March 2026 — Strategy Metrics¶

Data snapshot: 2026-03-29 22:02 UTC

Active Strategies (v10)¶

Strategy	P&L	Trades	Volume	Win Rate	Margin (bps)	Sharpe
StatsArb_v8_ETH_BTC	`+$1.02`	38	`$1,832`	42%	5.6	0.07
TrendFollowing_v10_ETH	`-$0.14`	8	`$400`	50%	-3.5	-0.02
TrendFollowing_v10_BTC	`-$0.30`	6	`$299`	50%	-9.9	-0.08
MeanReversion_v10_ETH	`+$1.04`	12	`$600`	75%	17.4	0.33
MeanReversion_v10_BTC	`+$1.24`	9	`$450`	89%	27.7	0.61

Active total: +$2.87

March-Created Strategies (all versions)¶

Incident P&L (Mar 22-23 OKX Fill Pipeline Failure)¶

Strategy	Status	P&L	Trades	Win Rate	Sharpe
TrendFollowing_v8_ETH	ERROR	`+$68.27`	4	50%	0.22
MeanReversion_v8_ETH	KILLED	`+$45.06`	4	100%	0.94

These two account for +$113 — this is incident luck, not strategy edge. OKX WS failed to deliver fills, strategies accumulated $3,170 in unintended positions, and ETH rallied during the 20-hour outage. A price move the other way = -$3,000+. Must be excluded from strategy P&L attribution.

Market Making (new strategy type)¶

Strategy	Status	P&L	Trades	Win Rate	Margin (bps)	Sharpe
MarketMaking_v1_ETH	ERROR	`-$0.76`	12	67%	-5.1	0.32
MarketMaking_v1_BTC	ERROR	`-$2.20`	2	50%	-44.0	-0.59
MarketMaking_v2_ETH	ERROR	`-$0.87`	8	25%	-10.8	-0.52
MarketMaking_v2_BTC	ERROR	`+$0.04`	2	50%	2.2	0.05

MarketMaking total: -$3.78. High trade count but margins eaten by fees. Both versions hit ERROR status.

Disabled v9 Strategies¶

Strategy	P&L	Trades	Win Rate
StatsArb_v7_ETH_BTC	`+$0.16`	1	100%
TrendFollowing_v9_ETH	`-$0.57`	1	0%
TrendFollowing_v9_BTC	`-$0.96`	1	0%
MeanReversion_v9_ETH	`-$3.74`	2	50%
MeanReversion_v9_BTC	`$0.00`	0	—

Earlier March Versions (v3–v8, disabled)¶

Strategy	P&L	Trades	Win Rate
StatsArb_v6_ETH_BTC	`-$0.86`	3	33%
TrendFollowing_v8_BTC	`-$0.73`	1	0%
StatsArb_v5_ETH_BTC	`-$2.52`	6	50%
SMATrendFollowing_v6_ETH	`-$1.01`	31	68%
SMATrendFollowing_v6_BTC	`-$0.71`	16	56%
SMAMeanReversion_v6_ETH	`-$1.31`	16	25%
SMAMeanReversion_v6_BTC	`-$0.62`	12	33%
SMATrendFollowing_v5_ETH	`-$4.40`	23	43%
SMATrendFollowing_v5_BTC	`-$0.90`	18	61%
SMATrendFollowing_v4_ETH	`-$1.16`	5	40%
SMATrendFollowing_v4_BTC	`-$1.58`	8	38%
SMATrendFollowing_v3_ETH	`-$0.53`	3	67%
SMATrendFollowing_v3_BTC	`-$1.41`	5	20%
SMAMeanReversion_v5_ETH	`-$0.07`	3	33%
SMAMeanReversion_v5_BTC	`-$0.33`	1	0%
StatsArb_v4_ETH_BTC	`$0.00`	0	—
SMAMeanReversion_v4_ETH	`-$1.53`	6	0%
SMAMeanReversion_v4_BTC	`-$1.48`	9	44%

Cumulative Summary¶

Category	P&L	Strategies
Active (v10 + v8 StatsArb)	`+$2.87`	5
Incident luck (Mar 22-23)	`+$113.33`	2
March-created disabled	`-$21.20`	20
Pre-March (Jan–Feb)	`-$52.57`	12
All-time total	`+$42.43`	39
Excluding incident	`~-$71`

Paper Strategies (stag-paper, as of Mar 29)¶

Strategy	P&L	Trades	Volume	Win Rate	Sharpe
StatsArb_v8_ETH_BTC	`-$4.00`	53	`$2,648`	32%	-0.22
TrendFollowing_v10_ETH	`+$0.17`	14	`$699`	50%	0.03
TrendFollowing_v10_BTC	`-$1.65`	9	`$450`	33%	-0.31
MeanReversion_v10_ETH	`-$0.68`	21	`$1,049`	38%	-0.15
MeanReversion_v10_BTC	`-$0.84`	19	`$949`	21%	-0.33
VolScaledDir_v1_mean_rev_ml_vol	`$0.00`	0	`$0`	—	—
VolScaledDir_v1_mean_rev	`$0.00`	0	`$0`	—	—
VolScaledDir_v1_ml_vol	`-$2.53`	25	`$2,941`	32%	-0.51
VolScaledDir_v1_BTC	`-$1.07`	23	`$2,791`	30%	-0.49

Paper total: -$10.60

Paper vs Live Divergence¶

Strategy	Live P&L	Paper P&L	Direction Match?
StatsArb_v8	`+$1.02`	`-$4.00`	No
TrendFollowing_v10_ETH	`-$0.14`	`+$0.17`	No
TrendFollowing_v10_BTC	`-$0.30`	`-$1.65`	Yes (both losing)
MeanReversion_v10_ETH	`+$1.04`	`-$0.68`	No
MeanReversion_v10_BTC	`+$1.24`	`-$0.84`	No

Live and paper disagree on direction for 4 of 5 strategies. MeanReversion winning live but losing paper. Consistent with Feb finding: paper vs live fidelity is strategy-dependent and unreliable.

ML Strategy Integration (Paper Only)¶

4 VolScaledDirectionalStrategy variants deployed to paper (created Mar 18). Two variants have zero trades (mean_rev, mean_rev_ml_vol). Two active variants losing (-$3.60 combined). Not deployed to live — need parameter tuning before promotion.

Strategy Type Evolution¶

Type	Feb Result	Mar Result	Trend
TrendFollowing (SMA→renamed)	`+$3.31` best performer	`-$0.44` active, `+$68` anomaly	Regressed in clean runs
MeanReversion (SMA→renamed)	Always losing	`+$2.29` active, best Sharpe	Flipped to profitable
StatsArb	Killed by incident	`+$1.02` active, modest	Stabilized
MarketMaking	N/A (new)	`-$3.78`	Needs work
CrossVenueArb	Near-zero	Not redeployed	Parked
Classification	50% accuracy	Not redeployed	Parked