Skip to content

Latest commit

 

History

History
245 lines (243 loc) · 21.2 KB

File metadata and controls

245 lines (243 loc) · 21.2 KB

AutoResearch — Experiment Index

237 experiments, 62 kept, 175 rejected Best: 8.176 (exp199) | Baseline: 0.421 Last updated: 2026-03-22T14:32:32.262Z

# Score Status Hypothesis
exp001 0.421
exp002 0.289
exp003 0.39
exp004 0.486
exp005 0.486
exp006 0.472
exp007 0.439
exp008 0.486
exp009 0.486
exp010 0.346
exp011 0.523
exp012 0.531
exp013 0.219
exp014 0.498
exp015 0.485
exp016 0.526
exp017 0.54
exp018 0.505
exp019 0.529
exp020 -999
exp021 -999
exp022 -0.381
exp023 -0.113
exp024 0.524
exp025 0.524
exp026 0.236
exp027 0.105
exp028 -0.313
exp029 0.564
exp030 0.61
baseline 0.61 Initial strategy — baseline measurement
exp032 -999 Unknown mutation
baseline 0.61 Initial strategy — baseline measurement
exp034 -999 LLM-proposed mutation
baseline 0.61 Initial strategy — baseline measurement
exp036 0.525 I am increasing the deviationThreshold from 0.025 to 0.03 to make entry condit
exp037 0.615 I will decrease the atrPeriod from 14 to 7 to make the position sizing more re
exp038 0.598 I will decrease the rsiPeriod from 14 to 7 to make the RSI indicator more reac
exp039 0.091 I will decrease the vwapPeriod from 20 to 15 to make the VWAP indicator more r
exp040 0.566 I will decrease the exitThreshold from 0.01 to 0.008 to make the exit conditio
baseline 0.615 Initial strategy — baseline measurement
exp042 0.484 I will increase the cooldown period from 3 to 5 bars. This aims to reduce pote
exp043 0.553 I will slightly increase the basePositionSize from 0.15 to 0.17 to modestly in
exp044 0.511 I will increase the deviationThreshold from 0.025 to 0.03 to make the entry co
baseline 0.615 Initial strategy — baseline measurement
exp046 -999 I will increase the deviationThreshold from 0.025 to 0.035 to filter out weake
exp047 0.542 I will increase the rsiPeriod from 14 to 20 to smooth out RSI readings and red
exp048 0.598 I will reduce the rsiPeriod from 14 to 10 to make RSI more responsive to recen
exp049 0.484 I will increase the cooldown from 3 to 5 bars to reduce excessive trading and
exp050 0.517 I will increase the vwapPeriod from 20 to 25 to use a longer-term VWAP baselin
exp051 0.596 I will reduce the basePositionSize from 0.15 to 0.10 to decrease leverage and
exp052 0.547 I will reduce the deviationThreshold from 0.025 to 0.018 to make entry signals
exp053 0.671 I will increase the exitThreshold from 0.01 to 0.015 to hold winning positions
exp054 -999 I will increase the deviationThreshold from 0.025 to 0.035 to make entry signa
exp055 0.653 I will increase the rsiPeriod from 14 to 10 to make RSI more responsive to rec
exp056 0.668 I will reduce the cooldown from 3 to 1 to allow more frequent trading signals
exp057 0.653 I will reduce the rsiPeriod from 14 to 12 to improve RSI responsiveness while
exp058 0.651 I will reduce the basePositionSize from 0.15 to 0.10 to decrease position sizi
exp059 0.557 I will increase the deviationThreshold from 0.025 to 0.030 to reduce false ent
exp060 0.536 I will increase the deviationThreshold from 0.025 to 0.028 to filter out margi
exp061 0.59 I will increase the vwapPeriod from 20 to 25 to use a longer-term VWAP baselin
exp062 0.562 I will decrease the deviationThreshold from 0.025 to 0.020 to tighten entry fi
exp063 0.646 I will reduce the vwapPeriod from 20 to 18 to create a more responsive baselin
exp064 0.632 I will increase the exitThreshold from 0.015 to 0.018 to hold positions even l
exp065 0.714 I will reduce the deviationThreshold from 0.025 to 0.022 to capture earlier me
exp066 0.542 I will tighten the deviationThreshold from 0.022 to 0.021 to capture earlier,
exp067 0.248 I will increase the basePositionSize from 0.15 to 0.18 to allow larger positio
exp068 0.681 I will reduce the deviationThreshold from 0.022 to 0.023 to slightly loosen en
exp069 0.695 Reduce the cooldown from 3 to 2 to allow more frequent trading signals while m
exp070 0.726 Reduce the rsiPeriod from 14 to 10 to make RSI more responsive to recent price
exp071 0.726 Increase the rsiPeriod from 10 to 12 to find the sweet spot between the overly
exp072 0.503 Increase vwapPeriod from 20 to 25 to smooth VWAP calculation and reduce false me
exp073 0.726 Reduce the rsiPeriod from 10 to 8 to accelerate momentum detection on mean-rev
exp074 0.74 Reduce the atrPeriod from 7 to 5 to make volatility-based position sizing more
exp075 0.732 Increase atrPeriod from 5 to 6 to find the optimal volatility responsiveness bet
baseline 0.74 Initial strategy — baseline measurement
exp077 0.653 Increase the deviationThreshold from 0.022 to 0.025 to reduce false mean-rever
exp078 0.727 Reduce the basePositionSize from 0.15 to 0.12 to decrease per-trade risk exposur
exp079 0.644 Reduce the cooldown from 3 to 2 bars to allow more frequent re-entry signals whi
exp080 0.683 Increase the exitThreshold from 0.015 to 0.018 to allow positions to stay open l
exp081 0.652 Reduce the exitThreshold from 0.015 to 0.012 to exit positions earlier when me
exp082 0.74 Reduce the rsiPeriod from 10 to 8 to make RSI respond even faster to recent pric
exp083 0.653 Increase the deviationThreshold from 0.022 to 0.025 to reduce false mean-reversi
exp084 0.693 Reduce the vwapPeriod from 20 to 15 to make VWAP respond faster to recent price
exp085 0.702 Reduce the vwapPeriod from 20 to 18 to balance VWAP responsiveness with stabilit
exp086 0.489 Increase the cooldown from 3 to 5 bars to reduce whipsaw trades and consecutive
exp087 0.652 Reduce the exitThreshold from 0.015 to 0.012 to close positions earlier when mea
exp088 0.696 Reduce the exitThreshold from 0.015 to 0.013 to close positions earlier when mea
exp089 0.653 Increase the deviationThreshold from 0.022 to 0.025 to reduce false mean-reversi
exp090 0.652 Increase the deviationThreshold from 0.022 to 0.024 to filter out marginal mean-
exp091 0.672 Implement adaptive RSI thresholds based on volatility regime (using ATR percenti
exp092 0.489 Increase the cooldown from 3 to 5 bars to reduce whipsaw trades and excessive tu
exp093 0.74 Increase the rsiPeriod from 10 to 12 to reduce noise in RSI signals while mainta
exp094 0.652 Reduce the exitThreshold from 0.015 to 0.012 to trigger exits earlier when price
exp095 0.727 Reduce the basePositionSize from 0.15 to 0.12 to lower per-trade risk exposure w
exp096 0.653 Increase the deviationThreshold from 0.022 to 0.025 to reduce false mean-reversi
exp097 0.513 Increase the vwapPeriod from 20 to 25 to reduce whipsaw trades by using a longer
exp098 ? Implement regime-adaptive volatility scaling by detecting trending vs mean-rever
exp099 0.654 Implement asymmetric RSI thresholds (buyRsiThreshold=35, sellRsiThreshold=65) to
exp100 0.652 Reduce the exitThreshold from 0.015 to 0.012 to lock in profits more aggressivel
exp101 0.672 Reduce the exitThreshold from 0.015 to 0.010 to exit positions more aggressively
exp102 0.653 Increase the deviationThreshold from 0.022 to 0.025 to reduce false mean-reversi
exp103 0.726 Implement dual-timeframe volatility confirmation by using a longer ATR period (7
exp104 0.513 Increase vwapPeriod from 20 to 25 to reduce false mean-reversion signals by usin
exp105 -999 Implement adaptive RSI thresholds based on volatility regime using ATR percentil
exp106 0.574 Increase vwapPeriod from 20 to 22 and reduce deviationThreshold from 0.022 to 0.
baseline 0.157 Initial strategy — baseline measurement
baseline -1.46 Initial strategy — baseline measurement
exp109 -2.18 Reduce cooldown from 3 to 2 bars to allow faster re-entry on new mean-reversion
baseline -1.46 Initial strategy — baseline measurement
exp111 -1.905 Tighten deviationThreshold from 0.022 to 0.020 while relaxing RSI entry from 40/
baseline -1.46 Initial strategy — baseline measurement
exp113 -1.285 Reduce exitThreshold from 0.015 to 0.012 to allow positions to capture more mean
exp114 -2.475 Increase basePositionSize from 0.15 to 0.18 to capture more profit from high-con
exp115 -2.598 Reduce exitThreshold from 0.015 to 0.010 while tightening deviationThreshold fro
baseline -1.46 Initial strategy — baseline measurement
exp117 2.838 Complete strategy redesign: switched from VWAP reversion (overfit to synthetic d
baseline 2.838 Initial strategy — baseline measurement
baseline 2.838 Initial strategy — baseline measurement
baseline 2.838 Initial strategy — baseline measurement
exp120 2.838 Manual mutation — implement via OpenClaw agent
exp121 2.838 Manual mutation — implement via OpenClaw agent
exp122 2.838 Manual mutation — implement via OpenClaw agent
exp123 2.838 Manual mutation — implement via OpenClaw agent
exp124 2.838 Manual mutation — implement via OpenClaw agent
baseline 2.838 Initial strategy — baseline measurement
exp126 2.919 Add regime-based position sizing by increasing exposure in trending regimes (hig
exp127 2.923 Reduce Hurst lookback from 50 to 30 bars to make regime detection more responsiv
exp128 3.668 Reduce ATR trail multiple from 2.0 to 1.5 to exit winners earlier and capture pr
exp129 1.261 Add a volume surge filter to entry signals, requiring current volume to exceed 1
exp130 3.668 Add a volume confirmation filter to exit signals, requiring volume to drop below
exp131 3.668 Add a momentum confirmation filter requiring price to be above the 50-period SMA
exp132 1.048 Replace the Hurst-based regime detection with a simpler trend strength filter us
exp133 3.57 Replace the Donchian breakout entry with a Volume Weighted Average Price (VWAP)
exp134 -999 Replace the Donchian breakout system with a pure EMA crossover trend-following a
exp135 3.262 Replace the complex Hurst regime detection with a simpler volatility-adjusted br
baseline 3.668 Initial strategy — baseline measurement
exp137 3.741 Replace the Hurst exponent regime detection with a simpler trend-following appro
exp138 2.967 Add a volatility regime filter that scales down position sizes by 50% during hig
exp139 3.538 Add a volume momentum filter that requires current volume to be above the 20-per
exp140 3.448 Add a dual-timeframe momentum filter using both short-term (5-period) and medium
exp141 3.739 Add a win-rate optimization mechanism that tracks recent trade outcomes per pair
exp142 3.705 Add a pair correlation filter that only enters new positions in pairs showing po
exp143 -0.421 Add a volume spike filter that only enters breakout trades when current volume e
exp144 2.889 Replace momentum-based position sizing with a simple volatility-normalized syste
exp145 3.476 Add a volume-weighted position sizing mechanism that scales positions by 0.6x to
exp146 3.496 Add a market regime filter that only takes long positions when the 50-period EMA
exp147 2.178 Replace the Donchian breakout entry with a simpler EMA crossover entry combined
exp148 1.717 Add a volume-based entry filter that requires current volume to exceed the 20-pe
exp149 -999 Replace the complex Donchian/ROC momentum system with a simpler high-conviction
exp150 3.655 Add a volatility-adjusted profit target that scales from 1.5x to 3.0x ATR based
exp151 3.777 Add a multi-timeframe trend filter that calculates a longer-term 50-period EMA s
baseline 3.777 Initial strategy — baseline measurement
exp153 3.663 Add a volatility breakout filter that requires current ATR to be above its 20-pe
exp154 -999 Replace the momentum-based position sizing with a regime-adaptive approach that
exp155 1.564 Replace the complex multi-filter entry system with a pure momentum breakout appr
exp156 0.306 Add a volume confirmation filter that requires current volume to be above the 20
exp157 -999 Replace the complex multi-filter entry system with a simpler volume-weighted mom
exp158 2.806 Replace momentum-based position sizing with a drawdown-aware risk reducer that c
exp159 1.457 Replace the multi-filter entry system with a pure volatility breakout strategy t
exp160 2.637 Replace the fixed ATR trailing stop with a dynamic regime-aware stop that tighte
exp161 4.512 STRUCTURAL: Ensemble voting (Donchian+RSI+MACD) + macro 100-EMA trend filter + c
baseline 4.512 Initial strategy — baseline measurement
exp163 5.31 Replace the ensemble voting system with a pure trend-following breakout strategy
baseline 5.31 Initial strategy — baseline measurement
exp165 4.717 Replace the fixed 3.0x ATR profit target with a dynamic adaptive target that sca
exp166 4.755 Replace the fixed 1.5x ATR trailing stop with a momentum-adaptive stop that tigh
exp167 4.114 Replace the single-stage volatility scaling with a multi-regime position sizing
exp168 3.934 Replace the fixed breakout lookback period with a volatility-adaptive lookback t
exp169 4.159 Replace the fixed 50-period trend EMA with a dual-timeframe trend confirmation s
exp170 7.327 Replace the fixed volatility expansion filter (ATR > ATR_SMA) with a percentile-
baseline 7.327 Initial strategy — baseline measurement
exp172 6.763 Replace the fixed 60th percentile ATR threshold with a dynamic adaptive threshol
exp173 4.875 Replace the fixed 3.0x ATR profit target with a momentum-adaptive profit target
exp174 7.01 Replace the fixed 1.5x ATR trailing stop with a momentum-adaptive trailing stop
exp175 6.837 Replace the fixed 0.998 breakout threshold (allowing near-misses) with a strict
exp176 5.372 Replace the single 50-period trend EMA with a dual-timeframe trend filter using
exp177 3.13 Replace the fixed 20-period Donchian breakout with a volume-weighted momentum fi
exp178 6.419 Replace the fixed 60th percentile ATR threshold with a dual-timeframe ATR expans
exp179 -999 Replace the rigid 20-period Donchian breakout with a volume-adjusted momentum en
exp180 7.875 Replace the fixed 3.0x ATR profit target with an adaptive profit target that sca
baseline 7.875 Initial strategy — baseline measurement
exp182 -999 Replace the fixed 20-period Donchian breakout with a volume-weighted price momen
exp183 7.991 Replace the fixed 20-period Donchian breakout with a dynamic breakout lookback t
exp184 7.082 Replace the trailing stop mechanism with a volatility-adjusted Chandelier Exit t
exp185 0.248 Replace the adaptive Donchian breakout with a dual-timeframe momentum confirmati
exp186 5.436 Replace the adaptive Donchian breakout with a strict triple-confirmation entry r
exp187 -0.022 Replace the adaptive Donchian breakout with a momentum-quality filter requiring
exp188 4.585 Replace the adaptive Donchian breakout with a momentum acceleration filter requi
exp189 1.439 Replace the percentile-based ATR expansion filter with a dual-condition volatili
baseline 7.991 Initial strategy — baseline measurement
exp191 5.389 Implement a regime-switching framework that runs the current pure-trend strategy
exp192 3.126 Implement a multi-timeframe momentum confirmation system that synthesizes fast (
exp193 7.991 Implement a multi-pair correlation divergence overlay that detects when correlat
exp194 1.649 Implement a triple-timeframe momentum cascade system that synthesizes fast (5-ba
exp195 -999 Implement a multi-stage exit cascade system that exits positions in thirds at pr
exp196 7.092 Implement a dual-strategy ensemble system that runs both the existing pure-trend
baseline 7.991 Initial strategy — baseline measurement
exp198 7.991 Implement a multi-pair correlation divergence detection system that identifies w
exp199 8.176 Implement a dual-strategy portfolio allocation system that runs a pure breakout
exp200 4.828 Implement a volatility-regime adaptive exit system that dynamically switches bet
exp201 8.176 Manual mutation — implement via OpenClaw agent
baseline 8.176 Initial strategy — baseline measurement
exp203 7.993 Implement a volatility-breakout filter that only enters mean-reversion shorts wh
exp204 7.668 Implement a triple-timeframe momentum confirmation system that synthesizes 4-bar
baseline 8.176 Initial strategy — baseline measurement
exp206 -3.345 Implement a multi-timeframe confluence system that synthesizes 6-hour, 12-hour,
baseline 8.176 Initial strategy — baseline measurement
baseline 8.176 Initial strategy — baseline measurement
exp209 5.269 Implement a cross-pair correlation-based entry filter that only triggers breakou
exp210 7.843 Implement a volatility-regime-adaptive position scaling system that dynamically
baseline 8.176 Initial strategy — baseline measurement
baseline 8.176 Initial strategy — baseline measurement
exp213 7.137 Implement a cross-pair momentum divergence system that detects when a pair shows
exp214 1.09 Implement a multi-pair relative strength ranking system that only enters breakou
exp215 -999 Implement a volatility-adaptive stop-loss layering system that uses multiple ATR
exp216 5.575 Implement a volatility-adaptive multi-timeframe momentum filter that synthesizes
baseline 8.176 Initial strategy — baseline measurement
baseline 8.176 Initial strategy — baseline measurement
exp219 4.457 Implement a cross-pair volume-weighted momentum scoring system that ranks pairs
exp220 -999 Implement a position-sizing pyramid system that adds to winning breakout trades
exp221 -999 Implement a volatility-adjusted partial profit-taking system that scales out 50%
baseline 8.176 Initial strategy — baseline measurement
exp223 8.15 Implement a cross-pair correlation-based signal generator that detects divergenc
baseline 8.176 Initial strategy — baseline measurement
baseline 8.176 Initial strategy — baseline measurement
exp226 7.758 Implement an adaptive volatility regime filter that uses dual ATR measurements (
baseline 8.176 Initial strategy — baseline measurement
baseline 8.176 Initial strategy — baseline measurement
exp229 6.262 Implement a cross-timeframe momentum confirmation system that synthesizes 4-hour
baseline 8.176 Initial strategy — baseline measurement
exp231 -999 Implement a cascading multi-level exit system for breakout positions that scales
exp232 -999 Implement a volatility-adaptive cascading exit system that scales out of breakou
exp233 5.392 Implement an ensemble voting system where breakout and mean-reversion strategies
baseline 8.176 Initial strategy — baseline measurement
exp235 6.019 Implement a three-regime adaptive strategy that detects market regimes using Hur
exp236 6.576 Implement a volatility-adjusted position sizing system combined with a three-reg