about what went wrong. 60s is enough. Do not record 10 minutes of speeches.
Storage and privacy
- 1 month of recordings is ~30-50GB. Keep 90 days locally, then archive only A and C trades (delete B).
- Do not upload to public YouTube – account data, P&L.
- Encrypt the drive (BitLocker / FileVault) – recordings are financial data.
- If you use a prop firm, check the rules – some firms forbid recording the DOM.
When the plan says DO NOT TRADE
- OBS is not recording or lags at >15% CPU – fix it, do not trade without evidence.
- You have no microphone and have not saved thought audio for 2 weeks – you have half the data; require audio on at least 3 days a week.
- You watch recordings 2h a day instead of 5+30 min – that is procrastination, not review. Cap review at 35 min a day.
- You watch only winning trades – bias. 1 worst C per day is mandatory.
- After 3 trades with a Deviation Rate >50% you do not take a break – the plan says a 30 min lock.
- You have >3h of recording a day – overtrading, not pro activity. Shorten the session to a 2-3h killzone.
The mistakes that cost the most
- Recording the whole desktop in 4K 60fps: 40% CPU, the platform lags, slippage grows. Record 1080p 30fps, trading monitors only, NVENC.
- No replay buffer hotkey: You want to show a friend 30s of absorption, but you have to search a 2h file. Set F11 = save the last 60s as a separate file immediately.
- Review without a measure: You watch and say “well, weak”. Zero numbers. Measure Execution Score, Hesitation, Deviation – otherwise there is no progress.
- Deleting recordings after a week: After 30 days you want to compute statistics and you have 3 files. Keep 90 days, not a week.
- Recording without thought audio: You lose 50% of the information about FOMO. Thought audio is the cheapest tilt indicator – turn on the microphone, even if talking to yourself feels silly at first.
Measurable exercise
A recording system – 14 days.
- Day 1: Install OBS, configure it as instructed, test 10 min of recording, check CPU and file size. Set the F9/F10/F11 hotkeys.
- Days 2-7: Record every killzone session (max 3h). After the session a 5-min EOD + mark the Hesitation Time for every trade.
- Days 8-14: Add thought audio. Twice this week do a 30 min Deep Dive (3 trades). Compute Execution Score avg and Deviation Rate.
- End of day 14: A 1-page report:
- Avg Hesitation Time
- Execution Score avg
- 1 main Deviation trigger (e.g. FOMO after a loss, entering before the close)
- 1 physical block for the next week
- Link 5 recordings to the journal (a link in the Video_Link field). Check whether replaying a trade takes <30s.
Check yourself
- Why does recording the whole desktop in 4K 60fps make execution worse, and what OBS settings are recommended for a trader?
- Define Execution Score and compute it for a trade: Setup Validity 1, Hesitation 12s, Deviation 1, exit according to plan 1.
- You have 3 trades: all C-quality, Deviation Rate 100%, Tilt Audio Events 4. What does your plan say and what block will you implement?
- How many GB will a month of recordings take at 2h a day x 20 days at 1.2GB/h, and how long do you keep the files?
- Why is thought audio more valuable than the chart image alone for detecting FOMO?
Frequently asked questions
Do I have to record audio? It is not mandatory, but it is a game changer for tilt detection. Start with 3 days a week, and you will see how many times you say “I have to make it back” – that is data, not an opinion.
OBS loads my machine and ATAS lags. Switch the encoder to NVENC (GPU) instead of x264 (CPU), lower the canvas to 1080p, record only 2 monitors, not 3. If it still lags – buy a used mini PC as a dedicated recorder via a capture card.
Can I record the screen with my phone? No – the quality is poor, timestamps and T&S are not visible. OBS/screen capture only.
How do I quickly find the entry moment in a 2h recording? Use the F11 replay buffer at the moment of the click, or write the session minute in the journal (e.g. Session Minute 32:15). Then you jump in the player instead of watching the whole thing.
What if I do not want to listen to myself? That is exactly what it is for – discomfort = data. If you cannot stand listening to your own explanations, that is a signal that you are explaining instead of playing mechanically.
A recording turns “it seems to me” into “you can see that”. Without a recording you have no feedback, you have a narrative. A professional measures execution like P&L.
Advertisement
Key takeaways
- Memory lies. After a session you remember a "nearly perfect entry" and "I got out a bit too late". The recording shows 3 seconds of hesitation, a click 2 bars too late, and a stop-loss moved by 5 ticks. This lesson turns review from a story into evidence: how to record, what to tag, and how to extract...
Structure sources
The material is based on source documents. Variable values — fees, limits, regulatory figures — are worth checking at the source before making a decision.
- TradingView — platform documentation
- Bookmap / ATAS / Sierra Chart — docs
Disclaimer
- Educational material. Not investment advice or a recommendation. Trading involves the risk of losing capital.
- The materials are educational. They are not investment advice or a recommendation within the meaning of financial instruments trading regulations. Numbers in examples are illustrative unless a measurement date and source are given.
Who writes this and how we check it
The materials are written and verified by Redakcja Kurs Tradera — the rules are described on about the site: every number has a source, and we show gaps openly. Report a factual error via contact — corrections are described in the content.
Progress is stored only in this browser (localStorage) — there is no account and no copy on the server. Clearing site data will remove it. Manage your data