Market analysis
The supplied market context describes a constructive but tactically uneven crypto tape. CoinGecko global data showed total crypto market capitalization near 2.87 trillion dollars and twenty-four-hour volume above 118 billion dollars, but the latest market-cap change was negative. That daily decline matters because several CoinRobot strategies operate on short intraday confirmation windows: a longer-term bull tape can still produce repeated micro pullbacks, stop-loss hits, and premature sell signals. Bitcoin’s ninety-day return was strongly positive, with a relatively contained maximum drawdown compared with many altcoin cycles and annualized volatility around the high-thirties. This combination is generally favorable for momentum and rotation systems, but not automatically favorable for high-frequency BUY/SELL systems, because intraday reversals can repeatedly interrupt otherwise constructive daily trends.
Binance data reinforced the longer-window strength. BTCUSDT, ETHUSDT, BNBUSDT and SOLUSDT all had positive ninety-day returns. ETH and SOL were particularly strong, with higher volatility than BTC and BNB. SOL’s return was large but came with the deepest maximum drawdown among the four supplied majors, which is exactly the environment where momentum rotation can work if it switches into strength early but can also suffer from whipsaws if entries lag or exits fire into pullbacks. BTC retained large daily quote volume and high trade counts, giving it relatively reliable execution context. ETH also showed strong volume and higher average funding, while SOL displayed high relative volatility and active participation. BNB was steadier but still experienced notable pullbacks. Across these assets, taker-buy share was close to balanced rather than persistently aggressive, which supports the report’s caution that the market was not a one-way intraday bid despite positive ninety-day performance.
CoinRobot’s market vector adds breadth context. Out of the supplied symbol set, the distribution showed more sideways and bull names than bear names. Median daily RSI was modestly above neutral, median daily NATR was near four percent, and median daily Bollinger width was above ten percent. The interpretation is a market with enough directional opportunity for momentum systems, but with enough volatility and dispersion to punish overly tight exits. Median trend strength was high, which favors trend-following formulas that require daily or intraday alignment. However, “sideways” was the largest single regime bucket, so range behavior remained common. That explains why many adaptive notes focused on avoiding sell-low churn, widening profit giveback thresholds, and adding confirmation to StopLoss rather than simply increasing risk.
The dated crypto event calendar was dense around the publication date, especially for SOL, ETH, ADA, HBAR, ZEC, ASTER and several infrastructure or DeFi names. The report does not infer causality from those events because the payload does not provide coin-level event-linked return studies. Still, event density raises the probability of idiosyncratic volatility around individual assets in strategy universes. A strategy rotating across or trading many symbols can benefit from event-driven dispersion, but only if formulas distinguish confirmed momentum from noisy spikes. This is directly relevant to strategy 935, which traded a broad trusted-asset set and generated many BUY/SELL events, and to strategy 933, whose selective filters still encountered elevated bad-buy and early-exit rates after the latest active revision.
A key tension is therefore visible across all supplied evidence: the market rewarded holding or rotating into winners over the longer window, but many strategies struggled with intraday exit precision. Strategy 934’s convert plans show that rotation can accumulate units when switches are sufficiently selective; strategy 912’s convert plan also had positive convert increase and better profit factor than its BUY/SELL cohort. Conversely, strategy 911’s historical rotation data showed weak convert increase and inactive current plans, while strategy 932’s validation failure indicates that formula complexity can become operationally dangerous if not scenario-tested. The market was favorable enough that inactive or overly restrictive systems risked opportunity cost, but volatile enough that loose entry expansion could compound churn.
The broad conclusion from market context is that this was not a bear-market de-risking problem. It was a quality-of-signal problem within a mostly constructive regime. Systems that preserved trend alignment, required MACD and EMA confirmation, and tolerated normal pullbacks tended to have better evidence than systems that exited into every short reversal. At the same time, the latest daily market-cap decline warns against indiscriminate relaxation. Adaptive changes should therefore be conservative: widen or confirm exits where evidence shows repeated early recovery, tighten entries where bad-buy rates are elevated, and preserve hard emergency exits for true breakdowns.
| Asset | 90d return | Max drawdown | Volatility | Taker buy share | Funding |
|---|
| BTCUSDT | 35.7765% | -6.9224% | 38.0591% | 49.7332% | 0.002021% |
| ETHUSDT | 58.2181% | -5.6161% | 52.0502% | 49.8734% | 0.005433% |
| BNBUSDT | 35.1081% | -7.5418% | 34.2811% | 49.2053% | 0.000923% |
| SOLUSDT | 47.12% | -12.6184% | 54.6582% | 50.7788% | 0.002359% |
Methodology
The report uses only supplied data and treats every data field according to its stated scope. Market analysis is based on the provided CoinGecko global snapshot, Bitcoin ninety-day series summary, Binance ninety-day asset summaries, the CoinRobot market vector and the supplied event calendar. No external news, macro data, or unsupplied price series are used. Because crypto news and macro-event feeds were disabled in the payload, the report does not infer macro drivers or news causality. Event-calendar items are treated as dated context only; they are not interpreted as causal explanations unless the payload also supplies supporting temporal price evidence, which it does not.
Strategy evaluation is performed in layers. First, each strategy’s declared formula is interpreted structurally: BUY conditions, SELL conditions and StopLoss conditions are separated, and indicators are translated into their decision roles. EMA alignment is read as trend direction, MACD histogram as momentum confirmation, RSI and RSI trend as participation or overextension filters, NATR and Bollinger terms as volatility-quality filters, and cooldown or last-price terms as anti-churn controls. Second, KPI windows are compared across seven, thirty and ninety days. The report explicitly distinguishes plan-level daily KPI context from active-revision trade-evaluation evidence. A thirty-day or ninety-day aggregate can include multiple plans and older behavior, while active-revision diagnostics describe the currently tested formula period.
Coverage metadata is treated as a data-quality constraint. When win-rate or profit-factor coverage is incomplete, unavailable, or absent, the report does not convert missing data into zeros. This is especially important for backtest plans, inactive plans, and long-running plans with no recent trades. Where the payload reports no stop-loss days, no convert-increase days, or no win-rate days, conclusions are qualified. Strategy 932 is a clear example: long-window KPI fields include very large historical figures, but the optimization was validation-blocked because scenario validation found the proposed buy formula always false. The operational validation result is therefore given high weight even when older aggregate KPIs appear favorable.
Execution-model analysis follows the rules supplied in the prompt. BUY/SELL is interpreted as an entry-and-exit model where realized profit, profit factor, win rate, early exits, bad buys and stop-loss behavior are primary quality measures. CONVERT is interpreted as a rotation model where unit accumulation, convert increase, switch frequency, target-versus-source post-switch returns and stalled-position evidence matter. A CONVERT entry is treated as a real executed entry even if the recorded formula evaluation was neutral. However, neutral CONVERT switch legs are not mislabelled as BUY or SELL formula triggers. Formula attribution is used only when evaluation success is available and result states are interpretable.
StopLoss is evaluated as a common risk-control layer across both execution models, but its effects are separated by stoploss_policy where the payload provides enabled versus disabled cohorts. A StopLoss exit is considered potentially problematic when the supplied evidence shows high early-exit rates, frequent exit-near-next-low behavior, or rapid recovery above the exit price within one, two or five minutes. Such evidence supports adding confirmation or widening thresholds, but the methodology requires preserving hard emergency branches. Therefore, the report distinguishes false-stop risk from catastrophic-loss protection.
Adaptive changes are not assumed successful merely because they were applied. They are classified according to evidence status: supported by active-revision evidence, supported by prior-revision evidence but still awaiting post-change confirmation, validation-blocked, or too inactive to judge. The hypotheses section converts these observations into measurable forward tests. Review triggers are framed in terms of event counts, window duration, and specific metrics such as bad-buy rate, early-exit rate, convert increase, target-minus-source returns and stop-loss recovery rates.
Observation windows: 7, 30 and 90 days. Formula changes remain attributable by strategy version and adaptive change ID.
BUY/SELL and CONVERT execution models
The supplied universe contains both pure CONVERT strategies, pure BUY/SELL strategies and shared strategies used by both execution models. This distinction is central because the same formula can have different practical meaning depending on execution. BUY/SELL seeks profitable entries followed by exit decisions that realize or protect fiat-denominated gains. CONVERT seeks to improve the number of units held across a set of named assets; fiat profit is useful, but convert increase, switch quality and lock-in risk are often more informative.
Strategy 934 is the cleanest dedicated CONVERT case. Its buy_sell execution_model_kpis are empty, while the convert cohort contains eight plans. Aggregate thirty-day data showed fifty-five trades, positive profit amount, a seventy-five percent win rate, profit factor above four, and positive convert increase. However, the ninety-day aggregate also contained a very large negative profit amount driven by backtest or historical plan effects, while realized profit and realized loss fields did not scale with the same magnitude. That mismatch requires caution. The revision history is more interpretable: the baseline had fourteen complete switches, all execution-confirmed, with positive convert increase, but target-versus-source returns were negative at short horizons. The newest active revision had no events, so the latest buy relaxation cannot yet be judged. This is classic CONVERT lock-in tension: strict formula gates may protect quality, but if no switch occurs, the held asset may become stale unless evidence later shows it remains competitive.
Strategy 935 is mostly BUY/SELL, with one small CONVERT plan. BUY/SELL produced six hundred-plus recent active-revision events, while CONVERT had only four events and two switches. The BUY/SELL cohort showed high trade volume and high win rates but profit factor below one across aggregate windows, meaning average or total losses outweighed wins despite frequent winning trades. This is a sign that exit sizing, stop behavior or loss distribution mattered more than headline win rate. The small CONVERT cohort showed positive convert increase and strong long-horizon post-switch evidence at one hundred eighty minutes, but its sample was too small to dominate conclusions. Because the strategy is shared across models, formula changes must be safe for both: tighter entries and more tolerant exits are safer than model-specific rewrites.
Strategy 932 also spans models, but the immediate execution-model conclusion is operational rather than performance-based. The optimization was validation-blocked because the proposed buy formula was always false under scenario validation. Regardless of historical convert increase or ninety-day turnover, a buy formula that cannot trigger in validation would break or freeze the strategy. This highlights why execution-model diagnostics must include formula validation and not only past KPIs. Inactive seven-day data and missing coverage also prevent confident short-term comparisons.
Strategy 912 shows a strong divergence between BUY/SELL and CONVERT. BUY/SELL had poor recent and thirty-day profit factors, low win rates and negative profit amounts, while the CONVERT plan had positive profit amount, positive convert increase and a much stronger thirty-day profit factor. Active execution diagnostics showed one convert switch and several StopLoss or initial-entry events, with true buy-formula initial entries having better short-horizon outcomes than false-state entries in the latest sample. This suggests the convert implementation of the swing logic was more robust than the BUY/SELL implementation, but sample size remains modest.
Strategy 933 is pure BUY/SELL in the supplied execution model. It has high trade count, high win rate and positive thirty-day aggregate profit, but recent active-revision evidence shows elevated bad-buy and early-exit rates. Since there is no CONVERT cohort, interpretation can focus on entry quality, sell timing and stop behavior without unit-rotation tradeoffs.
Strategy 911 is shared in design, but all listed plans were inactive at the report timestamp. Historical execution data showed both BUY/SELL and CONVERT events, weak profit factor, negative convert increase near zero, poor initial-entry outcomes for true buy states and high sell/stop-loss recovery rates. Because the plans were not actively trading, any formula change is a reactivation hypothesis rather than a live improvement claim. Across the universe, the best execution-model evidence favors selective CONVERT rotation in strategies 934 and 912, cautious BUY/SELL tightening in strategy 935, and operational validation discipline for strategy 932.
| Strategy | Execution model | Plans | 30d trades | 30d profit | Profit factor | Win rate | Unit increase |
|---|
| CONVERT - Relative Momentum Rotation | CONVERT | 8 | 55 | 196.6577 | 4.7361 | 75% | 3.6975% |
| BUY/SELL - Trusted Asset Recovery Profit Guard | BUY/SELL | 7 | 798 | 19686.6017 | 0.8339 | 68.4588% | — |
| BUY/SELL - Trusted Asset Recovery Profit Guard | CONVERT | 1 | 5 | 1.2681 | 13.7216 | 50% | 0.4219% |
| Universal - Flexible Momentum Quick Profit Guard | BUY/SELL | 6 | 0 | 0 | — | — | — |
| Universal - Flexible Momentum Quick Profit Guard | CONVERT | 8 | 31 | 52.9613 | 1.3706 | 66.6667% | 87.3735% |
| Universal - Swing Trend Guard | BUY/SELL | 2 | 210 | -5.1899 | 0.5191 | 24.2718% | — |
| Universal - Swing Trend Guard | CONVERT | 1 | 91 | 28.5079 | 2.437 | 47.0588% | 21.3793% |
| Universal - Selective Trend Momentum Profit Guard | BUY/SELL | 2 | 1750 | 52.1273 | 1.2824 | 74.0645% | — |
| Universal - Trend Pullback Guard | BUY/SELL | 3 | 0 | 0 | — | — | — |
| Universal - Trend Pullback Guard | CONVERT | 6 | 0 | 0 | — | — | — |
CONVERT decision-pattern outcomes
| Strategy | Pattern | Switches | Avg unit increase | Target beat source after 60m | Mean target − source return |
|---|
| BUY/SELL - Trusted Asset Recovery Profit Guard | sell+neutral | 2 | 5.230321% | 0% | -0.133112 pp |
| Universal - Swing Trend Guard | sell+neutral | 1 | 9.763491% | 0% | -0.087284 pp |
| Universal - Trend Pullback Guard | neutral+buy | 1 | 0.302572% | 100% | 1.321686 pp |
| Universal - Trend Pullback Guard | sell+neutral | 1 | -0.319725% | 0% | -0.84792 pp |
StopLoss analysis
StopLoss evidence was one of the strongest cross-strategy signals in the supplied data. The main pattern was not that stop-loss rules failed to protect; rather, many enabled stop exits appeared too reactive to intraday noise. The supplied diagnostics repeatedly reported early-exit rates, exits near the following sixty-minute low, and rapid recovery above the stop price within one, two or five minutes. Those are classic symptoms of wick-driven or pullback-driven stops in a volatile but generally constructive market. The appropriate response is not to remove StopLoss, because hard breakdown protection remains necessary, but to separate emergency exits from ordinary volatility confirmation.
Strategy 934 had a small but important prior stop-loss sample. Under the baseline revision, two enabled stop-loss exits were recorded, both early, both recovered within two to five minutes, and both saw reentry above the exit by six hours. The subsequent active revision widened the profit cushion and drawdown thresholds. After that change, there were no active-revision stop-loss events, so the latest evidence is unresolved rather than proven. Still, the direction of change was consistent with the data: this CONVERT strategy uses StopLoss as a profitable-peak guard, not as a conventional loss cutter, so demanding more confirmation before temporarily moving to cash is appropriate.
Strategy 935 had the broadest stop-loss evidence among active BUY/SELL systems. In the baseline evaluation, stoploss-enabled plans showed very high stop-loss early-exit behavior and frequent rapid recovery. The enabled cohort reported stop-loss exits that were all negative, but many recovered above the exit within minutes. The adaptive change therefore delayed the non-catastrophe stop while keeping a hard five-percent-style catastrophe branch. This is methodologically sound because a trusted-asset recovery strategy is designed to tolerate ordinary drawdowns and avoid repeated shallow stop-out churn.
Strategy 912 had a high weighted stop-loss early-exit rate and substantial exit-near-low behavior. The active revision underperformed recent aggregate context, but the live sample was short, so the applied change was limited to stop-loss logic. It widened drawdown confirmation and retained the hard daily or market-composite breakdown exits. This distinction matters: when stop exits are often followed by quick recovery, adding intraday confirmation can improve exit quality, but if daily NATR becomes extreme or price breaks a hard loss threshold, the system still needs to exit.
Strategy 933 offered a small but concentrated active-revision stop-loss sample. All three stop-loss exits were negative, early, near the next sixty-minute low, and recovered within one, two and five minutes. That is unusually direct evidence of false-stop risk, though the sample is small. The applied stop-loss-only adjustment required additional momentum confirmation around the hard loss branch while keeping the bear-regime and volatility-breakdown branches. This was preferable to changing normal SELL logic, because the negative evidence was specifically stop-loss related.
Strategy 911’s historical evidence also supported widening, but with more caution. StopLoss exits had high recovery rates and were often near local lows, yet the strategy also had weak historical profit factor and poor buy-entry outcomes. Widening stops without improving entries could increase loss duration. The applied change therefore widened stops only modestly and also widened normal sell thresholds. Strategy 932’s stop-loss discussion is secondary because the optimization was blocked at buy validation; no stop-loss change can rescue a formula that cannot enter safely. Overall, the stop-loss lesson is consistent: preserve emergency exits, add confirmation to ordinary stops, and review whether widened stops reduce churn without increasing tail losses.
Strategy analyses
Strategy 934: CONVERT rotation is promising, but the newest buy relaxation is still unproven.
Strategy 934 · version 1
Formula and indicators
Strategy 934 is a dedicated CONVERT relative momentum rotation system. The BUY formula requires multi-timeframe alignment: non-bear entries need daily trend strength, price above the daily EMA, price above intraday EMAs, stacked five-minute EMAs, RSI in a constructive but not overbought band, positive one-minute returns across several horizons, and positive MACD histograms on both one-minute and five-minute frames. Bear-market entries remain possible, but only through a stricter reversal path. A first-entry branch is broader, which makes sense for a rotation strategy that must get capital into an asset before normal switching can begin. The latest adaptive change slightly relaxed non-bear and first-entry thresholds, reducing the five-minute trend-efficiency and one-minute momentum requirements while preserving EMA, MACD, RSI-trend, NATR and cooldown gates. SELL is independent of entry price and focuses on confirmed deterioration: intraday EMA breakdown, negative MACD, falling short-horizon momentum, bear-regime daily weakness or volatility shock below the lower Bollinger band. StopLoss is not a loss stop; it is a profitable-peak guard activated only after profit exists and after drawdown from the post-buy high is confirmed by negative MACD and either RSI-trend or short momentum weakness.
BUY
((MarketRegime-1d != bear AND 5m-TrendEfficiency-24 >= 36 AND TrendStrength-1d >= 48 AND CurrentPrice > 1d-EMA-50 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-EMA-50 > 5m-EMA-100 AND 5m-RSI >= 51 AND 5m-RSI <= 70 AND 1m-2 > 0.06 AND 1m-10 > 0.26 AND 1m-20 > 0.48) OR (MarketRegime-1d = bear AND 5m-TrendEfficiency-24 >= 42 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-RSI >= 48 AND 5m-RSI <= 62 AND 1m-2 > 0.12 AND 1m-10 > 0.38 AND 1m-20 > 0.65) OR (TradeCount = 0 AND 5m-TrendEfficiency-24 >= 28 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-RSI >= 48 AND 5m-RSI <= 68 AND 1m-2 > 0.06 AND 1m-10 > 0.20 AND 1m-20 > 0.34)) AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-RSI-trend = buy AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.55 AND (TimeSinceLastSellMin = null OR TimeSinceLastSellMin > 45)
SELL
((5m-TrendEfficiency-24 >= 34 AND CurrentPrice < 5m-EMA-20 AND 5m-EMA-20 < 5m-EMA-50 AND 5m-MACD-hist < 0 AND 1m-MACD-hist < 0 AND 1m-2 < -0.10 AND 1m-10 < -0.32 AND 1m-20 < -0.55) OR (CurrentPrice < 5m-EMA-50 AND 5m-EMA-50 < 5m-EMA-100 AND 5m-RSI-trend = sell AND 5m-RSI < 45 AND 5m-MACD-hist < 0 AND 1m-10 < -0.25) OR (MarketRegime-1d = bear AND CurrentPrice < 1d-EMA-20 AND 1d-MACD-hist < 0 AND CurrentPrice < 5m-EMA-50 AND 5m-MACD-hist < 0 AND 1m-MACD-hist < 0) OR (5m-NATR-Wilder-14 > 5m-NATR-EMA-14 * 2.20 AND CurrentPrice < 5m-BB-lower AND 1m-2 < -0.35 AND 1m-10 < -0.70 AND 5m-MACD-hist < 0))
STOPLOSS
ProfitSinceBuyPct > 1.00 AND ((5m-NATR-Wilder-14 <= 5m-NATR-EMA-14 * 1.25 AND DrawdownFromHighPct > 0.95) OR (5m-NATR-Wilder-14 > 5m-NATR-EMA-14 * 1.25 AND 5m-NATR-Wilder-14 <= 5m-NATR-EMA-14 * 1.75 AND DrawdownFromHighPct > 1.35) OR (5m-NATR-Wilder-14 > 5m-NATR-EMA-14 * 1.75 AND DrawdownFromHighPct > 1.85)) AND 5m-MACD-hist < 0 AND 1m-MACD-hist < 0 AND (5m-RSI-trend = sell OR 1m-10 < -0.45)
Measured performance
The aggregate KPI profile is mixed but encouraging for the live convert objective. Over thirty days the strategy recorded fifty-five trades, positive profit amount, a seventy-five percent win rate, profit factor above four and positive convert increase. Plan-level data show large dispersion: plan 79 was excellent, with very high profit factor, strong win rate and double-digit convert increase, while plan 80 was negative and inactive plans contributed little useful evidence. The seven-day window had positive realized trade balance and strong profit factor, but aggregate profit amount was slightly negative, indicating mark-to-market or open-position effects. The ninety-day window is harder to interpret because a backtest plan created a very large negative profit amount while realized profit and realized loss remained limited. Coverage also differs across fields; convert increase is available for most but not all plan-days, and stop-loss-day coverage is absent. The active revision after the prior stop-loss change had no formula events for about one day across the active plans. The newer change that relaxes BUY therefore rests on inactivity evidence rather than post-change performance evidence.
| Window | Trades | Profit | Profit factor | Win rate | CONVERT increase | Turnover | StopLoss |
|---|
| 7d | 7 | -0.3725 | 6.1515 | 66.6667% | 3.695% | 3425.6929 | 0 |
| 30d | 55 | 196.6577 | 4.7361 | 75% | 3.6975% | 19960.5891 | 0 |
| 90d | 80 | -21076.3643 | 4.7361 | 75% | 5.1685% | 266756.856 | 0 |
Market alignment
The market setting broadly supports a relative-momentum rotation concept. Bitcoin and major Binance assets were positive over ninety days, CoinRobot breadth had more bull and sideways assets than bear assets, and median trend strength was high. A rotation strategy can exploit dispersion when large-cap assets rise at different speeds. However, the latest global market-cap change was negative and taker-buy participation was not uniformly strong, so the formula’s insistence on EMA alignment, positive MACD and volatility caps remains appropriate. The risk is not that the strategy is too cautious for a bear market; it is that it may become locked into an asset during a sideways pullback if no unit-increasing switch clears the gates. The supplied diagnostics do not include counterfactual best-alternative returns, so lock-in can only be treated as a proxy, not proven underperformance.
Problems found
- The newest active revision has no post-change execution events, so there is no direct evidence yet that the relaxed BUY thresholds improve switch quality or reduce inactivity.
- Prior baseline switch diagnostics showed positive convert increase but weak target-versus-source returns at fifteen, sixty and one hundred eighty minutes, implying switches were not consistently followed by immediate target outperformance.
- Plan-level dispersion is high: strong plans such as plan 79 coexist with inactive or negative plans, so aggregate averages can hide operational differences.
- StopLoss evidence was small but adverse: prior enabled stop-loss exits recovered quickly, supporting the previous widening but leaving little sample for final judgment.
Adaptive change and expected behavior
The applied change is defensible because it addresses the main active problem: inactivity after the stricter stop-loss revision. It does not weaken the bear-market branch, remove trend confirmation, or make SELL profit-dependent. By relaxing only small momentum and trend-efficiency thresholds in non-bear and first-entry paths, the strategy may admit more valid rotations during a constructive but choppy tape. SELL and StopLoss were left unchanged because normal SELL had no new evidence and prior StopLoss evidence already justified caution against re-tightening.
If the hypothesis is correct, the strategy should regain a moderate switch cadence without a spike in bad entries. Convert increase should remain positive, and target-versus-source post-switch returns should improve or at least stop deteriorating. The review focus should be the next several confirmed switches, especially whether neutral-plus-buy or sell-plus-neutral patterns produce unit gains and whether any newly admitted first entries show negative sixty-minute follow-through.
Strategy 935: high activity and win rate mask weak loss distribution in BUY/SELL.
Strategy 935 · version 1
Formula and indicators
Strategy 935 is a trusted-asset BUY/SELL strategy with one small CONVERT deployment. Its BUY logic requires three-hour cooldowns after both buys and sells, a non-bear daily regime, price above the daily EMA, stacked five-minute EMAs, positive MACD on five-minute and one-minute frames, a buy RSI trend, and a volatility cap through five-minute NATR versus its EMA. It distinguishes recovery entries near or below the last sell price from normal entries above that reference. The adaptive change tightened both branches: lower volatility tolerance, higher trend-efficiency and trend-strength thresholds, slightly higher RSI and one-minute momentum requirements. SELL has a profit-guard branch based on profit since buy and drawdown from high, plus confirmed breakdown exits and a bear-market capitulation exit. The adaptive change widened the profit-guard branch and added one-minute MACD confirmation, while leaving severe breakdown exits intact. StopLoss requires a meaningful open loss plus a hard catastrophe condition or multi-timeframe breakdown; the change delayed non-catastrophe loss stops but kept the hard protection.
BUY
(TimeSinceLastSellMin = null OR TimeSinceLastSellMin > 180) AND (TimeSinceLastBuyMin = null OR TimeSinceLastBuyMin > 180) AND MarketRegime-1d != bear AND CurrentPrice > 1d-EMA-50 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-EMA-50 > 5m-EMA-100 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-RSI-trend = buy AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.30 AND (((LastSellPrice != null AND CurrentPrice <= LastSellPrice * 99.5%) AND 5m-TrendEfficiency-24 >= 40 AND TrendStrength-1d >= 50 AND 5m-RSI >= 50 AND 5m-RSI <= 63 AND 1m-2 > 0.10 AND 1m-10 > 0.34 AND 1m-20 > 0.62) OR ((LastSellPrice = null OR CurrentPrice > LastSellPrice * 99.5%) AND 5m-TrendEfficiency-24 >= 44 AND TrendStrength-1d >= 54 AND 5m-RSI >= 54 AND 5m-RSI <= 67 AND 1m-2 > 0.14 AND 1m-10 > 0.42 AND 1m-20 > 0.78))
SELL
((ProfitSinceBuyPct >= 0.85 AND ((5m-NATR-Wilder-14 <= 5m-NATR-EMA-14 * 1.25 AND DrawdownFromHighPct > 0.70) OR (5m-NATR-Wilder-14 > 5m-NATR-EMA-14 * 1.25 AND 5m-NATR-Wilder-14 <= 5m-NATR-EMA-14 * 1.75 AND DrawdownFromHighPct > 1.00) OR (5m-NATR-Wilder-14 > 5m-NATR-EMA-14 * 1.75 AND DrawdownFromHighPct > 1.45)) AND 5m-MACD-hist < 0 AND 1m-MACD-hist < 0) OR (5m-TrendEfficiency-24 >= 42 AND CurrentPrice < 5m-EMA-50 AND 5m-EMA-20 < 5m-EMA-50 AND 5m-EMA-50 < 5m-EMA-100 AND 5m-MACD-hist < 0 AND 1m-MACD-hist < 0 AND 5m-RSI-trend = sell AND 1m-2 < -0.20 AND 1m-10 < -0.55 AND 1m-20 < -0.90) OR (MarketRegime-1d = bear AND CurrentPrice < 1d-BB-lower AND 5m-NATR-Wilder-14 > 5m-NATR-EMA-14 * 1.80 AND 5m-MACD-hist < 0 AND 1m-MACD-hist < 0 AND 1m-10 < -0.65))
STOPLOSS
ProfitSinceBuyPct <= -3.0 AND (CurrentPrice <= LastBuyPrice * 95.0% OR (ProfitSinceBuyPct <= -3.25 AND CurrentPrice < 5m-BB-lower AND CurrentPrice < 5m-EMA-100 AND 5m-MACD-hist < 0 AND 1m-MACD-hist < 0 AND 5m-RSI-trend = sell AND 1m-2 < -0.50 AND 1m-10 < -1.00 AND 1m-20 < -1.50) OR (5m-NATR-Wilder-14 > 5m-NATR-EMA-14 * 3.00 AND CurrentPrice < 5m-BB-lower AND 5m-MACD-hist < 0 AND 1m-MACD-hist < 0 AND 1m-2 < -0.70 AND 1m-10 < -1.30 AND 1m-20 < -1.85))
Measured performance
The strategy generated the most active BUY/SELL evidence among the selected systems. Aggregate seven-day and thirty-day trade counts were high, and win rates were near or above two-thirds. However, profit factor was below one in the aggregate BUY/SELL cohort across the supplied windows, meaning the strategy won often but lost more per losing trade or exited winners too small. Plan-level evidence confirms this dispersion. Plan 66 had a high win rate and near-breakeven longer-window profit, but plan 72 was notably negative over thirty days, plan 84 had heavy turnover and negative seven-day profit, and plan 85 had low win rate. The one CONVERT plan, plan 77, was small but positive, with strong lifetime profit percentage and convert increase. Active-revision diagnostics support the adaptive change: weighted bad-buy rate was elevated, recent bad-buy rate was especially high, sell early-exit evidence persisted, and stop-loss-enabled plans showed high stop-loss recovery behavior. The change therefore targeted all three points: stricter buys, more tolerant profit exits and delayed non-catastrophe stops.
| Window | Trades | Profit | Profit factor | Win rate | CONVERT increase | Turnover | StopLoss |
|---|
| 7d | 338 | -42.11 | 0.8557 | 70.1863% | 0.4219% | 24535.6762 | 0 |
| 30d | 803 | 19687.8698 | 0.8396 | 68.1979% | 0.4219% | 152872.6302 | 0 |
| 90d | 1889 | 25680.9598 | 0.8396 | 68.1979% | 0.4219% | 817062.6036 | 0 |
Market alignment
The market was favorable enough for trusted-asset recovery to work, but not smooth enough for shallow churn. CoinRobot breadth was dominated by sideways and bull names, not bear names, and Bitcoin plus major Binance assets were positive over ninety days. Yet the latest daily market-cap move was negative and intraday taker-buy shares were mixed. A trusted-asset strategy can tolerate pullbacks if assets remain structurally strong, but the evidence shows that frequent trades did not translate into strong profit factor. Tightening entries is aligned with a sideways-heavy market because choppy rebounds can look like recovery but fail quickly. Widening profit guards is also aligned with the high-volatility environment: a median daily NATR near four percent makes tiny drawdown triggers vulnerable to noise.
Problems found
- High win rate did not translate into positive aggregate profit factor, indicating unfavorable payoff asymmetry.
- Weighted bad-buy evidence was elevated, especially in shorter active-revision windows, so entry selectivity was insufficient for the recent tape.
- Normal sell exits frequently showed early-exit and recovery behavior, suggesting winners were being clipped during ordinary pullbacks.
- StopLoss-enabled cohorts had high recovery after stops, implying non-catastrophe stops were too sensitive for trusted-asset recovery behavior.
Adaptive change and expected behavior
The adaptive change is evidence-led and balanced. Tightening the BUY formula addresses bad-buy rates without changing the strategy’s trusted-asset premise. Widening only the profit-guard SELL branch reduces premature profit taking while preserving strong breakdown and bear-market exits. Delaying non-catastrophe StopLoss exits responds to rapid recovery evidence while retaining a hard loss floor. Because the strategy is shared across BUY/SELL and CONVERT, the changes avoid execution-model-specific assumptions and should be safe for the small convert cohort as well.
The desired behavior is fewer but higher-quality BUY/SELL entries, lower churn, and improved profit factor even if raw win rate falls. Normal sells should occur less often on shallow profit pullbacks and more often on confirmed deterioration. StopLoss should fire less on fast wicks but still protect sustained breakdowns. The key next test is whether loss magnitude declines relative to realized gains after the entry tightening, not whether trade count remains high.
Strategy 932: validation blocked an always-false BUY formula, preventing operational failure.
Strategy 932 · version 1
Formula and indicators
Strategy 932 is a flexible active-momentum system with bull, sideways and selective bear-rebound entries, plus continuation paths. The BUY formula is broad in concept but complex in implementation. It first requires either no recent last sell price constraint breach or enough time since sell, then requires five-minute trend efficiency, action-based buy confirmation for one branch, cooldowns, highest-since-sell anti-chase logic, and multiple regime-specific momentum conditions. It also includes continuation entries for non-bear regimes when one-minute momentum is strong and EMAs align. SELL requires action equals sell, a minimum profit threshold, and then profit-giveback, MACD momentum reversal, EMA weakness or overbought RSI pullback. StopLoss requires a meaningful loss plus hard catastrophe or confirmed breakdown conditions, with an additional profitable drawdown guard. The active formula state shows a modified StopLoss with slightly harsher hard-loss confirmation and wider profitable drawdown guard. However, the optimization attempt was validation-blocked because the proposed BUY formula evaluated always false under scenario validation.
BUY
(LastSellPrice = null OR CurrentPrice <= LastSellPrice * 1.003 OR TimeSinceLastSellMin > 360) AND (5m-TrendEfficiency-24 >= 35 AND ((action = buy AND ((TimeSinceLastSellMin = null OR TimeSinceLastSellMin > 90) OR (MarketRegime-1d = bull AND 5m-TrendEfficiency-24 >= 24 AND TrendStrength-1d >= 68 AND 1m-10 > 0.65 AND 1m-20 > 1.05 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-MACD-hist > 0)) AND (HighestSinceSell = null OR CurrentPrice < HighestSinceSell * 99.35% OR (MarketRegime-1d = bull AND 5m-TrendEfficiency-24 >= 24 AND TrendStrength-1d >= 70 AND 1m-10 > 0.75 AND 1m-20 > 1.20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-EMA-50 > 5m-EMA-100)) AND ((MarketRegime-1d = bull AND 5m-TrendEfficiency-24 >= 24 AND TrendStrength-1d >= 56 AND CurrentPrice > 1d-EMA-50 AND 1d-MACD-hist > -0.05 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 1m-2 > 0.08 AND 1m-10 > 0.24 AND 1m-20 > 0.03 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-RSI > 50 AND 5m-RSI < 67 AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.35) OR (MarketRegime-1d = sideways AND 5m-TrendEfficiency-24 >= 40 AND TrendStrength-1d >= 38 AND TrendStrength-1d <= 58 AND CurrentPrice > 1d-BB-middle AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 >= 5m-EMA-50 AND 1m-2 > 0.06 AND 1m-10 > 0.22 AND 1m-20 > -0.05 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-RSI > 49 AND 5m-RSI < 63 AND CurrentPrice < 5m-BB-upper) OR (MarketRegime-1d = bear AND TrendStrength-1d <= 30 AND 1d-RSI < 34 AND CurrentPrice > 1d-BB-lower AND CurrentPrice > 5m-EMA-20 AND 1m-2 > 0.10 AND 1m-10 > 0.32 AND 1m-20 > 0.10 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-RSI > 47 AND 5m-RSI < 57)) AND (TimeSinceLastBuyMin = null OR TimeSinceLastBuyMin > 120)) OR (MarketRegime-1d != bear AND (5m-TrendEfficiency-24 >= 30 OR 1m-20 > 1.20) AND TrendStrength-1d >= 45 AND CurrentPrice > 1d-EMA-50 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 1m-2 > 0.10 AND 1m-10 > 0.45 AND 1m-20 > 0.90 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-RSI >= 55 AND 5m-RSI <= 74 AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.50 AND (TimeSinceLastBuyMin = null OR TimeSinceLastBuyMin > 120))) OR (MarketRegime-1d != bear AND (5m-TrendEfficiency-24 >= 30 OR 1m-20 > 1.20) AND TrendStrength-1d >= 42 AND CurrentPrice > 1d-EMA-50 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-EMA-50 > 5m-EMA-100 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 1m-2 > 0.08 AND 1m-10 > 0.30 AND 1m-20 > 0.65 AND 5m-RSI >= 53 AND 5m-RSI <= 78 AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.65 AND (TimeSinceLastBuyMin = null OR TimeSinceLastBuyMin > 90)))
SELL
action = sell AND ProfitSinceBuyPct > 0.35 AND (((ProfitSinceBuyPct > 0.65 AND DrawdownFromHighPct > 0.50) OR (ProfitSinceBuyPct > 1.35 AND DrawdownFromHighPct > 0.34) OR (ProfitSinceBuyPct > 2.75 AND DrawdownFromHighPct > 0.26)) OR (5m-MACD-hist < 0 AND 1m-MACD-hist < 0 AND 1m-10 < -0.18 AND DrawdownFromHighPct > 0.20) OR (CurrentPrice < 5m-EMA-50 AND 1m-10 < -0.25) OR (5m-RSI > 78 AND DrawdownFromHighPct > 0.35))
STOPLOSS
(ProfitSinceBuyPct <= -2.5 AND ((CurrentPrice < LastBuyPrice * 94.8%) OR (CurrentPrice < LastBuyPrice * 95.2% AND 1m-10 < -0.45 AND 5m-MACD-hist < 0) OR (MarketRegime-1d = bear AND CurrentPrice < 5m-BB-lower AND CurrentPrice < 5m-EMA-100 AND 1m-2 < -0.55 AND 1m-10 < -1.00 AND 1m-20 < -1.50 AND 1m-MACD-hist < 0 AND 5m-MACD-hist < 0) OR (5m-NATR-Wilder-14 > 5m-NATR-EMA-14 * 2.90 AND CurrentPrice < 5m-BB-lower AND 1m-2 < -0.70 AND 1m-10 < -1.20 AND 1m-MACD-hist < 0 AND 5m-MACD-hist < 0))) OR (ProfitSinceBuyPct > 0.5 AND DrawdownFromHighPct > 1.45 AND 1m-MACD-hist < 0 AND 5m-MACD-hist < 0 AND (1m-10 < -0.20 OR CurrentPrice < 5m-EMA-50))
Measured performance
Historical aggregate KPIs are difficult to use for this strategy because execution-model and coverage fields vary materially. The seven-day window had no trades but negative profit amount in the convert cohort, suggesting mark-to-market or open-position impact. The thirty-day window showed thirty-one trades, positive profit amount, profit factor above one and very high convert increase. The ninety-day window included thousands of trades and very large turnover, but win-rate and profit-factor coverage was limited, and historical buy_sell figures appear dominated by older or differently scoped records. The most important evidence is therefore not the headline historical KPI but the validation result. A formula that is always false under scenario validation would prevent intended entries, freeze capital allocation, or cause the strategy to depend on fallback behavior outside the desired logic. Since the optimization status was validation_blocked, the system correctly prevented deployment of a broken change.
| Window | Trades | Profit | Profit factor | Win rate | CONVERT increase | Turnover | StopLoss |
|---|
| 7d | 0 | -10.6796 | — | — | 0% | 0 | 0 |
| 30d | 31 | 52.9613 | 1.3706 | 66.6667% | 87.3735% | 9479.0935 | 0 |
| 90d | 4171 | 22095.7938 | 0.8847 | 50% | 12.9669% | 1156419.8894 | 0 |
Market alignment
The broader market could have supported a flexible momentum system: large-cap returns were positive, breadth was mostly sideways or bull, and median trend strength was high. But flexible systems are especially vulnerable to formula complexity. In a sideways-heavy market, multiple branches and anti-chase conditions are useful only if they remain logically satisfiable. The validation failure shows that the proposed BUY logic became too constrained or syntactically inconsistent under the scenario suite. This is particularly dangerous because the market context included both opportunity and short-term weakness; an always-false formula would miss recoveries and trend continuations while still leaving existing positions subject to SELL or StopLoss logic.
Problems found
- The proposed BUY formula failed validation as formula_always_false, making performance interpretation secondary until logic is repaired.
- Recent seven-day live evidence showed no trades, so there is no current confirmation that the strategy is participating in the market.
- Historical KPI windows include materially different coverage and execution scopes, limiting direct comparison across seven, thirty and ninety days.
- The strategy’s broad conceptual flexibility increases the risk of logical conflicts among action gates, anti-chase filters, regime filters and cooldown terms.
Adaptive change and expected behavior
No applied optimization should be accepted until the BUY validation issue is resolved. The appropriate research response is to simplify and test the formula branch by branch. The existing StopLoss concept may be reasonable, but StopLoss refinements are secondary if entries cannot validate. Any next change should first restore at least one scenario-validated bullish and recovery-buy path while preserving volatility caps and anti-chase controls. The validation block is a positive governance outcome because it prevented deployment of a non-operational formula.
Until a corrected BUY formula is validated, the strategy should be treated as operationally unresolved. The next expected healthy behavior is not immediate profit but successful scenario validation: bullish and buy-signal cases should evaluate true, bearish and hard-stop cases should remain false for BUY, and missing values should fail safely. After validation, live evidence should be collected separately for action-based entries and continuation entries to determine whether the broad design actually improves participation without reviving stop-loss churn.
Strategy 912: CONVERT implementation outperformed BUY/SELL, but StopLoss still needs confirmation.
Strategy 912 · version 1
Formula and indicators
Strategy 912 is a swing trend guard originally described for multi-day to one-to-two-week moves. The BUY formula requires action equals buy and then selects among bull, sideways, bear-rebound and strong-trend branches. Bull entries demand stacked daily EMAs, positive daily MACD, controlled daily RSI and NATR, plus five-minute EMA, MACD, RSI-trend and NATR confirmation. Sideways entries require price above daily mid-band but below upper band, moderate trend strength and intraday confirmation. Bear entries are limited to rebound conditions with low daily RSI and price above the lower Bollinger band. SELL is regime-specific and uses daily NATR-scaled profit targets, giveback thresholds, overbought or breakdown conditions. StopLoss includes high daily NATR, hard loss, drawdown from high, profitable giveback, market-composite breakdown and lower-band breakdown branches. The latest adaptive change widened parts of StopLoss confirmation, especially drawdown-from-high exits, while preserving hard loss and daily breakdown protections.
BUY
action = buy AND ((MarketRegime-1d = bull AND TrendStrength-1d >= 63 AND CurrentPrice > 1d-EMA-20 AND 1d-EMA-20 > 1d-EMA-50 AND 1d-EMA-50 > 1d-EMA-100 AND 1d-MACD-hist > 0 AND 1d-RSI >= 51 AND 1d-RSI <= 67 AND 1d-NATR-Wilder-14 < 7.0 AND ((CurrentPrice < 1d-BB-upper AND 1d-RSI <= 63) OR (CurrentPrice >= 1d-BB-upper AND TrendStrength-1d >= 82 AND 1d-RSI >= 58 AND 1d-RSI <= 68 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-MACD-hist > 0 AND 5m-RSI-trend = buy AND 5m-RSI >= 56 AND 5m-RSI <= 63)) AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-MACD-hist > 0 AND 5m-RSI-trend = buy AND 5m-RSI >= 50 AND 5m-RSI <= 64 AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.10) OR (MarketRegime-1d = sideways AND TrendStrength-1d >= 38 AND TrendStrength-1d <= 62 AND CurrentPrice > 1d-EMA-50 AND CurrentPrice > 1d-BB-middle AND CurrentPrice < 1d-BB-upper AND 1d-RSI >= 47 AND 1d-RSI <= 58 AND 1d-MACD-hist > 0 AND 1d-NATR-Wilder-14 < 6.2 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-RSI-trend = buy AND 5m-MACD-hist > 0 AND 5m-RSI >= 50 AND 5m-RSI <= 60 AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.09) OR (MarketRegime-1d = bear AND TrendStrength-1d >= 20 AND TrendStrength-1d <= 36 AND 1d-RSI >= 31 AND 1d-RSI <= 41 AND CurrentPrice > 1d-BB-lower AND CurrentPrice < 1d-EMA-20 AND 1d-MACD-hist > 0 AND 1d-NATR-Wilder-14 < 5.6 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-RSI-trend = buy AND 5m-MACD-hist > 0 AND 5m-RSI >= 50 AND 5m-RSI <= 56 AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.00) OR (TrendStrength-1d >= 70 AND CurrentPrice > 1d-EMA-20 AND 1d-EMA-20 > 1d-EMA-50 AND 1d-EMA-50 > 1d-EMA-100 AND 1d-MACD-hist > 0 AND 1d-RSI >= 52 AND 1d-RSI <= 64 AND 1d-NATR-Wilder-14 < 6.8 AND CurrentPrice < 1d-BB-upper AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-MACD-hist > 0 AND 5m-RSI-trend = buy AND 5m-RSI >= 52 AND 5m-RSI <= 62 AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.07)) AND (TimeSinceLastBuyMin = null OR TimeSinceLastBuyMin > 1440)
SELL
((MarketRegime-1d = bull AND ((ProfitSinceBuyPct > 1d-NATR-Wilder-14 * 0.90 AND DrawdownFromHighPct > 1d-NATR-Wilder-14 * 0.45) OR ProfitSinceBuyPct > 1d-NATR-Wilder-14 * 1.35 OR 5m-RSI > 73 OR (CurrentPrice < 5m-EMA-50 AND 5m-MACD-hist < 0 AND DrawdownFromHighPct > 5m-NATR-Wilder-14 * 1.20) OR (CurrentPrice < 1d-EMA-20 AND 1d-MACD-hist < 0) OR DrawdownFromHighPct > 1d-NATR-Wilder-14 * 0.90)) OR (MarketRegime-1d = sideways AND ((ProfitSinceBuyPct > 1d-NATR-Wilder-14 * 0.65 AND DrawdownFromHighPct > 1d-NATR-Wilder-14 * 0.35) OR ProfitSinceBuyPct > 1d-NATR-Wilder-14 * 1.05 OR CurrentPrice > 1d-BB-upper OR 1d-RSI > 65 OR (CurrentPrice < 1d-BB-middle AND 5m-MACD-hist < 0) OR DrawdownFromHighPct > 1d-NATR-Wilder-14 * 0.70)) OR (MarketRegime-1d = bear AND ((ProfitSinceBuyPct > 1d-NATR-Wilder-14 * 0.45 AND DrawdownFromHighPct > 1d-NATR-Wilder-14 * 0.25) OR ProfitSinceBuyPct > 1d-NATR-Wilder-14 * 0.75 OR 1d-RSI > 43 OR (CurrentPrice < 5m-EMA-50 AND 5m-MACD-hist < 0) OR DrawdownFromHighPct > 1d-NATR-Wilder-14 * 0.50)))
STOPLOSS
(1d-NATR-Wilder-14 > 9.0 OR CurrentPrice < LastBuyPrice * 97.2% OR (DrawdownFromHighPct > 1d-NATR-Wilder-14 * 1.20 AND (CurrentPrice < 5m-EMA-50 OR 5m-MACD-hist < 0)) OR DrawdownFromHighPct > 1d-NATR-Wilder-14 * 1.45 OR (ProfitSinceBuyPct > 0.5 AND DrawdownFromHighPct > 1.35 AND (5m-MACD-hist < 0 OR CurrentPrice < 5m-EMA-50)) OR (market-composit-StopLoss = YES AND CurrentPrice < 1d-EMA-20 AND 1d-MACD-hist < 0) OR (CurrentPrice < 5m-BB-lower AND 5m-MACD-hist < 0 AND DrawdownFromHighPct > 5m-NATR-Wilder-14 * 1.60) OR (MarketRegime-1d = bear AND CurrentPrice < 1d-BB-lower AND 1d-MACD-hist < 0))
Measured performance
The execution-model split is crucial. BUY/SELL plan 1 had poor results: seven-day profit amount was negative, thirty-day profit amount was negative, win rate was low and profit factor was well below one. CONVERT plan 3 was much stronger, with positive seven-day and thirty-day profit amount, positive convert increase above twenty percent in daily KPIs, and profit factor above two over thirty days. Aggregate strategy KPIs therefore look better than BUY/SELL alone because the convert plan carried performance. Active-revision evidence after the previous change was short but informative: sixty-two events, twenty-one buys, fourteen sells and seventeen stop-loss exits. StopLoss early-exit rate was high, exit-near-low behavior was significant, and many stop exits recovered quickly. Weighted evidence supported a stop-loss-only adjustment rather than broad retuning. The newest change therefore gives drawdown exits more room while retaining emergency conditions. However, the active revision also showed weak BUY/SELL realized trade metrics, so the strategy should not be judged solely by the stronger CONVERT cohort.
| Window | Trades | Profit | Profit factor | Win rate | CONVERT increase | Turnover | StopLoss |
|---|
| 7d | 136 | 6.4873 | 1.0193 | 28.169% | 21.3793% | 12119.111 | 0 |
| 30d | 301 | 23.3179 | 1.6024 | 31.8182% | 21.3793% | 26670.6899 | 0 |
| 90d | 305 | 23.0582 | 1.6024 | 31.8182% | 21.3793% | 26710.4301 | 0 |
Market alignment
The swing-trend design aligns with a market where daily trends are positive but intraday volatility is high. Bitcoin and major assets advanced over ninety days, which benefits swing logic, but the latest daily market-cap decline and sideways-heavy breadth can create false intraday exits. The formula’s daily EMA, MACD, RSI and NATR filters are appropriate for avoiding weak swing entries. The problem is exit timing: when a swing strategy exits on short intraday drawdowns, it can abandon multi-day trends too early. The CONVERT plan’s stronger evidence suggests that rotating among assets, rather than repeatedly buying and selling into cash, may better capture this environment.
Problems found
- BUY/SELL performance was weak relative to CONVERT, with low win rate and profit factor below one in the buy_sell cohort.
- StopLoss exits in the active sample had high early-exit and recovery rates, suggesting ordinary pullbacks were being treated as risk events.
- The active-revision sample was short, so the newest StopLoss change is evidence-supported but not yet confirmed.
- Some initial CONVERT entries had neutral or false formula states, requiring careful attribution rather than assuming every entry was a BUY signal.
Adaptive change and expected behavior
The stop-loss-only adjustment is appropriate because the strongest active evidence pointed to premature stop exits, not bad buy logic. Widening drawdown confirmation and requiring more intraday weakness for ordinary stops should reduce sell-low churn. Keeping the hard loss, high-volatility and daily breakdown branches preserves protection against true adverse moves. SELL and BUY were left unchanged, which avoids overfitting a short live sample and preserves the swing structure.
The expected improvement is fewer StopLoss exits during normal intraday pullbacks and better retention of swing positions. BUY/SELL profit factor should improve only if fewer exits crystallize small losses before recovery. CONVERT should maintain positive unit accumulation if rotations remain selective. The next review should compare stop-loss recovery rates before and after the change and separately evaluate BUY/SELL and CONVERT because their historical performance profiles differ sharply.
Strategy 933: selective BUY/SELL is profitable longer term, but latest stop evidence is adverse.
Strategy 933 · version 1
Formula and indicators
Strategy 933 is a selective trend-momentum BUY/SELL strategy. BUY excludes bear regimes, requires five-minute trend efficiency, daily trend strength, price above the daily EMA and five-minute EMA, positive five-minute and one-minute MACD, NATR below its EMA-scaled volatility cap, and RSI in a controlled trend-participation band. It also applies cooldowns and anti-chase logic based on last sell price and highest since sell. There are two entry styles: an action equals buy path with moderate one-minute momentum and a continuation path requiring stronger trend strength, stacked five-minute EMAs and stronger one-minute returns. SELL requires action equals sell and at least a profit threshold, then exits on profit giveback, MACD reversal, EMA weakness or overbought RSI drawdown. StopLoss activates on deeper loss with either hard price damage, bear-regime breakdown or volatility shock, plus a profitable drawdown branch. The latest applied change was stop-loss-only: it required additional momentum confirmation around the hard-loss branch while leaving BUY and normal SELL unchanged.
BUY
(LastSellPrice = null OR CurrentPrice <= LastSellPrice * 1.003 OR TimeSinceLastSellMin > 360) AND MarketRegime-1d != bear AND 5m-TrendEfficiency-24 >= 36 AND TrendStrength-1d >= 48 AND CurrentPrice > 1d-EMA-50 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.35 AND 5m-RSI >= 53 AND 5m-RSI <= 68 AND (TimeSinceLastBuyMin = null OR TimeSinceLastBuyMin > 60) AND ((action = buy AND (TimeSinceLastSellMin = null OR TimeSinceLastSellMin > 60) AND 1m-2 > 0.08 AND 1m-10 > 0.28 AND 1m-20 > 0.20 AND (HighestSinceSell = null OR CurrentPrice < HighestSinceSell * 99.55% OR (TrendStrength-1d >= 68 AND 1m-20 > 0.95))) OR (TrendStrength-1d >= 56 AND 5m-EMA-50 > 5m-EMA-100 AND 1m-2 > 0.10 AND 1m-10 > 0.42 AND 1m-20 > 0.82 AND 5m-RSI <= 66))
SELL
action = sell AND ProfitSinceBuyPct >= 0.65 AND (((ProfitSinceBuyPct >= 0.65 AND DrawdownFromHighPct > 0.50) OR (ProfitSinceBuyPct >= 1.20 AND DrawdownFromHighPct > 0.38) OR (ProfitSinceBuyPct >= 2.20 AND DrawdownFromHighPct > 0.30)) OR (5m-MACD-hist < 0 AND 1m-MACD-hist < 0 AND 1m-10 < -0.12) OR (CurrentPrice < 5m-EMA-50 AND 1m-10 < -0.18) OR (5m-RSI > 74 AND DrawdownFromHighPct > 0.28))
STOPLOSS
(ProfitSinceBuyPct <= -2.5 AND (((CurrentPrice < LastBuyPrice * 95.2%) AND ((CurrentPrice < LastBuyPrice * 94.6%) OR (1m-2 < -0.35 AND 1m-10 < -0.65 AND 1m-MACD-hist < 0 AND 5m-MACD-hist < 0))) OR (MarketRegime-1d = bear AND CurrentPrice < 5m-BB-lower AND CurrentPrice < 5m-EMA-100 AND 1m-2 < -0.50 AND 1m-10 < -0.90 AND 1m-20 < -1.40 AND 1m-MACD-hist < 0 AND 5m-MACD-hist < 0) OR (5m-NATR-Wilder-14 > 5m-NATR-EMA-14 * 2.80 AND CurrentPrice < 5m-BB-lower AND 1m-2 < -0.65 AND 1m-10 < -1.10 AND 1m-MACD-hist < 0 AND 5m-MACD-hist < 0))) OR (ProfitSinceBuyPct > 0.5 AND DrawdownFromHighPct > 1.25 AND 1m-MACD-hist < 0 AND 5m-MACD-hist < 0)
Measured performance
The longer-window aggregate profile is better than many selected strategies. Over thirty days the strategy had high trade count, positive profit amount, win rate above seventy percent and profit factor above one. Plan 69 was especially strong over thirty days, while plan 68 was near breakeven to modestly positive longer term but negative over the latest week. The seven-day aggregate was slightly negative despite a high win rate, showing that recent payoff distribution weakened. Active-revision evidence after the previous sell-guard change was small but concentrated: thirty-nine events, eighteen buys, eighteen sells and three stop-loss exits. Bad-buy rate was elevated, but the AI notes correctly avoided changing BUY after only about fifty active-revision hours. Normal sells were early but not negative, while all three StopLoss exits were negative, early, near the next sixty-minute low, and recovered within one, two and five minutes. That very specific evidence supports changing StopLoss rather than normal SELL.
| Window | Trades | Profit | Profit factor | Win rate | CONVERT increase | Turnover | StopLoss |
|---|
| 7d | 286 | -4.732 | 1.4756 | 82.3077% | — | 6108.734 | 0 |
| 30d | 1750 | 52.1273 | 1.2824 | 74.0645% | — | 38260.0202 | 0 |
| 90d | 1999 | 41.8198 | 1.2041 | 68.764% | — | 43521.0473 | 0 |
Market alignment
The formula is well aligned with the supplied market breadth because it excludes bear-regime entries and requires trend and volatility quality. In a market with many sideways names and a meaningful number of bull names, selective trend participation is preferable to broad dip buying. However, the latest daily market decline and high intraday volatility can create false hard-loss signals even inside otherwise valid trends. The system’s recent issue was not lack of trend filters; it was stop-loss sensitivity in an environment where fast recoveries occurred after exits. This explains why the adaptive change preserved entry rules and focused on stop confirmation.
Problems found
- Recent active-revision bad-buy rate was elevated, but the sample was too small for a reliable buy rewrite.
- The seven-day aggregate profit amount was negative despite a high win rate, implying recent loss sizing or mark-to-market drag.
- All three active StopLoss exits were early, near local lows and quickly recovered, a concentrated false-stop signal.
- Plan-level dispersion remains important: the broad plan outperformed the more selective enabled plan over several windows.
Adaptive change and expected behavior
The stop-loss-only adjustment is proportional. It targets the branch that produced the clearest adverse evidence while avoiding overreaction to small-sample bad-buy data. Keeping normal SELL unchanged is justified because normal sells were not negative in the active sample and because the previous sell-guard widening had already been applied. The revised stop rule should reduce hard-loss triggers caused by transient wicks unless momentum confirms the move.
The strategy should continue to trade actively, but enabled StopLoss exits should become rarer and more confirmed. If the change works, stop-loss recovery within one to five minutes should fall, while catastrophic breakdown protection remains intact. If bad-buy rates remain high after a larger sample, the next adjustment should revisit entry momentum and trend-efficiency thresholds rather than widening stops further.
Strategy 911: inactive plans limit proof, but historical exits show severe sell-low churn.
Strategy 911 · version 1
Formula and indicators
Strategy 911 is a trend pullback guard using early momentum entries, profit pullback exits and fast deterioration exits. BUY has an action equals buy branch requiring five-minute trend efficiency, post-sell cooldown or a strong bull exception, highest-since-sell anti-chase logic, and regime-specific momentum filters. It also has a continuation branch for non-bear regimes with strong one-minute momentum, price above daily and five-minute EMAs, positive MACD and controlled volatility. SELL requires action equals sell and then checks profit giveback tiers, accelerating one-minute deterioration, bull-regime MACD or EMA weakness, sideways overbought or band-extension conditions, and bear-regime weakness. StopLoss originally had a relatively tight hard-loss branch, severe short-momentum branch, bear lower-band branch, volatility-spike branch and drawdown-from-high branch. The adaptive change widened SELL profit-giveback and momentum thresholds and modestly widened StopLoss hard-loss, volatility and drawdown triggers while keeping severe-drop guards.
BUY
(action = buy AND 5m-TrendEfficiency-24 >= 35 AND ((TimeSinceLastSellMin = null OR TimeSinceLastSellMin > 90) OR (MarketRegime-1d = bull AND 5m-TrendEfficiency-24 >= 24 AND TrendStrength-1d >= 68 AND 1m-10 > 0.65 AND 1m-20 > 1.05 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-MACD-hist > 0)) AND (HighestSinceSell = null OR CurrentPrice < HighestSinceSell * 99.35% OR (MarketRegime-1d = bull AND 5m-TrendEfficiency-24 >= 24 AND TrendStrength-1d >= 70 AND 1m-10 > 0.75 AND 1m-20 > 1.20 AND 5m-EMA-20 > 5m-EMA-50 AND 5m-EMA-50 > 5m-EMA-100)) AND ((MarketRegime-1d = bull AND 5m-TrendEfficiency-24 >= 24 AND TrendStrength-1d >= 56 AND CurrentPrice > 1d-EMA-50 AND 1d-MACD-hist > -0.05 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 1m-2 > 0.08 AND 1m-10 > 0.24 AND 1m-20 > 0.03 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-RSI > 50 AND 5m-RSI < 67 AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.35) OR (MarketRegime-1d = sideways AND 5m-TrendEfficiency-24 >= 40 AND TrendStrength-1d >= 38 AND TrendStrength-1d <= 58 AND CurrentPrice > 1d-BB-middle AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 >= 5m-EMA-50 AND 1m-2 > 0.06 AND 1m-10 > 0.22 AND 1m-20 > -0.05 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-RSI > 49 AND 5m-RSI < 63 AND CurrentPrice < 5m-BB-upper) OR (MarketRegime-1d = bear AND TrendStrength-1d <= 30 AND 1d-RSI < 34 AND CurrentPrice > 1d-BB-lower AND CurrentPrice > 5m-EMA-20 AND 1m-2 > 0.10 AND 1m-10 > 0.32 AND 1m-20 > 0.10 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-RSI > 47 AND 5m-RSI < 57)) AND (TimeSinceLastBuyMin = null OR TimeSinceLastBuyMin > 120)) OR (MarketRegime-1d != bear AND (5m-TrendEfficiency-24 >= 30 OR 1m-20 > 1.20) AND TrendStrength-1d >= 45 AND CurrentPrice > 1d-EMA-50 AND CurrentPrice > 5m-EMA-20 AND 5m-EMA-20 > 5m-EMA-50 AND 1m-2 > 0.10 AND 1m-10 > 0.45 AND 1m-20 > 0.90 AND 5m-MACD-hist > 0 AND 1m-MACD-hist > 0 AND 5m-RSI >= 55 AND 5m-RSI <= 74 AND 5m-NATR-Wilder-14 < 5m-NATR-EMA-14 * 1.50 AND (TimeSinceLastBuyMin = null OR TimeSinceLastBuyMin > 120))
SELL
action = sell AND (((ProfitSinceBuyPct > 0.45 AND DrawdownFromHighPct > 0.65) OR (ProfitSinceBuyPct > 0.90 AND DrawdownFromHighPct > 0.48) OR (ProfitSinceBuyPct > 1.50 AND DrawdownFromHighPct > 0.36)) OR (1m-2 < -0.35 AND 1m-10 < -0.70 AND 1m-20 < -1.05 AND 1m-2 < 1m-10 AND 1m-10 < 1m-20) OR (MarketRegime-1d = bull AND ((5m-MACD-hist < 0 AND 1m-MACD-hist < 0 AND 1m-2 < -0.14 AND 1m-10 < -0.14) OR (CurrentPrice < 5m-EMA-50 AND 1m-10 < -0.24) OR DrawdownFromHighPct > 5m-NATR-Wilder-14 * 1.40)) OR (MarketRegime-1d = sideways AND (5m-RSI > 68 OR (CurrentPrice > 5m-BB-upper AND 5m-RSI > 64) OR (5m-MACD-hist < 0 AND 1m-2 < -0.16 AND 1m-10 < -0.16) OR DrawdownFromHighPct > 5m-NATR-Wilder-14 * 1.15)) OR (MarketRegime-1d = bear AND ((CurrentPrice < 5m-EMA-20 AND 1m-10 < -0.12) OR (5m-MACD-hist < 0 AND 1m-MACD-hist < 0 AND 1m-2 < -0.08) OR 1m-10 < -0.35 OR DrawdownFromHighPct > 5m-NATR-Wilder-14 * 0.80)))
STOPLOSS
CurrentPrice < LastBuyPrice * 98.6% OR (1m-2 < -0.55 AND 1m-10 < -1.00 AND 1m-20 < -1.45 AND 1m-2 < 1m-10 AND 1m-10 < 1m-20) OR (MarketRegime-1d = bear AND CurrentPrice < 5m-BB-lower AND 5m-MACD-hist < 0) OR 5m-NATR-Wilder-14 > 5m-NATR-EMA-14 * 2.50 OR DrawdownFromHighPct > 3.5
Measured performance
The strategy’s current live relevance is limited because all listed plans were inactive at the report timestamp, with many last trades more than a month old. Seven-day and thirty-day KPI windows were empty. Ninety-day daily KPIs contained historical turnover and profit amounts, but win-rate and profit-factor coverage were unavailable for many plans. Revision KPI data were weak: ninety-day baseline profit factor was very low, trade win rate was low, and convert increase was slightly negative. Active baseline trade evaluation before inactivity showed eighty-nine events, high bad-buy rate, high early-exit rate, frequent exits near the next sixty-minute low, many negative exits, and strong short-term recovery after both normal sells and StopLoss exits. Convert diagnostics were also weak: only two switches, near-zero negative convert increase and true buy-state initial entries with high bad-entry rate and negative average sixty-minute return. This is not a strategy ready for aggressive expansion.
| Window | Trades | Profit | Profit factor | Win rate | CONVERT increase | Turnover | StopLoss |
|---|
| 7d | 0 | 0 | — | — | — | 0 | 0 |
| 30d | 0 | 0 | — | — | — | 0 | 0 |
| 90d | 921 | 1548.6742 | — | — | -2.0154% | 125125.7551 | 0 |
Market alignment
The concept of buying trend pullbacks can work in the supplied broader market because trend strength and ninety-day large-cap returns were positive. But the historical execution evidence suggests the strategy was too reactive inside noisy pullbacks. In a sideways-heavy market, an early pullback strategy must distinguish a real continuation from a range rebound. The current BUY formula already contains trend-efficiency gates, so the AI did not relax entries. That is appropriate because prior true buy entries performed poorly. The market context supports preserving quality filters and reducing exit churn, not broadening participation.
Problems found
- All plans were inactive at the report timestamp, so the applied change is a reactivation hypothesis rather than live-confirmed improvement.
- Historical profit factor and trade win rate were weak, and convert increase was slightly negative.
- Normal sells were frequently negative, early, near local lows and followed by rapid recovery, indicating severe sell-low behavior.
- StopLoss exits also recovered quickly, but widening stops without improving entries could increase exposure if buy quality remains poor.
Adaptive change and expected behavior
The change widened SELL and StopLoss only modestly while leaving BUY unchanged. This is defensible because exit-quality evidence was strong, but entry evidence was not good enough to justify looser buys. Wider profit-giveback and momentum thresholds should reduce exits caused by ordinary pullback noise. Wider StopLoss thresholds should reduce immediate wick exits, while severe momentum, bear-band and volatility branches remain available for true risk events.
If reactivated, the strategy should trade less frenetically, exit fewer positions at local lows and show lower short-term recovery after sells and stops. However, success also requires buy quality to improve under the existing filters. The next review should demand new live evidence before any further change: at minimum, true buy-state entries should show better sixty-minute follow-through and sell-negative-exit rates should decline materially.