Canonical Strategy Performance Report

Generated 2026-06-03 from a static snapshot of dev views · Source: trsports_tradingjournal_dev · Views: main.v_perf_by_canonical_slug, main.v_perf_by_canonical_slug_weekly, main.v_label_to_canonical_mapping · Sample: 310 trades through 2026-06-03 · 0 NULL canonical_strategy_slug · Snapshot — re-query the views for live numbers.

1. Executive summary

2. Canonical strategy performance — all-time

Slug Family n Total P&L Avg Median Min Max W / L / S Win %
vwap-pull-in -> trend edge vwap_pullback_continuation 15 +$367.67 +$24.51 +$12.25 -$11.76 +$136.29 14 / 1 / 0 93.3%
chase-extended sentinel off-book 56 +$200.12 +$3.57 +$3.50 -$100.51 +$199.07 32 / 24 / 0 57.1%
unclassified sentinel off-book 99 +$35.79 +$0.36 +$0.71 -$115.50 +$124.36 52 / 47 / 0 52.5%
backside-long counter_trend_scalp 13 -$74.56 -$5.74 -$3.65 -$62.55 +$52.56 6 / 7 / 0 46.2%
vwap-pull-in leak vwap_pullback_continuation 127 -$364.47 -$2.87 -$2.25 -$216.15 +$150.76 45 / 80 / 2 35.4%

Family rollup

FamilynTotal P&LWLWin %
off-book155+$235.91847154.2%
vwap_pullback_continuation142+$3.20598141.5%
counter_trend_scalp13-$74.566746.2%
vwap_pullback_continuation is essentially flat (+$3.20) — masks the +$367 / -$364 split between the two Pull-In variants. The conversion-to-trend management is the entire edge in this family.

3. Weekly performance by canonical strategy

Week startSlugnTotal P&LAvgWin %
Week of 2026-06-01
chase-extended2+$165.27+$82.6450.0%
vwap-pull-in -> trend2+$16.30+$8.15100.0%
backside-long4+$14.68+$3.6775.0%
unclassified5-$43.72-$8.740.0%
vwap-pull-in14-$248.86-$17.7828.6%
Week of 2026-05-25
unclassified13+$169.21+$13.0261.5%
vwap-pull-in -> trend7+$116.13+$16.5985.7%
backside-long1-$50.37-$50.370.0%
chase-extended9-$69.12-$7.6855.6%
vwap-pull-in21-$159.78-$7.614.8%
Week of 2026-05-18
chase-extended5-$38.80-$7.7640.0%
vwap-pull-in5-$39.56-$7.9160.0%
unclassified1-$26.57-$26.570.0%
Week of 2026-05-11
chase-extended8+$156.23+$19.5362.5%
unclassified9-$13.29-$1.4833.3%
vwap-pull-in24-$32.35-$1.3554.2%
backside-long3-$47.83-$15.9433.3%
Week of 2026-05-04
vwap-pull-in21+$9.35+$0.4533.3%
chase-extended15-$1.13-$0.0853.3%
unclassified24-$61.59-$2.5754.2%
Week of 2026-04-27
vwap-pull-in13+$157.46+$12.1153.8%
unclassified14+$99.36+$7.1078.6%
vwap-pull-in -> trend1+$8.58+$8.58100.0%
backside-long1-$11.60-$11.600.0%
chase-extended3-$38.08-$12.6966.7%
Week of 2026-04-20
chase-extended3+$46.19+$15.40100.0%
backside-long3+$33.61+$11.2066.7%
vwap-pull-in16-$58.07-$3.6318.8%
unclassified16-$148.61-$9.2937.5%
Week of 2026-04-13
vwap-pull-in -> trend4+$90.37+$22.59100.0%
unclassified15+$74.49+$4.9766.7%
chase-extended9+$21.33+$2.3755.6%
backside-long1-$13.05-$13.050.0%
vwap-pull-in11-$19.81-$1.8054.5%
Week of 2026-04-06
vwap-pull-in -> trend1+$136.29+$136.29100.0%
vwap-pull-in2+$27.15+$13.5850.0%
unclassified2-$13.49-$6.7550.0%
chase-extended2-$41.77-$20.8950.0%
Within most weeks, the same playbook prints both wins and losses — selection within the day is the discriminator, not the week itself. The 2026-05-25 and 2026-06-01 weeks are the worst plain-Pull-In stretches (-$408 combined) and contain the heaviest hot-tape concentration.

4. Display label → canonical slug mapping

Canonical slugDisplay label (raw)nTotal P&L
backside-longBackside Long13-$74.56
chase-extendedChase / Extended56+$200.12
unclassifiedUnclassified99+$35.79
vwap-pull-inVWAP Pull-In127-$364.47
vwap-pull-in -> trendVWAP Pull-In -> Trend15+$367.67
Clean 1:1 mapping — no display label maps to multiple slugs, no orphans. Post-Phase-2 backfill, every historical row resolves through either alias-match (Step A) or display-name match (Step C); Step D fallback never fired (0 rows).

5. VWAP Pull-In split — the headline finding

EDGE · vwap-pull-in -> trend

15 trades, +$367.67 total, +$24.51/trade, 93.3% win rate.

The highest-expectancy slug in the dataset. This isn't a separate entry technique — it's the management decision to hold a Pull-In through the +2 SD breakout instead of scratching at the original target. Today this conversion fires only 15 times in 142 Pull-In family attempts. The edge lives in the conversion.

LEAK · vwap-pull-in

127 trades, -$364.47 total, -$2.87/trade, 35.4% win rate.

Erases everything Pull-In → Trend earns and then some. The discriminator is RVOL + gap, not the entry technique:

Symbol concentration: IBIT alone is -$185 of -$364 total (51%). Plus AMD (-$70), ETHU (-$69), RGTI (-$36) = -$360 combined. Four symbols carry the leak.

Time-of-day concentration: 09–10 ET holds 91 of 127 trades and -$401 of P&L. Same hour, Pull-In → Trend was +$253 at 100% win. Selection within the hour, not the hour itself.

Plain VWAP Pull-In is only a clean setup in calm, controlled tape; in hot tape, it must either be managed for trend conversion or be passed.

6. Next actions

  1. Use the Quality Gate tomorrow. Open the cheatsheet (link below) pre-market. Paste the MGP block into the Morning edition. Print the intraday checklist; keep it next to the chart.
  2. Sample-review the 56 Chase / Extended trades over the coming week. Today they're net +$200 at 57% WR — that contradicts the classifier rule's "behavioral mistake" framing. Decide whether (a) the rule should be retuned, (b) some of these belong in fl2-fash-late, or (c) they stay as the off-book sentinel they are.
  3. Sample-review the 99 Unclassified trades separately. Net flat at 52.5% WR. Some likely belong in canonical slugs the classifier missed; some are legitimately off-book.
  4. Phase 4 (VCET-at-entry validation) is the next research step, but not today unless you approve. It would test whether VCET states empirically separate Pull-In → Trend winners (expected: EARLY_EXPANSION → EXPANSION_CONFIRMED → TREND_UP) from plain Pull-In losers (expected: ROTATION / FAILED_EXPANSION).
  5. Phase 3b (full Scorecards HTML refresh) remains on hold. The current Scorecards still group by raw display label; switching them to canonical_strategy_slug requires touching gen_scorecard.py's HTML render path and will trigger the next EOD to regenerate every scorecard. Hold until you want that.

7. Workflow artifacts

Live data — re-query at any time

The numbers in this report are a static snapshot taken 2026-06-03. The underlying views are queryable any time on dev:
SELECT * FROM main.v_perf_by_canonical_slug ORDER BY total_pnl DESC;
SELECT * FROM main.v_perf_by_canonical_slug_weekly ORDER BY week_start DESC;
SELECT * FROM main.v_label_to_canonical_mapping;
Static report · no DB schema changes · no VCET touches · no taxonomy changes · no gen_scorecard.py edits · Phase board unchanged: Phase 3b and Phase 4 remain ON HOLD pending your call.