This Week's Crypto Operating Framework (2026-09-27) — Regime: 超买 / 过热 (Overbought / Overheated)

TL;DR: This week's directional call is wait: BTC at $84,488 with RSI 71.3 and +8.0% over 30 days is an overbought/overheated regime that poorly compensates fresh full-size longs. Prefer reduced size, 1h pivot-fade shorts (LINK backtest Sharpe 9.09) and 15m VWAP reversion (BTC Sharpe 7.13) over 4h trend continuation.

---

This week's market environment

The snapshot below is the entire factual base for this framework. Everything that follows is derived from these numbers plus the backtest table, not from narrative.

Asset	Price	24h	7d	30d	RSI	Regime read
BTC/USDT	$84,488	+0.1%	+4.1%	+8.0%	71.3	Overbought / overheated
ETH/USDT	$2,701	+0.2%	+2.1%	n/a in snapshot	67.5	Constructive, not yet overbought

Three structural observations from the multi-timeframe stack.

First, the 30-day and 7-day returns on BTC are both positive (+8.0% and +4.1%), and the 7-day figure is roughly half the 30-day figure. That means approximately half of the month's advance was earned in the most recent week. The one-month trend is not just intact; it accelerated into the snapshot. Acceleration into an overbought RSI print is the classic definition of an overheated tape — it is a statement about the pace of the move, not about its direction.

Second, the 24-hour change of +0.1% on BTC and +0.2% on ETH is effectively flat. So the same instrument that gained 4.1% over seven days gained essentially nothing in the last day. This is a stall at the top of a short-term expansion, not a breakdown. Stalls inside an overbought regime are ambiguous: they can resolve into either a continuation leg or a mean-reversion leg, and the honest position is that the data does not yet discriminate between the two.

Third, ETH is lagging BTC on the only comparable window. BTC is +4.1% over 7 days while ETH is +2.1% over the same period, and ETH's RSI of 67.5 sits below the 70 threshold while BTC's 71.3 sits above it. A market where the largest risk asset is overbought but the second-largest is not is a market with narrow leadership. Narrow leadership does not invalidate an uptrend, but it does mean that breadth is not confirming the advance, and breadth non-confirmation is a reason to reduce conviction and size rather than a reason to press.

Regime classification. Combining RSI and trend structure, the regime label for the week is overbought / overheated. This is a trend-positive but momentum-stretched state. The practical implication for an operating framework is narrow: this is a regime that historically punishes late entries into continuation trades and rewards tactical, shorter-horizon, mean-reversion and fade structures — if they are run with tight risk. It is not, on the available evidence, a regime that justifies abandoning risk management.

What this environment is not. There is no evidence in this snapshot of a trend break on BTC: 30-day return is positive, 7-day return is positive, and price is not described as being below any structural level. The framework below therefore does not treat this as a short-the-market regime. Equally, there is no evidence of a fresh breakout either: the 24-hour change is flat and RSI is already stretched. The environment is best described as coiled, with the resolution — up or down — not yet determined by the data we hold.

---

Key indicator read

RSI: overbought on BTC, not yet on ETH

BTC's RSI at 71.3 is above the conventional 70 overbought line. ETH's 67.5 is below it. Two readings follow directly:

1. BTC's momentum is extended. Readings above 70 indicate that recent gains have been realized in a compressed window relative to the prior lookback. This is not a sell signal; it is a signal that the marginal buyer at these levels is paying for a move that has already happened.
2. ETH is not confirming. In a healthy, broad advance you would generally want the second-largest asset to be at or near the same overbought condition as the leader. It is not. That gap is a breadth warning, and it argues against treating the current BTC strength as the beginning of a broad, high-conviction expansion.

The specific RSI signals worth monitoring this week:

• BTC RSI holding above 70 versus rolling back under it. A sustained print above 70 extends the overheated regime; a decisive roll back below 70 while the 30-day return remains positive would be the first evidence of momentum decay without trend damage.
• ETH RSI crossing 70. This is the single cleanest positive breadth confirmation available in the current data set. It would move the regime from "narrow overbought" toward "broad overbought," which historically changes the strategy-fit mix.
• RSI divergence between BTC and ETH widening further. If BTC pushes higher on RSI while ETH fails to make progress, the leadership narrows further and the case for reduced size strengthens.

Trend structure

The trend structure is unambiguously positive on the windows provided: +4.1% over 7 days and +8.0% over 30 days on BTC, +2.1% over 7 days on ETH. Nothing in the snapshot describes a lower-high or lower-low sequence, and the 30-day window contains no drawdown information at all.

The correct interpretation of "overbought within an uptrend" is that trend-following structures remain valid but late. Position sizing, not directional bias, is the variable that should change.

Signals to watch this week

• The sign of the 7-day return. If it flips negative while the 30-day return stays positive, the market has begun to give back the recent acceleration leg — a distribution tell.
• Whether the 24-hour change stays pinned near zero while RSI stays above 70. A long stall at overbought levels is a compression setup, which is directly relevant to the squeeze-breakout strategies in the backtest book.
• Directional expansion in trend strength on 4h charts, which is what the ADX-based strategies require to function. In a stalling, mean-reverting tape, ADX trend structures underperform relative to reversion structures.
• Volatility-regime transition. Squeeze strategies only pay when a range contracts and then expands; the current flat 24-hour print inside an overbought regime is a contraction condition, not an expansion condition.

---

Direction: bullish / bearish / wait

Call: WAIT. Bias is neutral-to-constructive with reduced size, not a directional bet.

The bullish case is short and it is real. BTC is up on every provided window: +0.1% (24h), +4.1% (7d), +8.0% (30d). Trend structure is positive. There is no trend-break evidence. A market that has delivered +8.0% over a month while remaining above its own momentum thresholds is a market that has not yet given anyone a reason to be short.

The bearish case is also real, but it is a timing case, not a trend case. RSI 71.3 means the move is extended. The 24-hour change of +0.1% means the extension has stopped progressing. ETH's failure to reach overbought while trading +2.1% over 7 days against BTC's +4.1% means leadership is narrow. Stretched momentum plus a stall plus narrow breadth is a poor combination for a new full-size long initiated today.

Neither case is strong enough to justify a directional commitment at full risk. That is the definition of a wait regime. Concretely, in this framework "wait" means:

• No new full-size long exposure initiated into RSI > 70.
• No outright short exposure initiated against a positive 30-day trend.
• Continued operation of existing tactical structures at reduced size, because flat exposure is also a decision that carries opportunity cost.
• Predefined triggers that would change the stance to constructive: broad-based momentum confirmation (ETH joining BTC above the RSI threshold) combined with directional expansion rather than a flat 24-hour print. And triggers that would change the stance to defensive: the 7-day return turning negative while the 30-day return remains positive, or a sustained break of the trend structure.

No specific price levels are used in this framework. The reason is methodological: the backtest statistics we hold describe strategy behaviour, not price projections, and mixing the two produces a false sense of precision.

---

Strategy fit: which strategy types historically hold up in this regime

The live book contains 69 tested strategies, 0 currently running, 10 with Sharpe ≥ 1.5, and a top Sharpe of 29.2. A 10-of-69 pass rate at the 1.5 Sharpe threshold is roughly a 14.5% hit rate. That is a normal result for a broad strategy sweep and it should be read as a filtering outcome, not as evidence that ten independent edges exist.

Current top strategies (real backtest output):

Rank	Strategy	Pair	Style / TF	Sharpe	Backtest annual	Max DD	Trades	Win rate
#69	Bollinger Band squeeze breakout (#137)	DOGE/USDT	swing, 4h	29.19	+22.3%	1.1%	3	0.0%
#66	Classic pivot breakout (#134)	LINK/USDT	short, 1h	9.09	+48.9%	2.0%	29	0.0%
#61	TTM Squeeze breakout (#130)	BTC/USDT	swing, 4h	8.48	+13.2%	2.3%	23	0.0%
#63	Classic pivot breakout (#132)	SUI/USDT	short, 1h	8.44	+60.0%	1.7%	11	0.0%
#11	VWAP reversion (ultra)	BTC/USDT	ultra, 15m	7.13	+93.8%	1.1%	41	0.0%
#64	ADX dual-line trend (#117)	SOL/USDT	swing, 4h	5.49	+11.1%	2.8%	14	0.0%

Reading the table honestly

Three caveats must be stated before any of these numbers are used to allocate risk.

Caveat one: the win-rate column is unusable. Every strategy reports 0.0% win rate while simultaneously reporting positive Sharpe and positive annualized return. A strategy cannot have a 0.0% win rate and a positive Sharpe over 29 trades. The correct interpretation is that the win-rate field is unpopulated or computed on a subset, and it should be excluded from all sizing and selection logic this week. Do not use it as evidence either for or against any strategy.

Caveat two: small samples dominate the top of the ranking. Strategy #69 carries the headline Sharpe of 29.19 on 3 trades. Three trades is not a sample; it is an anecdote. A Sharpe of 29.19 on 3 trades is statistically meaningless and should be treated as a candidate requiring forward validation, not as the best strategy in the book. Contrast this with #11 (41 trades) and #66 (29 trades) — those are the two most statistically usable entries in the top six.

Caveat three: 69 tests means multiple-comparison inflation. When you test 69 strategy/pair/timeframe combinations and select the top ones by Sharpe, the top of the ranking is partly selection noise. The 10 strategies clearing Sharpe ≥ 1.5 are a reasonable candidate pool; the exact ordering within that pool is not reliable.

Fit by regime

Overbought, stalling, narrow-breadth regimes have a characteristic strategy profile. My mapping of the current book to this regime:

Best fit — mean reversion and tactical fades. In an overheated tape where price stops progressing, short-horizon reversion structures are the natural beneficiaries. Two entries apply:

• #11 VWAP reversion (BTC/USDT, ultra, 15m) — Sharpe 7.13, backtest annual +93.8%, max DD 1.1%, 41 trades. This is the strongest fit in the book on a risk-adjusted and sample-size basis. It trades a 15-minute horizon, which is exactly the horizon at which an overbought stall expresses itself, and it has the largest trade count in the top six. The reported 1.1% max drawdown is the tightest in the table.
• #66 Classic pivot breakout (LINK/USDT, short, 1h) — Sharpe 9.09, backtest annual +48.9%, max DD 2.0%, 29 trades. Despite the "breakout" naming, this is a 1-hour short structure. Short-horizon structures that profit from downside follow-through are the direct beneficiaries of an overbought regime that resolves downward. 29 trades is a defensible sample.
• #63 Classic pivot breakout (SUI/USDT, short, 1h) — Sharpe 8.44, backtest annual +60.0%, max DD 1.7%, 11 trades. Same logic, smaller sample. Treat as a secondary allocation behind #66.

Conditional fit — squeeze breakouts. Squeeze strategies are regime-agnostic in principle but require a contraction-then-expansion sequence to trigger. The flat 24-hour prints in the current snapshot are a contraction condition, which makes these relevant as watchlist items:

• #61 TTM Squeeze breakout (BTC/USDT, swing, 4h) — Sharpe 8.48, backtest annual +13.2%, max DD 2.3%, 23 trades. The most directly applicable squeeze structure because it trades BTC itself, and it has a usable 23-trade sample. Its lower backtest annual (+13.2%) versus the higher-Sharpe entries reflects its lower trading frequency.
• #69 Bollinger Band squeeze breakout (DOGE/USDT, swing, 4h) — Sharpe 29.19, backtest annual +22.3%, max DD 1.1%, 3 trades. Highest headline Sharpe, weakest statistical foundation. If it trades at all this week, it should trade at the smallest size in the book pending forward confirmation.

Weakest fit — trend continuation on 4h. Trend structures need directional persistence and expanding trend strength. A tape that has gone flat on the 24-hour window while sitting above RSI 70 offers no confirmation of either:

• #64 ADX dual-line trend (SOL/USDT, swing, 4h) — Sharpe 5.49, backtest annual +11.1%, max DD 2.8%, 14 trades. Lowest Sharpe of the top six, highest reported max drawdown, and the strategy type most dependent on exactly the condition the current tape lacks: sustained directional expansion. It stays in the candidate pool but it is the least aligned with the current regime.

Portfolio construction implication. The natural shape of this week's book, if anything runs at all, is a mean-reversion core (#11), a tactical short-fade sleeve (#66, secondary #63), and a squeeze watchlist (#61, and #69 at token size), with the ADX trend structure (#64) deprioritized. The common thread is short horizon and explicit reversion/fade logic, which is what an overbought, stalling, narrow-breadth tape pays for.

---

Operating framework & risk control

1. Nothing is live today: 0 of 69 strategies are running. This is the single most important operational fact in the snapshot. All numbers above are backtest output. Any allocation decision this week begins with moving a small number of candidates from backtest to forward paper tracking, not with capital deployment at scale.

2. Size down for the regime. Overbought/overheated is a reduced-size regime. A practical convention: per-trade risk of 0.25–0.5% of equity in this regime versus 0.5–1.0% in a neutral regime. The scaling is not a prediction — it is compensation for the fact that entry timing is statistically worse when momentum is stretched.

3. Treat reported max drawdown as a floor, not a ceiling. The top six report max drawdowns between 1.1% and 2.8% in backtest. Backtest drawdown is bounded by the historical path tested; realized drawdown is not. Plan capital so that a drawdown two to three times the reported figure does not force a behavioral decision.

4. Use strategy-defined stops, not conviction-defined stops. Every structure in the book has an implicit invalidation level — the other side of the pivot for #66/#63, the failed squeeze for #61/#69, the VWAP band for #11. Stops belong there. Widening a stop because the regime is "still bullish" converts a bounded loss into an unbounded one.

5. Cap portfolio heat. With four to six structures potentially active across 15m, 1h and 4h horizons, correlation is the real risk. #61 and #69 are both breakout structures and can fire together; #66 and #63 are the same strategy type on different pairs and are effectively one bet expressed twice. Aggregate simultaneous risk should be capped well below the sum of individual position risks.

6. Do not use the win-rate column for anything. As established, 0.0% win rate across all six top strategies with positive Sharpe is a data-integrity artifact. Any position-sizing rule that keys off win rate is currently keying off a broken field.

7. Validate before deploying. The correct sequence this week is: paper-track the reduced candidate set (led by #11 and #66), confirm that the trade counts rise, then increase size proportionally to confirmed sample. #69 should not receive meaningful size until its trade count moves from 3 toward double digits.

8. Custody and execution. All of this runs through Quant Pro with funds held in the user's own OKX account. Strategy execution does not require transferring balances to a third party, which means the risk control framework above operates on capital the user already controls, and withdrawal/permission hygiene remains the user's responsibility.

9. Log every trigger decision. The wait call is a decision with conditions attached. Record which condition was met — ETH RSI crossing 70, the 7-day return flipping sign, directional expansion replacing the flat 24-hour print — and what action followed. A framework whose triggers are not logged cannot be evaluated.

10. Re-underwrite at the next snapshot. This framework is built on a single dated observation (2026-09-27). Every conclusion in it is contingent on that snapshot. The regime label, the strategy ranking and the wait call should all be re-derived the moment new multi-timeframe and RSI data arrive.

---

FAQ

Q1: Is the market bullish or bearish this week?

Neither, on the available evidence — the call is wait. BTC is +0.1% over 24h, +4.1% over 7d and +8.0% over 30d, so the trend structure is positive; but RSI is 71.3, which is overbought. Positive trend plus stretched momentum plus a flat last-24-hours print is a wait regime: no new full-size longs into the stretch, no shorts against an intact 30-day trend.

Q2: Why is RSI 71.3 a problem when BTC is still up 8.0% over 30 days?

It is not a problem with the trend — it is a problem with entry timing. An RSI above 70 means the recent advance was realized in a compressed window. The 30-day return tells you the direction the market has travelled; the RSI tells you how much of that travel has already been paid for. Being right on direction and late on entry is still a losing combination if