QLV90802 EEMV holdout 10-session half-profit base
BacktestsrejectQuant replicationEEMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 3.92e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.469122,0.117191]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-6.33
drawdown10.53
win rate55.56
sample36
average win0.57
average loss1.11
mean-0.18
mean CI low-0.47
mean CI high0.12
quarter Kelly0
slippage bp5
actual size0
QLV90802 EEMV train 10-session half-profit base
BacktestsrejectQuant replicationEEMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 3.55e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.566171,0.097840]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-7.38
drawdown12.63
win rate53.13
sample32
average win0.59
average loss1.17
mean-0.23
mean CI low-0.57
mean CI high0.1
quarter Kelly0
slippage bp5
actual size0
QLV90802 EEMV recent 10-session fixed-exit base
BacktestsrejectQuant replicationEEMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 6.53e-14 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-1.591759,1.210684]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-1.12
drawdown5.59
win rate57.14
sample7
average win1.42
average loss2.23
mean-0.14
mean CI low-1.59
mean CI high1.21
quarter Kelly0
slippage bp5
actual size0
QLV90802 EEMV holdout 10-session fixed-exit base
BacktestsrejectQuant replicationEEMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 2.43e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.581388,0.097071]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-8.66
drawdown12.39
win rate38.89
sample36
average win0.98
average loss1.02
mean-0.25
mean CI low-0.58
mean CI high0.1
quarter Kelly0
slippage bp5
actual size0
QLV90802 EEMV train 10-session fixed-exit base
BacktestsrejectQuant replicationEEMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 3.27e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.647948,0.107154]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-8.53
drawdown14.15
win rate40.63
sample32
average win0.97
average loss1.12
mean-0.27
mean CI low-0.65
mean CI high0.11
quarter Kelly0
slippage bp5
actual size0
QLV90802 EFAV holdout 10-session half-profit stress
BacktestsrejectQuant replicationEFAV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+20bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 4.69e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.849032,-0.327937]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-20.8
drawdown20.45
win rate33.33
sample39
average win0.37
average loss1.07
mean-0.59
mean CI low-0.85
mean CI high-0.33
quarter Kelly0
slippage bp20
actual size0
QLV90802 EFAV holdout 10-session fixed-exit stress
BacktestsrejectQuant replicationEFAV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+20bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 3.13e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.851523,-0.199385]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-18.78
drawdown19.18
win rate41.03
sample39
average win0.67
average loss1.36
mean-0.53
mean CI low-0.85
mean CI high-0.2
quarter Kelly0
slippage bp20
actual size0
QLV90802 EFAV recent 10-session half-profit base
BacktestsrejectQuant replicationEFAV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 2.22e-15 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-1.705693,0.823750]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-2.19
drawdown4.08
win rate50
sample6
average win1.03
average loss1.74
mean-0.36
mean CI low-1.71
mean CI high0.82
quarter Kelly0
slippage bp5
actual size0
QLV90802 EFAV holdout 10-session half-profit base
BacktestsrejectQuant replicationEFAV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 4.74e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.554674,-0.034357]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-10.97
drawdown11.38
win rate48.72
sample39
average win0.53
average loss1.08
mean-0.29
mean CI low-0.55
mean CI high-0.03
quarter Kelly0
slippage bp5
actual size0
QLV90802 EFAV train 10-session half-profit base
BacktestsrejectQuant replicationEFAV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 3.41e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.450050,0.041509]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-8.26
drawdown13.66
win rate43.9
sample41
average win0.61
average loss0.84
mean-0.21
mean CI low-0.45
mean CI high0.04
quarter Kelly0
slippage bp5
actual size0
QLV90802 EFAV recent 10-session fixed-exit base
BacktestsrejectQuant replicationEFAV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 8.44e-15 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-1.656603,1.016394]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-1.77
drawdown3.66
win rate50
sample6
average win1.18
average loss1.74
mean-0.28
mean CI low-1.66
mean CI high1.02
quarter Kelly0
slippage bp5
actual size0
QLV90802 EFAV holdout 10-session fixed-exit base
BacktestsrejectQuant replicationEFAV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 4.83e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.551373,0.100114]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-8.7
drawdown10.67
win rate41.03
sample39
average win0.97
average loss1.06
mean-0.23
mean CI low-0.55
mean CI high0.1
quarter Kelly0
slippage bp5
actual size0
QLV90802 EFAV train 10-session fixed-exit base
BacktestsrejectQuant replicationEFAV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 3.84e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.507839,0.083461]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-8.6
drawdown14.53
win rate36.59
sample41
average win0.97
average loss0.9
mean-0.21
mean CI low-0.51
mean CI high0.08
quarter Kelly0
slippage bp5
actual size0
QLV90802 USMV holdout 10-session half-profit stress
BacktestsrejectQuant replicationUSMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+20bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 4.19e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.895933,-0.350419]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-24.22
drawdown24.78
win rate27.27
sample44
average win0.57
average loss1.07
mean-0.62
mean CI low-0.9
mean CI high-0.35
quarter Kelly0
slippage bp20
actual size0
QLV90802 USMV holdout 10-session fixed-exit stress
BacktestsrejectQuant replicationUSMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+20bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 2.74e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.810655,-0.145435]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-19.42
drawdown20.23
win rate40.91
sample44
average win0.79
average loss1.37
mean-0.48
mean CI low-0.81
mean CI high-0.15
quarter Kelly0
slippage bp20
actual size0
QLV90802 USMV recent 10-session half-profit base
BacktestsrejectQuant replicationUSMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 7.77e-14 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.834121,0.226791]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-2.35
drawdown3.24
win rate42.86
sample7
average win0.46
average loss0.94
mean-0.34
mean CI low-0.83
mean CI high0.23
quarter Kelly0
slippage bp5
actual size0
QLV90802 USMV holdout 10-session half-profit base
BacktestsrejectQuant replicationUSMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 3.27e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.598364,-0.048365]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-13.52
drawdown16.09
win rate45.45
sample44
average win0.59
average loss1.09
mean-0.33
mean CI low-0.6
mean CI high-0.05
quarter Kelly0
slippage bp5
actual size0
QLV90802 USMV train 10-session half-profit base
BacktestsrejectQuant replicationUSMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 3.27e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.384520,0.171548]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-5.01
drawdown9.88
win rate48.89
sample45
average win0.73
average loss0.92
mean-0.11
mean CI low-0.38
mean CI high0.17
quarter Kelly0
slippage bp5
actual size0
QLV90802 USMV recent 10-session fixed-exit base
BacktestsrejectQuant replicationUSMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 1.15e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.717164,0.723895]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-0.53
drawdown3.15
win rate42.86
sample7
average win1.08
average loss0.94
mean-0.07
mean CI low-0.72
mean CI high0.72
quarter Kelly0
slippage bp5
actual size0
QLV90802 USMV holdout 10-session fixed-exit base
BacktestsrejectQuant replicationUSMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 4.16e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.510007,0.151176]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-8.05
drawdown13.22
win rate40.91
sample44
average win1.1
average loss1.07
mean-0.18
mean CI low-0.51
mean CI high0.15
quarter Kelly0
slippage bp5
actual size0
QLV90802 USMV train 10-session fixed-exit base
BacktestsrejectQuant replicationUSMV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 4.93e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.366656,0.239295]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-3.14
drawdown9.34
win rate44.44
sample45
average win0.99
average loss0.91
mean-0.07
mean CI low-0.37
mean CI high0.24
quarter Kelly0
slippage bp5
actual size0
QLV90802 SPLV holdout 10-session half-profit stress
BacktestsrejectQuant replicationSPLV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+20bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 2.91e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.925563,-0.343834]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-23.56
drawdown24.01
win rate23.81
sample42
average win0.58
average loss1.01
mean-0.63
mean CI low-0.93
mean CI high-0.34
quarter Kelly0
slippage bp20
actual size0
QLV90802 SPLV holdout 10-session fixed-exit stress
BacktestsrejectQuant replicationSPLV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. Unmanaged parent of swing exit comparison; no partial profit. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+20bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 4.37e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-1.037164,-0.383701]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-26.32
drawdown27.5
win rate33.33
sample42
average win0.76
average loss1.46
mean-0.72
mean CI low-1.04
mean CI high-0.38
quarter Kelly0
slippage bp20
actual size0
QLV90802 SPLV recent 10-session half-profit base
BacktestsrejectQuant replicationSPLV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 8.70e-14 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-1.050851,0.827689]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-0.8
drawdown2.58
win rate50
sample6
average win1.11
average loss1.36
mean-0.13
mean CI low-1.05
mean CI high0.83
quarter Kelly0
slippage bp5
actual size0
QLV90802 SPLV holdout 10-session half-profit base
BacktestsrejectQuant replicationSPLV2026-09-08 02:14
Frozen note877. 2016-20 training (2013-15 unavailable),2021-25 new-rule holdout but historically inspected market period,2026 completed cohorts only. Monthly firstsessionopen following priormonthclose>100dayMA; previousbar ATR14 stop1ATR,target1.5ATR,timeexit10thsessionclose. At+1ATR sellhalf thenstopbreakeven; adverse samebar ordering. Gapstopopen,stopfirst if ambiguous,target capped. Alladjusted SIP daily;5bpfee+5bpslippage eachfill. Fractional continuous sizing diagnostic,100initial cash,oneposition, no leverage. Vectorbt exactfill reconstruction agrees independentcompounding within 5.13e-13 percentagepoints. Drawdown uses synthetic perdayfill/close slots,not actual15minpath; positions normalizedfullallocation for comparison,not proposal. IIDtrade bootstrap10000seed90802,95% mean interval [-0.628014,-0.042305]%. Historical p/W/L sizing only; failuresin training/holdout keepactual0 evenpositive2026counterexample. Source https://runway.adityaperswal.workers.dev/data/stock-bars ; authorpaper https://pages.stern.nyu.edu/~afrazzin/pdf/Betting%20Against%20Beta%20-%20Frazzini%20and%20Pedersen.pdf . Not exactleveragedBAB. Full252metricsobservationsQLV90802metrics;deterministicstudy.py/ledgers local.
return-13.3
drawdown15.22
win rate50
sample42
average win0.53
average loss1.2
mean-0.33
mean CI low-0.63
mean CI high-0.04
quarter Kelly0
slippage bp5
actual size0