ExitLiquidty

Wallet: 0xeb6789ca6b1425ff908a69a2a5469c38532cd696 Window: 2026-03-30 to 2026-05-09 (41 calendar days, 34 active) Universe: 1,234 trades across 73 markets and 24 unique events. $382,813 BUY notional, $524,163 SELL notional, $906,977 gross turnover. Net resolved-BUY P/L: +$185,765 on $299,809 resolved-buy notional = +86.4% ROI over 41 days

The Punchline

ExitLiquidty is a specialist in crypto token launch markets and public sale commitment markets, two niche Polymarket categories that most participants treat as too opaque or too risky to trade systematically. The strategy is not directional speculation in the conventional sense. It is threshold calibration: the operator builds a ladder of Yes positions across FDV thresholds for a single launch event, sizes heavily on the thresholds they assess as near-certain to clear, and accepts bounded losses on the stretch thresholds that miss.

The result over 41 days: 560 resolved BUYs with an 82.5% win rate, +$185,765 in realized P/L, +$330,940 in total realized cash flow including SELL proceeds. 73 markets, 24 events. The book is extraordinarily concentrated: the MegaETH launch family (multiple FDV threshold markets across a single launch) accounts for the majority of both volume and P/L. This is not diversification. It is deep specialization in a narrow niche where the operator has genuine informational and analytical advantages.

The wallet name is self-referential: ExitLiquidty provides exit liquidity to panicking sellers in illiquid niche markets, then collects when the fundamentals were right all along.

What This Trader Trades

The universe is three overlapping categories of prediction markets:

1. Token launch FDV threshold markets. Markets of the form "Will [Token]'s FDV exceed $[X]M one day after launch?" The key series from the top-markets table:

MegaETH >$600M one day after launch    156 trades   $57,495 volume   +$11,609 P/L   24 wins / 0 losses resolved
MegaETH >$800M one day after launch    131 trades  $143,193 volume   +$50,968 P/L   70 wins / 0 losses resolved
MegaETH >$1B one day after launch      151 trades  $159,607 volume   +$51,555 P/L  134 wins / 0 losses resolved
MegaETH >$1.5B one day after launch     19 trades   $14,790 volume   +$43,715 P/L   16 wins / 0 losses resolved
MegaETH >$2B one day after launch       19 trades   $25,486 volume   -$21,665 P/L    0 wins / 18 losses resolved
EdgeX >$400M one day after launch       36 trades    $7,781 volume    +$3,548 P/L    1 win  / 0 losses (partially unresolved)

The pattern is a calibrated ladder: massive positions on the thresholds that cleared (up to $1.5B), correct rejection of $2B. The wallet's P/L is essentially the net of a well-structured options ladder on the MegaETH launch outcome.

2. Public sale commitment markets. Markets asking whether a project's token sale will reach a specific funding level. The Printr public sale series dominates the late-April portion of the CSV. These markets ask questions like "Over $6M committed to the Printr public sale?" and "Over $60M committed to the Printr public sale?" The operator buys "No" on the high-threshold questions when the running commitment total makes those thresholds mathematically impossible to reach by the deadline, then sells at near-$1.00 when the position is confirmed correct.

3. Token launch yes/no binary markets. "Will [Token] launch a token by [Date]?" markets, typically bought on the "Yes" side when launch is imminent.

What is NOT in this book:

• No BTC/ETH Up/Down markets
• No sports, politics, or current events
• No DeFi protocol performance markets
• No high-frequency microstructure plays
• 0% both-sides participation (confirmed by both_sides.rate = 0.0)

This is a pure directional specialist. The operator bets one side on every market, and the side they choose is almost always the one that wins.

The Order of Operations: One Launch, Trade by Trade

The EdgeX launch (March 31, 2026) is the cleanest single-event trace in the dataset, showing the full entry-accumulation-exit cycle.

Event: EdgeX token launch. Markets: FDV above $300M, $400M, $500M.

Walk-through: The operator enters EdgeX >$400M early with inventory carried from the prior day (the earliest CSV entries show SELLs at $0.65, implying BUYs happened before the CSV sample begins). As the launch day progresses and real-time DEX data confirms EdgeX's FDV is tracking above $400M, the price rises from $0.61 to $0.63+. Rather than simply holding to settlement, the operator actively sells blocks of shares into the rising price, converting unrealized gains into realized cash. When the price briefly dips (11:24 UTC, sell at $0.55), they accept the lower price to maintain liquidity. Then at 12:32 UTC, they re-enter with a large block at $0.70, apparently confident the $400M threshold will clear. The market resolves Yes. Total net P/L on EdgeX >$400M: approximately +$3,548.

The parallel $500M position (BUYs at $0.59-$0.64) ultimately returned additional P/L. The $300M position was sold entirely at $0.80-$0.81 before settlement, capturing a slightly lower price than the $1.00 settlement but securing the gain risk-free.

The key behavioral signature: This trader both buys AND sells the same position within a single launch day, adjusting size and direction based on how DEX prices evolve during the measurement window. It is not pure buy-and-hold. It is active portfolio management within each launch event.

Why It Works: The Math

The strategy's positive EV comes from a structural mispricing that persists in thin FDV markets:

Mechanism: FDV launch markets price "Yes" at a discount
            to what real-time DEX data implies

Example (MegaETH >$1B):
  Polymarket "Yes" price at entry (avg):   ~$0.40-$0.60
  Implied probability from CLOB:           40-60%
  
  Actual FDV outcome:                      ~$1.2B (cleared $1B threshold)
  Resolved settlement:                     $1.00 per "Yes" share
  
  ROI per share:   ($1.00 - $0.50) / $0.50 = +100% per share
  Aggregate ROI (134 wins, 0 losses):      +86.4% on resolved positions

Why the market was wrong:
  1. Thin orderbook liquidity = high bid-ask spread = lazy pricing
  2. Retail uncertainty premium: most buyers don't know DEX pre-launch data
  3. Settlement lag: the market closes >24h after launch, 
     but DEX price is visible within minutes of token generation

For the public sale "No on stretch thresholds" play:

Example (Printr "Over $60M committed"):
  Commitment at time of entry:    ~$12M (clearly below $60M)
  Time to deadline:               <24 hours
  "No" share price at entry:      $0.99
  Settlement:                     $1.00
  
  Gross P/L per $1,000 deployed:  ~$10 (1% spread, but replicated
                                   many times with large clips)
  CSV shows clips of $499-$1,996 per fill across 10+ fills per market
  Aggregate on Printr series:     Still unresolved but positioned for profit

The EV calculation for the FDV ladder:

MegaETH ladder P/L (realized):
  >$600M threshold:  +$11,609 (24 wins, 0 losses, avg entry ~$0.42)
  >$800M threshold:  +$50,968 (70 wins, 0 losses, avg entry ~$0.32)
  >$1B threshold:    +$51,555 (134 wins, 0 losses, avg entry ~$0.40)
  >$1.5B threshold:  +$43,715 (16 wins, 0 losses, avg entry ~$0.25)
  >$2B threshold:    -$21,665 (0 wins, 18 losses, avg entry ~$0.21)
  
  Net MegaETH ladder:  +$136,182 (approx, combining best markets)
  
  The $2B position was a $21K loss against $136K+ in gains
  from the clearing thresholds. Loss/gain ratio: ~15.4%.
  This is the correct structure: small losses on stretch bets,
  large wins on near-certain bets.

Phase 1: Trader Profile

Scale and Activity

Trade Size Distribution

The size distribution is power-law with a fat right tail. The max ($52,200) is 321x the median ($162.59). The top 5% of trades carry 47.9% of capital. This is the signature of a conviction-scaling strategy: small exploratory positions, large positions when conviction is high. The $52,200 maximum is almost certainly a large block on the MegaETH >$1B or >$800M threshold at a moment of high conviction about the impending resolution.

Execution Signature

The gap distribution is bimodal: 10.9% of fills are under 10 seconds (bot-like bursts, visible in the CSV as same-second fan-outs on a single market), but the median is 24 minutes and the mean is 8.4 hours. This is not a bot in the SirMartingale sense. It is a semi-automated, primarily human-directed operation that fires quick bursts of orders (walking the orderbook with 5-15 fills in quick succession on a single market), then goes quiet for hours or days between events.

Trading Hours (UTC)

Peak activity: 02:00-11:00 UTC (Asia/Europe morning). The hour histogram shows:

02:00 UTC: 96 trades   (peak)
03:00 UTC: 76 trades
04:00 UTC: 84 trades
05:00 UTC: 92 trades
07:00 UTC: 96 trades   (peak)
09:00 UTC: 85 trades
10:00 UTC: 87 trades
11:00 UTC: 92 trades

Activity drops sharply after 14:00 UTC (falling to 16-32 trades per hour) and nearly stops after 20:00 UTC (5-8 trades). This pattern suggests Asia-Pacific timezone operation or a deliberate strategy of trading before US market hours open, when liquidity in these niche markets is thinnest and spreads are widest.

Archetype: Concentrated event-specialist directional bettor with semi-automated execution. Niche: crypto token launch FDV markets and public sale commitment markets.

Phase 2: Core Strategy Identification

Both-sides participation: 0.0%

Zero markets had both Yes and No sides purchased. The wallet is purely directional on every single position. This immediately eliminates market-making, spread capture, and hedging as strategy components.

Classification: Pure directional betting, archetype B, with a specialist niche focus that functions as a form of stale-price arbitrage (archetype C) when the operator has real-time DEX data that the Polymarket orderbook has not yet priced in.

The strategy is not:

• Market making (0% both-sides, confirmed)
• Copy trading (the markets are too niche and too event-specific for copy-following to explain timing)
• DCA accumulation (single-event concentration, not long-term accumulation)
• Pure longshot hunting (82.5% win rate refutes this; most positions are on high-probability outcomes)

The strategy is:

• Calibrated FDV threshold betting: the operator maps a launch outcome to a specific FDV level and positions accordingly
• Information-edge directional trading: real-time DEX data allows the operator to form a high-confidence FDV estimate before the market prices it correctly
• Public sale near-certainty capture: buying near-certain "No" positions on stretch thresholds in commitment markets

Phase 3: Dominance Ratio Analysis

Both-sides participation is 0.0%. The dominance ratio framework is structurally inapplicable. There are no paired markets to analyze.

What replaces dominance analysis here is threshold ladder architecture: the operator's "conviction" is expressed through the capital allocation across different FDV threshold levels. They allocate heavily to the thresholds they expect to clear and lightly (or not at all) to the stretch thresholds they consider unlikely but possible. The MegaETH example shows this explicitly: heaviest capital at the $1B threshold ($159K volume), with decreasing capital at $1.5B ($14K) and $2B ($25K stretch bet that lost).

Phase 4: Entry Price Analysis

Two outlier bands require explanation:

The $0.10-$0.20 band is the single most extraordinary finding in the price data. 43 resolved trades, 88.4% win rate, and +$46,872 P/L on $16,381 of capital = +286% ROI. These are positions where the operator paid 10-20 cents for shares that resolved at $1.00 with 88% probability. The MegaETH >$1.5B position (CSV: megaeth-market-cap-fdv-1pt5b-one-day-after-launch) is the primary driver: entries at $0.15-$0.20 on a threshold that ultimately cleared at a $1.2B+ FDV. The "best markets by P/L" table confirms this: >$1.5B returned +$43,715 on just $14,790 of volume (16 wins, 0 losses resolved). At a $0.15 entry, each winning share pays $0.85 net, a 5.7x return.

The $0.20-$0.30 band is the only deeply negative band after sub-$0.10: 24 trades at 29.2% win rate, -$6,907 P/L, -40.7% ROI. This band likely captures the MegaETH >$2B and >$3B stretch positions, where entries in the 0.21-0.29 range reflected a small probability assessment of super-high FDV outcomes that didn't materialize.

The $0.00-$0.10 band shows 0% win rate on 46 trades. These are pure longshot bets (or token launch "No" positions bought at deep discounts that resolved No). The -72.6% ROI reflects that shares at $0.05 that resolve at $0 return -100%, while shares at $0.05 that resolve at $1 would return +1,900%. With 0% wins in this band, every trade was a complete loss.

Sub-bucket concentration check: The dominant single price point is $0.42 for the MegaETH >$600M early entries, $0.32 for the >$800M entries, and $0.20-$0.15 for the >$1.5B entries. No single tick dominates the entire book (the operator uses a wide range of prices as they walk the orderbook), but within individual market series, entries cluster at specific price levels reflecting the operator's fair-value estimate at time of entry.

Phase 5: Category and Vertical Breakdown

The report_data.json collapses the entire book into a single "Other" category (+$185,765 P/L, +61.9% ROI). The standard classification framework doesn't apply because these are niche crypto launch markets not captured by the standard keyword groups. The meaningful breakdown is by event family:

The clear verdict: MegaETH is the dominant alpha source. The Printr series represents the next major event bet but is unresolved in the data window. The LoL eSports losses (-$2,327 total across two trades) are clearly off-thesis and represent the only cases where the operator strayed from the niche they know.

Phase 6: Timing and Execution

Hourly P/L Distribution

The best absolute P/L hours cluster in the early-morning UTC window:

The only negative hour with meaningful sample is 11:00 UTC (-$1,517 on 53.8% win rate). This is a weak but real signal. Hours 19-23 UTC have low trade counts and mixed results, which is the period when this operator appears less active.

Day of Week P/L

Wednesday is massively dominant: +$165,835 P/L at +168.3% ROI. This is not a structural day-of-week effect. It almost certainly reflects that the MegaETH token launched on a Wednesday (the launch date falls within this dataset window), and the Wednesday session captured the bulk of the MegaETH ladder resolution. The operator entered MegaETH positions over several preceding days and collected the bulk of the P/L on Wednesday launch day. This confirms the event-concentration nature of the strategy.

Burst Patterns

The CSV shows characteristic bursts: 5-15 fills on a single market in rapid succession (often multiple fills per second), followed by hours of silence. Examples from the EdgeX series on March 31: fills at 12:34:55, 12:35:01, 12:35:49 (three times), 12:36:53, 12:37:05, 12:37:21, 12:37:25, 12:37:33, 12:37:39, 12:39:33, 12:39:51, 12:41:19 (twice), 12:41:25. Fourteen fills in seven minutes as the operator walks the orderbook at market open. This is automated order execution within a human-directed strategy.

Phase 7: Filter Experiments

The price filter destroys -$54,285 of P/L. The mechanism is the same as for SirMartingale: the highest-ROI band ($0.10-$0.20 at +286% ROI) is stripped out by the $0.30-$0.70 filter, losing the single most profitable trade cluster in the dataset. The $0.10-$0.20 band holds 43 trades and +$46,872 of P/L. The filter removes those entirely.

The hour filter does modest damage (-$6,868) mostly by cutting the 11:00 UTC hour (negative P/L, -$1,517) along with several other marginal hours. The net improvement from removing the bad hour is actually positive (+$1,517 rescued), but the combined effect of also cutting good hours 19, 20, and 23 creates a small net negative.

The high-conviction filter is structurally inapplicable (0% both-sides).

Conclusion: The standard filter battery is either inapplicable or destructive for this trader. The single most important filter insight is the same as Phase 4 found: the $0.10-$0.20 entry zone is the alpha concentrate. Do not filter it out.

Phase 8: Rolling Window Consistency

The trajectory is not consistent: three strong weeks then two weeks of losses. The cumulative P/L peaked at $195,508 on April 19 and declined to $185,765 by May 1. The rolling 15-day window went negative from May 2 onward. This is the concentrated event-risk manifesting: when the events under bet (Printr, and possibly later launches) resolve unfavorably, the rolling window bleeds. The strategy's consistency depends entirely on the quality of the operator's launch event selection.

Phase 9: P/L Decomposition

The total realized cash flow is +$330,940 against +$185,765 in resolved-BUY P/L. The difference (+$145,175) represents SELL proceeds captured above the BUY cost basis, reflecting the operator's active exit management: selling positions into the orderbook as prices rise toward settlement rather than holding everything to $1.00. The Printr public sale series shows this vividly: dozens of sells at $0.98-$0.999 on "No" positions that should resolve at $1.00, capturing near-certainty value slightly early.

The SELL/BUY ratio is $524,163 / $382,813 = 1.37x, meaning the operator extracts $1.37 in SELL proceeds for every $1.00 of BUY notional. This is the active exit management fingerprint: they are not passive hold-to-settlement traders.

Phase 10: Strategy Specification

One-sentence summary: A concentrated specialist in crypto token launch FDV threshold markets and public sale commitment markets who builds calibrated position ladders across FDV thresholds, enters using real-time DEX data as an information edge, sizes heavily on near-certain clearing thresholds, actively manages exits into the orderbook as prices rise, and accepts small bounded losses on the stretch thresholds that miss.

Edge sources:

1. Information advantage on FDV outcomes via real-time DEX data not yet reflected in Polymarket CLOB pricing
2. Thin liquidity in niche launch markets creates persistent mispricings vs fundamental value
3. Public sale commitment tracking allows near-certainty "No" capture on mathematically impossible stretch thresholds

What works: MegaETH-type launches where the FDV is visible on DEXes during the measurement window. The $0.10-$0.20 entry zone (286% ROI). Early UTC hours (02:00-11:00). Near-certain "No" positions on public sale stretch thresholds. Active SELL exit management (1.37x SELL/BUY ratio adds $145K vs holding to settlement).

What drags: The $2B+ stretch thresholds (-$21,665 on MegaETH >$2B). The sub-$0.10 longshot zone (0% win rate, -$3,772). Late-window weeks 17-18 show the event-concentration risk: when the operator's events don't clear, rolling windows go red quickly. Off-niche bets (LoL eSports) are pure losses.

Key rebuild parameters: FDV threshold ladder structure (buy Yes on expected-clearing thresholds, skip/short stretch thresholds). Entry price range $0.15-$0.70 (sweet spot by ROI). Active SELL management: sell 30-50% of position as price rises to $0.80+, hold rest to settlement. Position sizing: largest clips on highest-conviction threshold levels, small exploratory clips on stretch thresholds. Market selection: token launches with pre-announced tokenomics and DEX price discovery. Avoid eSports, avoid crypto Up/Down, avoid any market without a quantifiable real-time signal.

Wallet: 0xeb6789ca6b1425ff908a69a2a5469c38532cd696 Window: 2026-03-30 → 2026-05-09 (34 active / 41 calendar days) Methodology: Cash-flow P/L = -buy_usdc + sell_usdc + remaining_share_payout. Resolved shares settle at $1 (win) / $0 (loss); open positions marked at last price.

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: 0.00% (0 of 73 markets)

Dominance buckets

Phase 4 - Entry-Price Analysis

Phase 5 - Category & Vertical Breakdown

Phase 6 - Timing & Execution

Net P/L by hour (UTC)

Phase 8 - Rolling Window Consistency

• Rolling 7-day windows green: 27 of 41 (65.9%)
• Rolling 7-day P/L range: -$21,593 → +$83,715
• Rolling 15-day windows green: 33 of 41 (80.5%)
• Rolling 15-day P/L range: -$22,201 → +$140,864

Weekly P/L

Phase 9 - P/L Decomposition

Phase 10 - Top Markets by Volume

Top 10 winners by P/L

Top 10 losers by P/L

Report generated 2026-05-14 10:00 UTC.

Wallet: 0xeb6789ca6b1425ff908a69a2a5469c38532cd696 Window: 2026-03-30 to 2026-05-09 Baseline: 560 resolved BUYs, 82.5% WR, $299,809 deployed, +$185,765 P/L, +86.4% ROI Total realized cash flow (including SELLs): +$330,940

The headline result

Two filters do meaningful damage. One does modest improvement. The rest are inapplicable or no-ops.

The $0.30-$0.70 price filter cuts 29% of P/L (-$54,285). The high-conviction dominance filter is structurally inapplicable (0% both-sides). The hour filter does small net harm (-$6,868) because it discards a modestly bad hour (11:00 UTC) alongside several good ones. The category filter is an identity operation because 100% of trades fall in "Other."

The single most useful genuine filter for this strategy is not in the standard battery: event selection. The two LoL eSports trades posted a combined -$2,327 at a 0% win rate. The off-niche trades are cleanly separable from the core FDV/public-sale strategy and should simply be excluded. But that is a strategy rule, not a quantitative filter.

Filter results table

Filter-by-filter commentary

1. Price $0.30-$0.70 filter

DESTRUCTIVE

Delta: -$54,285 (-29.2% of baseline P/L)

The price filter removes 186 trades from the baseline. The critical loss is the $0.10-$0.20 band: 43 trades, 88.4% win rate, +$46,872 P/L at +286% ROI. These are the MegaETH >$1.5B positions entered at $0.15-$0.20, which cleared the threshold and resolved at $1.00 each. The filter strips out the single highest-ROI band in the entire book.

The $0.20-$0.30 band also gets removed: 24 trades at 29.2% win rate and -$6,907 P/L. The filter correctly removes those losers. But the $0.10-$0.20 winners more than outweigh the $0.20-$0.30 losers by 6.8x. The net effect of removing both bands is a large loss.

Additionally, the filter removes the $0.70-$1.00 zone (76 trades, 98.7% combined win rate, +$22,406 P/L). Those were near-certainty positions with thin but reliable spreads. Removing them subtracts real profit with minimal variance reduction.

The sweet-spot filter is directly misaligned with this strategy. The operator's edge peaks at sub-$0.30 entries where information advantage creates large mispricings. Filtering to $0.30-$0.70 cuts out exactly where the alpha lives.

2. High-conviction dominance ≥ 2x filter

NOT APPLICABLE

Delta: -$185,765 (total elimination)

The wallet has 0% both-sides participation. Zero markets have paired Yes and No positions. The dominance ratio framework requires a pairing dimension to compute. With no pairs, the filter returns an empty set and the full baseline P/L disappears. This is structural, not a tuning issue.

The absence of both-sides trades is itself strategically informative: this operator is so confident in their directional calls that they never hedge. The 82.5% win rate across 560 resolved positions supports that confidence empirically.

3. Top category filter (Other)

NO-OP

Delta: $0 (identity operation)

100% of trades fall into the "Other" category because the standard keyword classifier doesn't have entries for "token launch FDV markets" or "public sale commitment markets." The category filter is a complete identity operation here. It neither helps nor hurts.

For replication, the useful category-equivalent filter would be market type whitelist: include only markets with slugs matching *-fdv-above-*, *-market-cap-fdv-*, *committed-to-the-*-public-sale*, and *launch-a-token*. This is a strategy rule, not captured by the standard category framework, but would cleanly exclude the LoL eSports trades and any other off-niche positions.

4. Exclude worst hours (11, 19, 20, 23 UTC)

MILD HARM

Delta: -$6,868 (-3.7% of baseline P/L)

The worst-performing hourly buckets are 11:00 UTC (-$1,517, 53.8% WR), 19:00 UTC (50.0% WR), 20:00 UTC (62.5% WR), and 23:00 UTC (60.0% WR). Excluding those four hours removes 49 resolved trades and saves $1,517 in losses from the 11:00 slot, but costs $8,385 in positive P/L from the other three hours that also contain winning trades.

The net is -$6,868. The filter does mild harm. The operator's worst-performing hourly bucket (11:00 UTC) happens to contain some of their large fills on markets that didn't resolve well in that specific hour, but it's not a structural pattern worth filtering on. It reflects which events happened to resolve at which times, not a consistent hour-of-day edge effect.

One genuine refinement the hour filter doesn't capture: the 02:00-07:00 UTC window shows consistently strong performance (+$26K in hour 02, +$28K in hour 03) and appears to be when the operator enters fresh positions on live launches in the Asia session. If capacity-constrained, over-weighting the 02:00-07:00 window (where the operator seems freshest and most active) would preserve the best trades.

5. Combined filter (price $0.30-$0.70 + exclude worst hours)

DESTRUCTIVE

Delta: -$66,055 (-35.6% of baseline P/L)

The combined filter inherits the damage from the price filter and adds to it. ROI improves modestly to 65.3% from 86.4%, but absolute P/L falls from +$185,765 to +$119,710. This is the standard case where stacking filters on an already-optimized strategy removes more good trades than bad.

The combined filter's "higher win rate" (92.0% vs 82.5%) is a statistical artifact: removing the correctly-priced low-probability bets ($0.10-$0.20 band with 88.4% WR, lower than 92.0%) makes the average win rate rise while destroying absolute dollar value.

6. What filters would actually add value

The standard battery misses the two filters that would genuinely improve this strategy's replication:

The first filter (market slug whitelist) is directly computable from the trade CSV and would be the only one in this list that a replicator can implement immediately. It prevents the wallet's two worst categories of loss: LoL eSports bets and any other non-niche trades that occasionally appear.

Bottom line for replication

The base strategy is already highly optimized within its niche. Three concrete filter recommendations:

1. DO NOT apply the $0.30-$0.70 price filter. The $0.10-$0.20 band is the alpha concentrate, generating +$46,872 (25% of P/L) on 7.7% of trades. Filtering it out destroys the strategy's highest-ROI wing.

1. DO implement a market-type whitelist replacing the category filter. Accept only: *-fdv-above-*, *-market-cap-fdv-*, *committed-to-the-*-sale*, *launch-a-token*. Reject LoL, NFL, and anything else outside the niche.

1. DO track event-level concentration risk rather than applying per-trade filters. The biggest risk in this strategy is single-event concentration, not individual bad trades. When more than 40% of deployed capital is in one event (e.g., MegaETH), apply an event-level cap at 35% of bankroll.

The correct "filter" for this strategy is understanding which events the operator has a genuine information edge on, and trading only those. That is a research process, not a mechanical filter, which is why the standard PR&R battery is largely inapplicable here.

Source wallet: 0xeb6789ca6b1425ff908a69a2a5469c38532cd696 Strategy: Crypto token launch FDV threshold ladder + public sale commitment near-certainty capture Reference book: $382,813 BUY notional, +$330,940 total realized cash flow, +$185,765 resolved-BUY P/L, +86.4% ROI in 41 days

One-paragraph operator brief

Build a research-driven prediction market position system focused exclusively on two market types: (1) token launch FDV threshold markets ("Will [Token] FDV exceed $X one day after launch?") and (2) public sale commitment markets ("Over $Y committed to the [Project] sale?"). For FDV markets, monitor DEX price discovery during the first hours post-launch and build a calibrated position ladder across Yes thresholds you assess as likely to clear. Size the heaviest capital on the highest-confidence, mid-range thresholds. Take small or no positions on stretch thresholds above your expected FDV. For public sale markets, track real-time commitment data and buy "No" on thresholds that the current commitment total cannot mathematically reach by the deadline. In both cases, sell 30-50% of positions actively into the orderbook as prices rise rather than holding everything to settlement. Expect concentrated wins when your launch thesis is correct and bounded losses when stretch thresholds miss. Run during 02:00-11:00 UTC when orderbook liquidity in these niche markets is thinnest and mispricings are largest.

1. Market selection

Why exclude eSports and sports: The two LoL trades in this wallet lost 100% (-$2,327). The operator's information edge is entirely derived from crypto-specific data feeds. Sports markets require entirely different signal sources and the operator has none. The wallet name "ExitLiquidty" is self-descriptive in these markets: you are the exit liquidity, not the informed buyer.

Event pipeline process:

Weekly research loop:
  1. Monitor Polymarket for new FDV markets on upcoming launches
     - Source: Polymarket market feed filtered by slug pattern
  2. Identify launch date, token economics (total supply, initial circulation)
  3. Find comparable launch FDV data from similar projects
  4. Map DEX that will host initial price discovery (Uniswap, Raydium, etc.)
  5. Set expected FDV range and identify clearing vs. stretch thresholds
  6. Queue positions for the launch date

2. Entry logic

For FDV threshold markets:

def should_enter_fdv(market, threshold, expected_fdv_range):
    # Whitelist check
    if not matches_fdv_slug_pattern(market.slug):
        return None
    
    # Eligibility: do we have DEX data to form a view?
    if not has_live_dex_feed(market.token):
        return None
    
    # Pre-launch entry (before DEX data available):
    # Use tokenomics model only. Enter small exploratory position.
    if hours_until_launch(market) > 0:
        if threshold <= expected_fdv_range[0] * 0.85:  # clearly expected to clear
            return "pre_launch_small_entry"
        return None  # wait for DEX data on uncertain thresholds
    
    # Post-launch entry (DEX data available):
    current_fdv = get_dex_fdv(market.token)
    
    # Clear-clearing threshold: FDV > threshold by >15% margin
    if current_fdv > threshold * 1.15:
        if market.yes_price < 0.85:  # still mispriced
            return "high_conviction_buy_yes"
    
    # Near-clearing threshold: FDV within 5-15% of threshold
    if threshold * 0.95 < current_fdv < threshold * 1.15:
        if market.yes_price < 0.60:
            return "moderate_conviction_buy_yes"
    
    # Stretch threshold: FDV well below threshold
    if current_fdv < threshold * 0.85:
        return None  # don't buy yes on a threshold that won't clear
    
    return None

For public sale commitment markets:

def should_enter_commitment(market, threshold, current_commitment, 
                             time_to_deadline):
    # Is the threshold mathematically impossible to reach?
    implied_rate = current_commitment / elapsed_time
    projected_total = current_commitment + (implied_rate * time_to_deadline)
    
    if projected_total < threshold * 0.80:  # clearly won't reach it
        if market.no_price > 0.90:  # "No" priced at >90 cents
            return "near_certainty_no_buy"
    
    # Standard liquidity check
    if market.no_side.depth_at_price(0.99) < 5000:
        return None  # not enough depth to make it worth the transaction
    
    return None

3. Exit logic

The operator is not a passive hold-to-settlement trader. The SELL/BUY ratio of 1.37x and +$145K gap between resolved-BUY P/L and total realized cash flow confirm active exit management.

def manage_position(position, current_market_price, time_to_resolution):
    # Tranche 1: Sell 30% when price reaches entry_price * 2.0
    # (locked in a double, reduce risk)
    if current_market_price >= position.avg_entry * 2.0:
        if not position.tranche_1_sold:
            sell_shares(position, fraction=0.30)
            position.tranche_1_sold = True
    
    # Tranche 2: Sell another 20% when price reaches $0.85+
    if current_market_price >= 0.85:
        if not position.tranche_2_sold:
            sell_shares(position, fraction=0.20)
            position.tranche_2_sold = True
    
    # Tranche 3: Sell 20% more in the final hours before resolution
    # (convert to cash, avoid resolution risk on the full position)
    if time_to_resolution < 3600 and current_market_price >= 0.90:
        if not position.tranche_3_sold:
            sell_shares(position, fraction=0.20)
            position.tranche_3_sold = True
    
    # Remaining 30%: hold to resolution at $1.00
    # If market resolves Yes, residual pays $1.00 per share
    # If thesis changes (DEX FDV drops below threshold): sell all immediately
    
    if get_current_fdv() < position.threshold * 0.90:
        sell_all(position)  # thesis broken, exit at market

Why stagger exits instead of holding everything to settlement: The operator captures three advantages from active selling: (1) realized gains are not subject to resolution-day market risk; (2) the SELL proceeds can be redeployed into fresh positions on the same or similar events; (3) on events where the FDV is close to a threshold, early selling locks in profit before the DEX data shifts.

4. Sizing model

This strategy uses conviction-scaling sizing calibrated to the expected probability of the threshold clearing.

Bankroll scaling:

Reference book ($380K deployed over 41 days):
  Estimated peak working capital:     ~$150,000-$200,000
  (capital cycles within a launch event, 24-48 hour resolution)
  
  Monthly P/L from resolved BUYs:     ~+$136,000 (extrapolated from 41-day window)
  ROI on peak working capital:        ~+90% monthly (86.4% / 1.37 months)

Scaled-down version ($50,000 bankroll):
  Scale all clip sizes by 50K/200K = 25%
  High-conviction clip:  $1,250-$5,000
  Expected monthly P/L:  ~$34,000 (rough extrapolation)
  
Key constraint: position size is bounded by orderbook depth
in these niche markets. Above ~$50K per threshold tier,
you start moving your own entry price materially.

Event concentration risk rule:

Max single-event exposure: 40% of total bankroll
  (MegaETH peaked at ~41% of the reference book's total
   buy notional; keep it below this level to preserve
   the ability to recover from a bad launch)

Max stretch threshold exposure: 10% of single-event allocation
  (The $2B stretch cost $21K against $160K+ in MegaETH
   clearing threshold gains, a ~13% loss/gain ratio.
   Keep stretch bets small: 5-15% of event allocation.)

5. The position ladder structure

This is the defining architecture of the strategy. For each launch event, map the FDV threshold markets and allocate capital in an inverse-confidence pyramid:

FDV ladder example (based on MegaETH structure):
  Expected FDV range from tokenomics model: $1.0B-$1.5B

  Threshold    Expected    Entry    Allocation  Logic
  ---------------------------------------------------------------
  >$600M       Near-cert   $0.42    25% of budget  Almost certainly clears
  >$800M       Near-cert   $0.32    30% of budget  High confidence
  >$1.0B       High conf   $0.40    30% of budget  Core position
  >$1.5B       Moderate    $0.15    10% of budget  Upper end of range
  >$2.0B       Longshot    $0.21     5% of budget  Stretch, small only
  >$3.0B       Very long   $0.058    0% of budget  Skip entirely

The math of the correct ladder: If your FDV estimate is $1.2B with a standard deviation of $300M:

Prob(>$600M):  ~97%   Entry at $0.42 → Expected ROI: [(0.97 × $0.58) - (0.03 × $0.42)] / $0.42 = +122%
Prob(>$800M):  ~90%   Entry at $0.32 → Expected ROI: [(0.90 × $0.68) - (0.10 × $0.32)] / $0.32 = +191%
Prob(>$1.0B):  ~77%   Entry at $0.40 → Expected ROI: [(0.77 × $0.60) - (0.23 × $0.40)] / $0.40 = +108%
Prob(>$1.5B):  ~35%   Entry at $0.15 → Expected ROI: [(0.35 × $0.85) - (0.65 × $0.15)] / $0.15 = +98%
Prob(>$2.0B):  ~10%   Entry at $0.21 → Expected ROI: [(0.10 × $0.79) - (0.90 × $0.21)] / $0.21 = -52%

Decision: Buy all thresholds except >$2B (negative EV unless
          the market prices it far below your probability estimate)

The $2B threshold at $0.21 is negative EV if your true probability is 10%. The reference wallet entered it anyway (and lost $21,665), which suggests either a lower entry price was targeted, or the operator's FDV model was slightly optimistic. At any entry price below $0.10, the $2B threshold becomes positive EV at a 10% probability. The lesson: stretch threshold entries are only defensible at deeply discounted prices.

6. Hour scheduling

The 02:00-11:00 UTC concentration in the reference wallet is not arbitrary. Launch events (especially for Asian and European crypto projects) often go live during Asia/Europe hours, and the Polymarket orderbook in niche FDV markets takes 30-120 minutes to price in the DEX data. That window is the entry opportunity.

7. Operational requirements

Critical distinction from high-frequency bots: This strategy does not require sub-second latency or co-location. The edge comes from having better fundamental information (tokenomics model + DEX data), not from being faster than other participants. A 30-second DEX polling interval is sufficient. The opportunity window for each launch event is typically 6-24 hours, not milliseconds.

8. Risk profile

Maximum loss scenario: A launch where the FDV lands well below even the lowest threshold (e.g., a token launches at $200M FDV against a $600M+ ladder) would produce near-total losses on the entire event allocation. Against a $200K event allocation, that's a -$200K single-event loss. The reference wallet avoided this because MegaETH launched within the expected range. Implement the 40% event cap and the stretch-threshold sizing rules to bound the worst-case loss at roughly 40% of bankroll per bad event, not 100%.

9. Diagnostic checklist: is the strategy still working?

Run after each resolved event:

Run monthly:

• Scan for upcoming token launches with Polymarket FDV markets
• Update tokenomics comparable database (comparable recent launches by FDV at similar supply/float)
• Review whether bid-ask spreads on FDV thresholds at launch time have compressed (competition signal)

10. What this playbook deliberately does NOT include

No $0.30-$0.70 sweet-spot filter. The $0.10-$0.20 band is the alpha concentrate (+286% ROI, +$46,872 P/L). Applying this filter cuts 25% of the strategy's realized profit. The low-price entries are not lottery tickets; they are deeply discounted near-certainties in the MegaETH >$1.5B case where the operator correctly identified the launch outcome.

No diversification across categories. The reference wallet's two off-niche trades (LoL eSports) lost 100%. This strategy's edge is entirely niche-specific. Spreading into sports, politics, or crypto Up/Down markets dilutes the edge without adding any. Concentrate in the niche you know.

No automated fully-systematic execution. The entry decisions require genuine research and judgment about tokenomics, comparable FDV analysis, and real-time DEX monitoring. A system that fires on slug patterns alone would enter every FDV market including ones where the operator has no information edge. The semi-automated model (human-approved entry, automated orderbook walking for execution) matches what the CSV evidence suggests the reference wallet is doing.

No hold-to-settlement default. Holding 100% of positions to settlement leaves money on the table when the orderbook is willing to pay $0.90+ for shares well before the 24-hour measurement window closes. The 1.37x SELL/BUY ratio in the reference wallet proves that active exit management adds substantial realized cash flow above what pure settlement would provide.

No recurring daily bot rhythm. This strategy fires episodically when launches happen, not on a daily schedule. There may be no qualifying trades for an entire week, followed by a day where $150K of capital is deployed in a 6-hour window. The scheduler should be event-driven, not calendar-driven.

The whole point of this strategy is that it requires real research and real conviction. Every mechanism the operator uses, from the ladder structure to the active exits to the early-UTC entry timing, exists to extract maximum value from a genuine information advantage in a narrow niche. A replicator who builds the research pipeline gets the edge. A replicator who tries to systematize the entries without the research gets the same win rate as a random buyer in these markets, which is not 82.5%.