The last three candles look tired. That sentence has ended more practice accounts than a trend that never turned. On a finished NSE chart, exhaustion is obvious: the spike, the stall, the snap back. On the live bar — or on a replay that has not revealed tomorrow — "tired" is a feeling competing with "one more push."
You need a tree you can freeze before the next candle, not a debate you reopen after it.
Two bets, not one mood
Continuation is a bet that the directional pressure that produced the last impulse is still in charge. Invalidation is a loss of that structure — pause break, failed resume, time without follow-through.
Exhaustion is a bet that the impulse is spent at a location you already respected. Invalidation is acceptance beyond the extreme you thought was spent.
They are not two readings of the same candle. They are opposite hypotheses. If you can hold both at once, you do not have a setup. You have a comment.
This is the same fork as trend following versus mean reversion, taken down to a single bar. Style comparison happens in weekly blocks. This tree happens at the pause.
The Exhaustion/Continuation Tree (ECT)
Answer in order. No skipping to the leaf you like.
Gate 0 — Location allowed?
If price is not at a pre-marked extreme, opening-range edge, or written stretch zone, exhaustion is auto-off. Continuation may still be on, but only if impulse + pause rules from your momentum practice are already green.
No location, no exhaustion leaf. Mid-range "it looks extended" is not a level.
Gate 1 — Has a pause formed, or is this still the impulse?
If the bar in front of you is the expansion, you are late for continuation and early for exhaustion. Tag still_impulse and wait. Entering here is chasing dressed as a theory.
Gate 2 — What would continuation require on the next candle?
Write one trigger: resume through pause high (long), or resume through pause low (short). Write one kill: pause structure breaks the other way, or N bars with no resume.
Gate 3 — What would exhaustion require on the next candle?
Write one trigger: failed new extreme, reversal close back inside the prior range, or a rejection rule you can measure. Write one kill: a close beyond the extreme you faded.
Gate 4 — Commit a leaf before you click forward
Allowed outputs:
C-enter— continuation trigger defined; take it if the next bar prints itE-enter— exhaustion trigger defined; take it if the next bar prints itwait-one— one bar only, then re-run the tree; not an open-ended sitskip— neither leaf is legal
You do not get maybe both. You do not get I'll know when I see it. The next candle is the exam, not the teacher.
Why the freeze has to happen before the bar
Without a freeze, the next candle always "confirms" whatever you secretly wanted. A strong close becomes continuation you "called." A rejection wick becomes exhaustion you "saw coming." That is not a decision tree. That is a caption.
Hidden-date, candle-by-candle replay is the only honest laboratory for this. If you can see the 14:00 bar while judging 11:20, both leaves look smart.
Side-by-side: what you are actually choosing
| Question | Continuation leaf | Exhaustion leaf |
|---|---|---|
| Who is in charge? | The impulse still | The impulse is spent |
| Entry waits for | Resume of pause | Failure at the extreme |
| Stop lives | Beyond pause structure | Beyond the extreme faded |
| Common NSE trap | Midday resume that dies | Open spike that was not spent |
| Emotional tell | FOMO on the strong close | Hero urge at the high/low |
Hypothesis, not market law. Your logs decide which trap you walk into.
A 24-decision block
Twelve randomised NSE intraday sessions, liquid names, one timeframe. Two tree-commits per session maximum — otherwise you will force a leaf on every pause.
Rules of the block:
- Mark impulse using a written range-expansion rule, not "felt fast."
- Run ECT at the first pause only. Second pauses are a different sample.
- Speak the leaf out loud or write it then advance one bar.
- Score process, not outcome:
2if the leaf matched the written gates;0if you changed leaf after the bar printed.
Review questions (one sentence each):
- How often was Gate 0 the whole trade — exhaustion auto-off?
- How often did I pick
E-enterbecause the high felt important, not because a trigger was defined? - Did
C-entercluster in 09:15–10:30 whileE-enterclustered in lunch — and did I still take exhaustion in the open?
If Gate 0 kills most exhaustion attempts, you are learning location. That is a better result than a handful of pretty fades.
Failure modes the tree is meant to catch
Leaf-switching: continuation until the bar goes red, then "obvious exhaustion." Tag leaf_switch. This is the live tell that will hurt you.
Impulse trading: no pause, still picking a leaf. Tag still_impulse.
Location shopping: moving the "extreme" to wherever price stalled. Pair with the double-top census if your exhaustion leaf keeps appearing at unmarked highs.
Time denial: sitting through six bars of chop because neither leaf wants to die. Use the N-bar kill. Exhaustion that needs all afternoon is usually continuation you refused to admit.
Open special-pleading: first 15 minutes after 09:15 produce large candles that look exhausted because they are large. They are often just the open. Auto-skip exhaustion before a time you write down, or tag every open leaf open_ect and review that bucket alone.
What the tree will not do
It will not tell you which leaf is "right" for Indian markets. Nifty and Bank Nifty can trend through midday and mean-revert the same week. The tree is a commitment device.
It will not turn 24 decisions into a style identity. It can show whether you violate exhaustion rules more than continuation rules — setup-fit data, nothing more.
It will not survive blending with pullback versus breakout labels in the same session. One taxonomy per block.
Replay Trader's job here is mechanical: hidden dates, bar-by-bar reveal, a place to log the leaf before the next print. AI review after the session can help you spot leaf_switch tags you stopped seeing. It cannot certify that your tree has an edge.
Freeze the leaf, then roll the next NSE candle. Join the waitlist.
Related reading
- Momentum Trading Practice on Historical NSE Charts
- Trend Following vs Mean Reversion — Which to Practise First
- Reversal Trading: How to Practise Without Hero Narratives
- Double Tops and Bottoms Without Hindsight
Educational note
Educational practice content only. Not investment advice. No guaranteed returns. Practice ≠ live profitability.