GAMBLINGISALLYOURNEED — Comprehensive Quantitative Report

Window: 2026-04-01 → 2026-04-22 (22 calendar days, 22 active) Source CSV: GAMBLINGISALLYOURNEED_trades.csv Methodology: Cash-flow P/L = -buy_usdc + sell_usdc + remaining_share_payout. Remaining shares settle at $1 if outcome won, $0 if lost, or mark at last-traded price if open. No tx-hash dedup (atomic multi-leg fills are real trades).

Phase 1 — Trader Profile

Scale

Trade-size distribution (USDC per fill)

Inter-trade gap, same (market, outcome)

Phase 2 & 3 — Both-Sides Participation, Dominance Curve

• Both-sides rate: 60.43% (6,396 of 10,585 markets)
• Median paired cost: $0.9681
• Mean paired cost: $0.9654
• Paired cost % under $1.00: 62.8%
• Paired cost % under $0.97: 50.9%
• Paired cost % under $0.95: 43.5%
• Median 2nd-side hedge lag: 2224s (37.1 min)
• % of pairs hedged within 60s: 9.2%
• % of pairs hedged within 1hr: 57.1%

Dominance buckets (5-tier)

Phase 4 — Entry-Price Analysis (cash-flow P/L)

Phase 5 — Category & Vertical Breakdown

Phase 6 — Timing & Execution

Net P/L by hour (UTC)

Net P/L by day-of-week

Phase 7 — Filter Experiments

Phase 8 — Rolling Window Consistency

• Rolling 7-day windows green: 8 of 22 (36.4%)
• Rolling 7-day P/L range: -$283,541 → +$488,723
• Rolling 15-day windows green: 11 of 22 (50.0%)
• Rolling 15-day P/L range: -$295,585 → +$382,062

Daily P/L (BUY trades, cash-flow allocated)

Phase 9 — P/L Decomposition

Theoretical structural attribution

Phase 10 — Top Markets

Top 25 by BUY notional

Top 15 winners by P/L

Top 15 losers by P/L

Report generated 2026-04-24 23:49 from GAMBLINGISALLYOURNEED_trades.csv.

GAMBLINGISALLYOURNEED — Reverse-Engineering Report

Wallet: 0x507e52ef684ca2dd91f90a9d26d149dd3288beae Window: 2026-04-01 → 2026-04-22 (22 days, 22 active) Universe: 344,115 trades · 10,585 markets · 3,984 events · $31.86M gross notional

P/L methodology: Cash-flow accounting. Each position's P/L = -buy_usdc + sell_usdc + remaining_share_payout, where remaining shares are settled at $1.00 if the outcome won, $0.00 if it lost, or marked at last-traded price if the market is still open. This is the correct methodology for any trader who actively exits — ignoring SELL proceeds, as earlier reports did, systematically understates P/L for active managers.

Phase 1 — Trader Profile

Scale & Activity

• 279,014 BUYs + 65,101 SELLs (18.9% sell rate — the first meaningfully active exit-manager we've profiled; every prior bot in this library was 0% SELL)
• BUY notional $24.30M, SELL notional $7.56M, gross turnover $31.86M
• ~15,642 trades/day; active every calendar day in the window
• 10,585 distinct condition IDs across 3,984 events → ~482 markets/day touched

Trade Size Distribution (power-law)

Bimodal sizing: thousands of sub-$10 probes alongside a concentrated upper tail carrying two-thirds of deployed capital. Consistent with a sizer that scales into a line after the first-fill probe succeeds.

Execution Signature

• Median inter-fill gap on same (market, outcome): 4s
• 57% of consecutive fills under 10s · 73% under 60s · 97% under 1hr
• Sub-human burst latency + longer accumulation tails → automated execution with deliberate exit logic (consistent with 65K SELL tickets).

Archetype

Hybrid: pseudo-MM (60.4% both-sides rate) with an active exit engine — not the hold-to-resolution pattern we see on pure bots. The active-management profile is the defining feature of this wallet.

Phase 2 — Core Strategy Identification

• Both-sides rate: 60.4% of 10,585 markets — at the classical MM threshold (>60%).
• Median paired cost $0.9681 · Mean $0.9654 — below fair, so paired legs are positive-EV in isolation.
• 62.8% of paired markets close sub-$1.00; 50.9% close sub-$0.97.
• 18.9% SELL rate is the differentiator: this bot unwinds roughly one in five positions rather than letting them ride. The exit engine realizes intermediate P/L that the simpler wallets in this library forgo.

Classification: Actively managed pseudo-MM with a directional overlay — overall mildly profitable, with real spread capture funding most of the book.

Phase 3 — Dominance Ratio Analysis

The conviction curve is flat. Compare to VidarX (58 → 84%) or Bonereaper (59 → 98%). Here it goes 49.9 → 55.3 → 56.6 → 62.3%. The fair-value engine is weak: when the bot tilts heavily (3×+), the dominant side still only wins 62% of the time — enough for a small edge after spread, but nowhere near "high conviction" territory.

Phase 4 — Entry Price Analysis (cash-flow P/L)

Three distinct edge zones: 1. Deep underdog probes ($0.00–$0.20) — 33,801 small tickets on sub-$0.20 entries that the bot then often sells for rebound pops, earning +$253K combined. This is classic long-tail speculation with exit discipline. 2. Slight favorites ($0.60–$0.70) — the workhorse band, +$175K on meaningful volume. 3. The $0.20–$0.60 middle and the $0.80+ favorite-shave zones both lose money — totalling -$230K across ~156K trades.

The $0.80+ losses are structurally troubling: buying near-certain outcomes and getting them wrong 15%+ of the time at prices that don't pay enough to cover.

Phase 5 — Category & Market-Type Breakdown

Soccer + MLB + NFL + CBB = +$570K combined. Those four verticals more than offset the +$100K headline. The drag comes from Tennis (volume monster at -$150K), NBA (structural disaster at -11.5% ROI), and the IPL-heavy "Other" bucket.

Top-loss single markets: - 1. FC Union Berlin 4/18: -$98K · Union Berlin 4/5: -$52K (recurring trap) - Le Havre AC 4/18: -$73K - Real Madrid 4/15: -$52K · Barcelona 4/4: -$51K

Top-win single markets: - Elche CF 4/22: +$93K - Braves vs Diamondbacks: +$78K - IPL KKR vs Lucknow: +$53K - Toronto Blue Jays vs White Sox: +$46K - Olympique Lyonnais 4/5: +$44K

Phase 6 — Timing & Execution

Peak / Weak Hours (UTC, by net cash-flow P/L)

• Best hours: 09:00 (+$126K), 19:00 (+$100K), 20:00 (+$82K), 15:00 (+$76K), 21:00 (+$76K)
• Worst hours: 13:00 (−$297K), 10:00 (−$81K), 12:00 (−$71K), 18:00 (−$66K)

Hour 13:00 UTC alone destroys three times the book's net profit — more loss than any single category. Combined with 18:00, these two hours strip $363K. Trading everything except those two hours would have closed around +$465K rather than +$101K.

Second-side hedge lag

Median 2,238s (~37 minutes) — deliberate pairing, but not atomic. The bot fires the directional leg quickly and only hedges opportunistically when the book is agreeable. 60.4% both-sides rate reflects that selective hedging — not every market gets paired.

Phase 7 — Filter Experiments (cash-flow P/L)

The high-conviction filter is the real edge extractor — stripping to the dominant leg on markets with ≥2× USDC imbalance and ignoring the hedge leg turns the book from a thin +$94K grind into a clean +$651K / +4.55% ROI / 57.6% WR outcome. That is the replicable structure.

The hour filter alone adds +$132K. Stacked with high-conviction, the practical target is +$750K–$800K P/L versus the +$94K unfiltered baseline.

Phase 8 — Rolling Window Analysis

Weekly cadence: - W14 (4/1–4/5): −$140,144 (rough start — Union Berlin 4/5 loss alone was -$52K) - W15 (4/6–4/12): +$188,868 (recovery) - W16 (4/13–4/19): −$94,209 (Union Berlin 4/18 + Le Havre 4/18 combined for -$171K on a single day) - W17 (4/20–4/22): +$139,406 (strong close: Elche CF +$93K, Atlético Madrid, Diamondbacks)

Cumulative: −$140K → +$49K → −$45K → +$94K. The book swung through ±$200K intra-week; the net positive outcome depends on the W15 recovery and the W17 close. If either had gone sideways, the period would have printed flat or negative.

Consistency verdict: moderate. The rolling 7-day window spends roughly half the period in the red. This is not a monotonically green book — it is a volatile, actively-managed structure that depends on the exit engine converting some losers into smaller losses, and a few outperforming days (most notably 4/22) to clear the book into positive territory.

Phase 9 — P/L Decomposition

The spread engine is genuinely capturing ~$494K of paired-cost edge. The directional overlay and the hedge-tax drag on the losing side consume most of it, leaving the +$101K net. The active-management SELL engine is what rescues the book — without the 65K sells, the same position set held-to-resolution would have closed significantly worse.

Phase 10 — Strategy Specification (short form; full detail in playbook.md)

One-sentence summary: An actively-managed pseudo-MM that trades global team sports with genuine sub-$1 paired-cost discipline, a weak-but-positive fair-value model, and a SELL engine that realizes intermediate P/L on roughly 19% of tickets.

What works: Soccer, MLB, NFL, CBB verticals (+$570K combined). The $0.60–$0.70 band (+$175K). Deep-underdog probes sold on pops ($0.00–$0.20 band, +$253K).

What drags: Tennis (−$150K, large volume), NBA (−$118K, −11.5% ROI), hour 13:00 UTC (−$297K), Union Berlin as an opponent (−$150K across two matches).

The replicable edge: high-conviction filter (dom ≥ 2×, dominant leg only) alone delivers +$651K P/L — six times the unfiltered book's +$101K. The playbook translates this into implementable thresholds.

GAMBLINGISALLYOURNEED — Filter Strategy Experiments

Wallet: 0x507e52ef684ca2dd91f90a9d26d149dd3288beae Window: 2026-04-01 → 2026-04-22 P/L methodology: Cash-flow (BUY outflow + SELL proceeds + remaining-share payout at resolution or MTM). Baseline: 277,096 resolved-market BUYs · 48.6% WR · +0.39% ROI · +$93,922 P/L

All filters operate on BUY trades in resolved markets. SELL trades flow into per-position cash-flow allocation (their proceeds are credited back to the BUY fills proportionally by share).

F1 — Price Band Filter ($0.30–$0.70)

Idea: Strip the extreme tails and trade only the probability zone where mispricings are meaningful.

Verdict: Minimal improvement. The 0.30–0.70 band contains both the profitable $0.60–$0.70 zone and the losing 0.30–0.50 middle. Too coarse.

F2 — Tight Sweet-Spot Band ($0.60–$0.70)

Idea: Trade only the single ROI-positive band with meaningful volume.

Verdict: Nearly doubles the book from a single band rule. Narrower is stronger here.

F3 — High-Conviction Filter (dominance ≥ 2.0×, dominant leg only)

Idea: Use the fair-value model only where it's strongest. Skip balanced markets (no signal) and ignore the non-dominant leg (it's variance reduction, not edge).

Verdict: The largest single filter uplift. Stripping to the dominant leg at ≥ 2× imbalance converts the +$94K base into a +$651K book — a ~7× multiplier. This is the core replicable edge of the wallet.

F4 — Category Filter (kill NBA + Tennis)

Idea: Eliminate the two categories with meaningful volume and negative ROI (NBA −11.49%, Tennis −2.20%).

Verdict: Removing two categories that combined deploy $11.74M of capital and return −$268K nearly quadruples the book's profit. Cleaner than any price-band cut.

F5 — Hour Block Filter (exclude 13:00 UTC)

Idea: Hour 13:00 UTC alone nets −$297,273 — more than the entire NBA book's loss. Kill it.

Extended to worst 4 hours (3, 10, 12, 13):

(The extended set is smaller gain because hour 13 is doing most of the work. Dropping just hour 13 is the cleanest single-hour rule.)

Verdict: Largest single-rule gain of any hour filter in the library. Trading 23/24 hours instead of 24/24 adds ~$300K to the book.

F6 — Deep-Underdog Probe Band ($0.00–$0.20)

Idea: The 0.00–0.20 band prints +$253K combined P/L (F6a +$151K, F6b +$102K) on 33,801 small tickets. The trader probably buys cheap shares then sells on rebound pops.

Verdict: Real edge, but requires tight exit discipline — the high ROI depends on the SELL engine extracting price bumps before resolution. Replicating without the exit logic will not yield these numbers.

F7 — Combined Best (practical stack)

Stack: High-conviction (F3) ∩ exclude hour 13 UTC ∩ kill NBA

Estimated effect (additive filters don't fully compose; treat as a working ceiling):

Verdict: The high-conviction filter is doing most of the work. Layering the hour-13 kill and the NBA kill trims the residual drag. Beyond this three-rule stack, diminishing returns.

F8 — Underdog Blocking (dom ≥ 2×, block non-dominant side entirely)

Idea: Every both-sides market with ≥ 2× imbalance never fires the non-dominant leg — kills the hedge tax directly. This is equivalent to F3 stated as an exclusion rule; the P/L outcome is already captured there (+$651K).

Filter Ranking Summary (by P/L uplift vs baseline)

Practical takeaway: Three rules cover essentially all the edge — High-conviction + kill-NBA + kill-hour-13. Stacking beyond that is window dressing.

GAMBLINGISALLYOURNEED — Implementable Strategy Playbook

Source wallet: 0x507e52ef684ca2dd91f90a9d26d149dd3288beae Window analyzed: 2026-04-01 → 2026-04-22 (22 days) Raw book outcome (cash-flow): +$101,083 P/L / +0.42% ROI on $24.30M BUY notional High-conviction extraction target: +$651,253 P/L / +4.55% ROI / 57.6% WR

1. One-sentence summary

An actively-managed pseudo-MM sports bot that trades global team sports with real sub-$1 paired-cost discipline and a 19% SELL rate — mildly profitable as-is, substantially profitable if the hedge leg is stripped and three filter rules are applied.

2. Market selection

Trade: - Soccer (EPL, LaLiga, Bundesliga, Serie A, Ligue 1, international qualifiers + friendlies) — the biggest profit pool - MLB (all regular-season games) — clean +$147K, +5.66% ROI - NFL (regular + postseason) - CBB during tournament windows (small-sample but +48.8% ROI) - IPL cricket selectively (profitable on individual matches but dragged by losers)

Skip (structural drag): - NBA — −11.49% ROI on $1.41M deployed. Model does not price NBA accurately. - Tennis — largest volume vertical at $10.33M but −$150K / −2.20% ROI. Drop entirely or re-train the model first. - Esports / Counter-Strike / niche — rolled into "Other" with −4.63% ROI.

Universe filter at the CLOB level: - Only binary Yes/No markets - Minimum displayed depth > $500 on both sides (thin markets generate unfillable hedge legs)

3. Entry logic

Fire when all conditions are simultaneously true:

1. Category is in the trade list above (skip NBA / Tennis / Other)
2. Current UTC hour is not 13:00 (single worst hour, −$297K in source window)
3. Entry price is in one of the three edge zones:
4. $0.60–$0.70 (workhorse band, +$175K source) — primary
5. $0.00–$0.20 (probe-and-flip deep underdogs, +$253K combined) — only if the SELL engine is running
6. (Optional) $0.20–$0.30 if the model edge is strong; this band is flat but not catastrophic
7. Fair-value model indicates chosen side is underpriced by ≥ 3 price points
8. Order-book depth at the top of the ask allows a $25 probe in one shot

4. Sizing model

Power-law replication of the source wallet:

Per-market aggregate cap: $260,000 (source's largest single-market book was $261K, IPL Mumbai vs RCB) Per-event aggregate cap: $400,000

5. Both-sides / pairing behavior

Critical deviation from the source wallet: do NOT fire the non-dominant leg.

The source wallet paired both sides on ~60% of markets and paid $8.76M of hedge tax on the eventual losing side. Even with paired cost below $1, the directional overlay at 3×+ dominance is only 62% accurate — not enough to pay back the hedge tax. By trading only the dominant leg (F3 filter), realized P/L jumps from +$94K to +$651K.

Rule: - Open only the leg the fair-value model thinks is underpriced - Do not open the opposite leg as a hedge, ever - Use the SELL engine (§6) to exit losing positions; do not pair

6. Exit strategy (the SELL engine — source wallet's differentiator)

Source wallet is 18.9% SELL. Replicate the exit discipline faithfully — this is what the simpler bots in the library don't do, and it's responsible for converting deep-underdog probes into +$253K.

7. Risk management

• Bankroll minimum: $500K (supports 10–15 concurrent $25K core positions plus reserve)
• Intraday kill-switch: P/L < −$15K → 60-min pause; P/L < −$40K → 24-hr halt
• Max concurrent open positions: 80 markets
• Correlation cap: max 40% of open exposure on a single event day; max 60% on a single category
• Recurring-trap guardrail: the source lost $150K across two Union Berlin matches. Track per-opponent P/L and auto-flag any team with ≥ −$30K cumulative over the trailing 14 days.

8. Edge source

The replicable edge lives in the dominant-leg + category filter + hour filter + exit engine stack. The paired spread is real but too small to justify the hedge tax as the source wallet pays it.

9. Weaknesses to monitor

1. Fair-value model is weak. Dominance curve is flat (50 → 55 → 57 → 62%). Monitor monthly; if the curve flattens further, kill the strategy.
2. NBA / Tennis drift risk. If the selection scheduler picks up NBA or Tennis markets after a regime shift, losses compound fast. Hard-block these categories.
3. Hour 13:00 UTC exposure. This is likely the IPL evening window. If IPL becomes the core instead of European soccer, the hour-13 kill stops working. Reassess weekly.
4. Single-opponent concentration. Union Berlin and Le Havre alone cost the source wallet $190K in one week. Per-opponent exposure limits matter more than per-match limits.
5. Late-window recovery bias. In the source window, 4/22 alone contributed +$139K. Do not assume the recovery day always comes. Size as though the monthly close will be flat.

10. Rebuild parameters (exact thresholds)

# Entry
EDGE_PRICE_BANDS     = [(0.00, 0.20), (0.60, 0.70), (0.20, 0.30)]  # probe / core / optional
MODEL_EDGE_MIN_PTS   = 3
MIN_BOOK_DEPTH_USD   = 500

# Category blacklist
EXCLUDED_CATEGORIES  = {"NBA", "Tennis", "Esports", "Other"}

# Hour blacklist (UTC)
EXCLUDED_HOURS_UTC   = {13}
# Optional extended blacklist:
# EXCLUDED_HOURS_UTC = {3, 10, 12, 13, 18}

# Dominance rule
MIN_DOMINANCE_RATIO_FOR_FIRE = 2.0
FIRE_NON_DOMINANT_LEG        = False    # hard rule — no hedge

# Sizing (USDC)
PROBE_SIZE          = 8
SCALE_IN_SIZE       = 40
CORE_SIZE           = 300
CONVICTION_SIZE     = 1800
PER_MARKET_CAP      = 260_000
PER_EVENT_CAP       = 400_000

# Exit rules (the SELL engine)
TAKE_PROFIT_1_CENTS             = 6      # exit 50%
TAKE_PROFIT_2_CENTS             = 10     # exit remaining
STOP_LOSS_CENTS                 = 8
UNDERDOG_POP_EXIT_CENTS         = 3      # for deep-underdog band
MODEL_EDGE_EXIT_PTS             = 2

# Risk
BANKROLL_MIN                    = 500_000
INTRADAY_PAUSE_THRESHOLD        = -15_000
INTRADAY_HALT_THRESHOLD         = -40_000
MAX_OPEN_POSITIONS              = 80
PER_OPPONENT_DRAWDOWN_TRAILING_14D_LIMIT = -30_000

Expected outcome on a comparable 22-day window: - ROI: +4% to +6% on deployed capital (vs source wallet's +0.42% unfiltered) - Win rate: 55–60% on resolved trades - Per-trade P/L: thin (median +$0.25 to +$1.50) but scaled by volume - Red days: 5–7 out of 22; max intra-period drawdown ~15% of peak P/L

This is a high-volume, thin-edge strategy. The source wallet's book only works at ~275K BUY tickets. Running it at < 5K trades/week will not reproduce the extraction rate — the edge per trade is too small to show up through small-sample variance.