Find out what works.

Samples up to five pump.fun launches from their first 10 minutes for 30 minutes each, pricing them with a ~$20 Jupiter buy quote every cycle. Entry (v4): data gate passed over 3 minutes, the quote-implied price is at least 0.5% above its prior high in that window, and curve progress (the liquidity proxy: SOL in the curve) is steady or rising over the window: at least two curve checks 2+ minutes apart, latest not below the first, within 1 point of the high, no drop over 2 points. Round-trip cost 0–5%. Stop and targets are measured from the sell-back value at entry.

Data gate (all entries): at least 4 observations, an observation 3–4 minutes old, no gap over 150 seconds, latest at most 90 seconds old. Tokens failing the gate are only watched.

Same as Group 3: $20 entry budget, three accounts with identical entries, maximum three open positions per account, one new entry per cycle, no re-entry. Exits: −5% loss trigger for all plans (measured from the sell-back value right after entry, so round-trip costs alone cannot trigger it; reported P/L still includes every cost); take10 / take20 exit at +10% / +20% net or after 15 minutes; split40 sells half at +40% then trails 10% from its peak, 30-minute limit. Assumed 0.00005 SOL per transaction plus 0.00204 SOL setup at entry. Quote-implied prices include swap fees and price impact for a ~$20 buy and are comparable only to each other.

2026-10-26 19:20:12 UTC → 2026-10-26 19:50:27 UTC · exits: 15-minute time limit, 30-minute time limit · final

• Trading costs took a large share of the position before the market moved (about 5.0% lost to costs (market moved -1.03%, worst plan -6.01%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Momentum never came; the trade drifted until the time limit (exit: 15-minute time limit, 30-minute time limit; market moved -1.03%).
  Better: The entry lacked follow-through; demand stronger momentum, or shorten the time limit to cut dead trades sooner.

2026-10-26 13:17:45 UTC → 2026-10-26 14:09:37 UTC · exits: 15-minute time limit, 30-minute time limit · final

• Trading costs took a large share of the position before the market moved (about 5.2% lost to costs (market moved -0.66%, worst plan -5.88%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Momentum never came; the trade drifted until the time limit (exit: 15-minute time limit, 30-minute time limit; market moved -0.66%).
  Better: The entry lacked follow-through; demand stronger momentum, or shorten the time limit to cut dead trades sooner.

2026-10-26 11:08:28 UTC → 2026-10-26 11:09:27 UTC · exits: loss trigger · final

• Price collapsed between checks, so the loss went far past the stop (worst plan lost -17.61%, more than 10 points past its stop).
  Better: Check open positions more often than entries, and avoid tokens whose liquidity or price is unstable before entry; a stop only protects at the moment it is checked.
• Trading costs took a large share of the position before the market moved (about 3.4% lost to costs (market moved -14.21%, worst plan -17.61%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price came back above the entry price after the exit (reached 18.77% vs entry within 15 minutes of the exit).
  Better: The stop was too tight for this token's volatility; size the stop to the typical 1-minute move (e.g. 2×) or confirm the stop on two samples.
• Price kept falling after the exit, so the entry was wrong, not the exit (-52.87% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.

2026-10-26 10:36:28 UTC → 2026-10-26 10:37:26 UTC · exits: loss trigger · final

• Price collapsed between checks, so the loss went far past the stop (worst plan lost -40.74%, more than 10 points past its stop).
  Better: Check open positions more often than entries, and avoid tokens whose liquidity or price is unstable before entry; a stop only protects at the moment it is checked.
• Price kept falling after the exit, so the entry was wrong, not the exit (-82.9% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.

2026-10-26 06:50:10 UTC → 2026-10-26 06:52:09 UTC · exits: loss trigger · final

• Price collapsed between checks, so the loss went far past the stop (worst plan lost -16.55%, more than 10 points past its stop).
  Better: Check open positions more often than entries, and avoid tokens whose liquidity or price is unstable before entry; a stop only protects at the moment it is checked.
• Trading costs took a large share of the position before the market moved (about 8.2% lost to costs (market moved -8.38%, worst plan -16.55%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price came back above the entry price after the exit (reached 271.4% vs entry within 15 minutes of the exit).
  Better: The stop was too tight for this token's volatility; size the stop to the typical 1-minute move (e.g. 2×) or confirm the stop on two samples.
• The trade was in profit before it turned into a loss (peaked at +7.13% while held).
  Better: Protect gains: a trailing stop once the trade is up ~5%, or a lower first target.

2026-10-26 06:02:38 UTC → 2026-10-26 06:04:17 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 4.9% lost to costs (market moved -0.16%, worst plan -5.09%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Stopped out quickly by a small price move (exited after 99 s on a -0.16% market move).
  Better: Give new entries room: no stop in the first 2–3 minutes unless the loss passes ~10%, or set the stop from recent volatility instead of a fixed −5%.
• The entry "breakout" was a tiny move above the recent high, within normal noise (entry price only 0.05% above the prior 5-minute high).
  Better: Require a larger breakout (≥1–2%) or a new high on two consecutive samples.

2026-10-26 06:00:37 UTC → 2026-10-26 06:04:16 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 4.9% lost to costs (market moved -0.11%, worst plan -5.05%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• The entry "breakout" was a tiny move above the recent high, within normal noise (entry price only 0.02% above the prior 5-minute high).
  Better: Require a larger breakout (≥1–2%) or a new high on two consecutive samples.

2026-10-26 05:51:36 UTC → 2026-10-26 05:52:35 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 4.9% lost to costs (market moved -0.12%, worst plan -5.05%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Stopped out quickly by a small price move (exited after 59 s on a -0.12% market move).
  Better: Give new entries room: no stop in the first 2–3 minutes unless the loss passes ~10%, or set the stop from recent volatility instead of a fixed −5%.
• The entry "breakout" was a tiny move above the recent high, within normal noise (entry price only 0.1% above the prior 5-minute high).
  Better: Require a larger breakout (≥1–2%) or a new high on two consecutive samples.

Samples up to five curves at ≥50% progress (not complete, not migrated, checked in the last 30 minutes) for 30 minutes each, pricing them with a ~$20 Jupiter buy quote every cycle. Entry (v4): data gate passed over 3 minutes, quote-implied price at least 0.5% above its high in that window, curve progress steady or rising over the window (as Group 1), and latest curve progress ≥55% and not below its value at slot start. Round-trip cost 0–5%. Stop and targets are measured from the sell-back value at entry.

Data gate (all entries): at least 4 observations, an observation 3–4 minutes old, no gap over 150 seconds, latest at most 90 seconds old. Tokens failing the gate are only watched.

Same as Group 3: $20 entry budget, three accounts with identical entries, maximum three open positions per account, one new entry per cycle, no re-entry. Exits: −5% loss trigger for all plans (measured from the sell-back value right after entry, so round-trip costs alone cannot trigger it; reported P/L still includes every cost); take10 / take20 exit at +10% / +20% net or after 15 minutes; split40 sells half at +40% then trails 10% from its peak, 30-minute limit. Assumed 0.00005 SOL per transaction plus 0.00204 SOL setup at entry. Quote-implied prices include swap fees and price impact for a ~$20 buy and are comparable only to each other.

2026-10-26 23:46:19 UTC → 2026-10-26 23:48:19 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 4.0% lost to costs (market moved -5.43%, worst plan -9.46%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price came back above the entry price after the exit (reached 15.97% vs entry within 15 minutes of the exit).
  Better: The stop was too tight for this token's volatility; size the stop to the typical 1-minute move (e.g. 2×) or confirm the stop on two samples.

2026-10-26 22:42:19 UTC → 2026-10-26 23:07:29 UTC · exits: 15-minute time limit, loss trigger · final

• Trading costs took a large share of the position before the market moved (about 6.6% lost to costs (market moved -2.78%, worst plan -9.4%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.

2026-10-26 23:06:19 UTC → 2026-10-26 23:07:29 UTC · exits: loss trigger · final

• Price collapsed between checks, so the loss went far past the stop (worst plan lost -84.81%, more than 10 points past its stop).
  Better: Check open positions more often than entries, and avoid tokens whose liquidity or price is unstable before entry; a stop only protects at the moment it is checked.
• Trading costs took a large share of the position before the market moved (about 87.6% lost to costs (market moved 2.81%, worst plan -84.81%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Stopped out quickly by a small price move (exited after 69 s on a 2.81% market move).
  Better: Give new entries room: no stop in the first 2–3 minutes unless the loss passes ~10%, or set the stop from recent volatility instead of a fixed −5%.
• Price kept falling after the exit, so the entry was wrong, not the exit (-85.67% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.

2026-10-26 21:35:19 UTC → 2026-10-26 21:35:29 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 21:11:17 UTC → 2026-10-26 21:26:29 UTC · exits: 10% net target, 15-minute time limit, loss trigger · final

• Trading costs took a large share of the position before the market moved (about 17.4% lost to costs (market moved 5.14%, worst plan -12.25%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price kept falling after the exit, so the entry was wrong, not the exit (-10.57% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.
• The trade was in profit before it turned into a loss (peaked at +17.81% while held).
  Better: Protect gains: a trailing stop once the trade is up ~5%, or a lower first target.

2026-10-26 20:33:44 UTC → 2026-10-26 20:34:09 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 20:23:01 UTC → 2026-10-26 20:27:20 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 8.5% lost to costs (market moved -4.15%, worst plan -12.65%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price kept falling after the exit, so the entry was wrong, not the exit (-17.95% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.

2026-10-26 18:55:12 UTC → 2026-10-26 18:55:22 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 16:31:00 UTC → 2026-10-26 16:31:58 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 4.2% lost to costs (market moved -4.35%, worst plan -8.57%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price came back above the entry price after the exit (reached 201.37% vs entry within 15 minutes of the exit).
  Better: The stop was too tight for this token's volatility; size the stop to the typical 1-minute move (e.g. 2×) or confirm the stop on two samples.

2026-10-26 12:29:21 UTC → 2026-10-26 13:11:42 UTC · exits: 15-minute time limit, 30-minute time limit · final

• Trading costs took a large share of the position before the market moved (about 5.9% lost to costs (market moved -0.74%, worst plan -6.67%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Momentum never came; the trade drifted until the time limit (exit: 15-minute time limit, 30-minute time limit; market moved -0.74%).
  Better: The entry lacked follow-through; demand stronger momentum, or shorten the time limit to cut dead trades sooner.

2026-10-26 12:11:20 UTC → 2026-10-26 12:11:45 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 11:47:35 UTC → 2026-10-26 11:49:31 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 4.0% lost to costs (market moved -6.19%, worst plan -10.2%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.

Entry (v5): four or more observations spanning at least three minutes, gaps no greater than 150 seconds, same pool, price above the prior observed high, reported liquidity at least $5,000 and stable or rising over the window (not below its start, within 3% of its high, no drop over 10% between samples), and latest rolling five-minute volume at least $500. Open positions are checked every 15 seconds. This is sampled-price momentum, not candle or unique-wallet analysis.

Sampling (v3): five tokens hold 30-minute slots; a slot is released early after three straight observations under $1,000 reported liquidity, so a newer graduation can be sampled.

$20 total entry budget; maximum three positions per account; one new shared entry per cycle; no re-entry. Quote-based round-trip loss must be between 0% and 3%, including assumed costs. Tokens with absent data are skipped. Ownership / social checks are not connected.

All plans attempt a loss exit at −5% of entry cost in USD. Plans 1 and 2 exit at +10% / +20% net or after 15 minutes. Plan 3 sells half when estimated whole-position net value reaches +40%; the remainder exits on a 10% decline from its highest observed unit quote, or 30 minutes from entry. Targets are triggers, not guaranteed prices.

Assumed costs: 0.00005 SOL each transaction plus 0.00204 SOL token-account setup at entry, with no rent refund. Provider quote outputs include their modeled swap costs. No extra 1% haircut is added. Fill delay, MEV, failed-transaction fees and wallet-specific execution are not modeled. Failed exit checks retain the position; no sale is invented. These accounts are separate simulations, not a combined portfolio.

2026-10-27 01:10:16 UTC → 2026-10-27 01:40:29 UTC · exits: 15-minute time limit, 30-minute time limit · final

• The entry "breakout" was a tiny move above the recent high, within normal noise (entry price only -0.02% above the prior 5-minute high).
  Better: Require a larger breakout (≥1–2%) or a new high on two consecutive samples.
• The safety screen had flagged this token (mint authority set: more supply can be created).
  Better: Compare flagged and unflagged results; if flagged tokens keep losing, block them as a new rule version.
• Momentum never came; the trade drifted until the time limit (exit: 15-minute time limit, 30-minute time limit; market moved -0.04%).
  Better: The entry lacked follow-through; demand stronger momentum, or shorten the time limit to cut dead trades sooner.

2026-10-27 01:08:15 UTC → 2026-10-27 01:38:29 UTC · exits: 15-minute time limit, 30-minute time limit · final

• The entry "breakout" was a tiny move above the recent high, within normal noise (entry price only 0.02% above the prior 5-minute high).
  Better: Require a larger breakout (≥1–2%) or a new high on two consecutive samples.
• The safety screen had flagged this token (mint authority set: more supply can be created; freeze authority set: holder accounts can be frozen; permanent delegate can move or burn any holder's tokens; new holder accounts can start frozen; transfers can be paused; transfer hook runs custom code on every transfer).
  Better: Compare flagged and unflagged results; if flagged tokens keep losing, block them as a new rule version.
• Momentum never came; the trade drifted until the time limit (exit: 15-minute time limit, 30-minute time limit; market moved -0.28%).
  Better: The entry lacked follow-through; demand stronger momentum, or shorten the time limit to cut dead trades sooner.

2026-10-27 00:32:15 UTC → 2026-10-27 01:02:29 UTC · exits: 15-minute time limit, 30-minute time limit · final

• Trading costs took a large share of the position before the market moved (about 2.7% lost to costs (market moved 2.29%, worst plan -0.37%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• The entry "breakout" was a tiny move above the recent high, within normal noise (entry price only 0.06% above the prior 5-minute high).
  Better: Require a larger breakout (≥1–2%) or a new high on two consecutive samples.
• Momentum never came; the trade drifted until the time limit (exit: 15-minute time limit, 30-minute time limit; market moved 2.29%).
  Better: The entry lacked follow-through; demand stronger momentum, or shorten the time limit to cut dead trades sooner.

2026-10-27 00:10:14 UTC → 2026-10-27 00:11:14 UTC · exits: loss trigger · final

• Price collapsed between checks, so the loss went far past the stop (worst plan lost -25.97%, more than 10 points past its stop).
  Better: Check open positions more often than entries, and avoid tokens whose liquidity or price is unstable before entry; a stop only protects at the moment it is checked.
• Trading costs took a large share of the position before the market moved (about 23.4% lost to costs (market moved -2.57%, worst plan -25.97%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Stopped out quickly by a small price move (exited after 59 s on a -2.57% market move).
  Better: Give new entries room: no stop in the first 2–3 minutes unless the loss passes ~10%, or set the stop from recent volatility instead of a fixed −5%.
• Price came back above the entry price after the exit (reached 117.79% vs entry within 15 minutes of the exit).
  Better: The stop was too tight for this token's volatility; size the stop to the typical 1-minute move (e.g. 2×) or confirm the stop on two samples.

2026-10-27 00:03:14 UTC → 2026-10-27 00:03:29 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 23:26:15 UTC → 2026-10-26 23:27:29 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 16.1% lost to costs (market moved 8.14%, worst plan -7.97%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price kept falling after the exit, so the entry was wrong, not the exit (-98.49% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.
• The trade was in profit before it turned into a loss (peaked at +8.14% while held).
  Better: Protect gains: a trailing stop once the trade is up ~5%, or a lower first target.
• The entry "breakout" was a tiny move above the recent high, within normal noise (entry price only -16.63% above the prior 5-minute high).
  Better: Require a larger breakout (≥1–2%) or a new high on two consecutive samples.

2026-10-26 22:27:14 UTC → 2026-10-26 22:28:14 UTC · exits: 10% net target, loss trigger · final

• Price kept falling after the exit, so the entry was wrong, not the exit (-53.72% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.

2026-10-26 21:39:16 UTC → 2026-10-26 21:39:58 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 21:03:11 UTC → 2026-10-26 21:03:25 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 20:56:12 UTC → 2026-10-26 20:56:25 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 20:39:43 UTC → 2026-10-26 20:40:12 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 19:30:10 UTC → 2026-10-26 19:30:22 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

v4: stablecoins and issuer-controlled tokenized assets (xStocks-style permanent delegate + pausable, or scaled-amount extension) are excluded. Everything is scaled to the token's volatility: the median 15-minute price range over the last 6 hours (needs 45 minutes of prices; tokens under 0.4% are too calm). Entry: pullback of 2–8× volatility (max 25%) then an uptick; the round trip must cost less than one volatility. Exits replace the shared ones: stop −2× volatility (1–5%); take10 +2×, take20 +4×, split40 half at +6× then a 2× trailing decline; time limits 60/60/90 minutes. Earlier v3 rule, for reference: samples up to five tokens that pass the Group 4 filters for 60 minutes each with batched DEX Screener prices. Entry (v3): data gate passed over 15 minutes (pool switches allowed), reported liquidity ≥$50,000 and stable or rising over the window (not below its start, within 3% of its high, no drop over 10% between samples), a 3–25% pullback from an observed peak, and the latest price above the previous sample. Round-trip cost 0–3%.

Data gate (all entries): at least 6 observations, an observation 15–22 minutes old, no gap over 150 seconds, latest at most 90 seconds old. Tokens failing the gate are only watched.

Same as Group 3: $20 entry budget, three accounts with identical entries, maximum three open positions per account, one new entry per cycle, no re-entry. Exits: −5% loss trigger for all plans; take10 / take20 exit at +10% / +20% net or after 15 minutes; split40 sells half at +40% then trails 10% from its peak, 30-minute limit. Assumed 0.00005 SOL per transaction plus 0.00204 SOL setup at entry. Quote-implied prices include swap fees and price impact for a ~$20 buy and are comparable only to each other.

2026-10-27 01:58:16 UTC → 2026-10-27 03:02:14 UTC · exits: 60-minute time limit, volatility stop (−1.6%) · final

• The entry "breakout" was a tiny move above the recent high, within normal noise (entry price only 0.52% above the prior 5-minute high).
  Better: Require a larger breakout (≥1–2%) or a new high on two consecutive samples.
• The safety screen had flagged this token (transfer hook runs custom code on every transfer).
  Better: Compare flagged and unflagged results; if flagged tokens keep losing, block them as a new rule version.

v1: candidates come from the Group 5 social screen (DEX Screener token profile or community takeover, liquidity ≥$20,000, 1-hour volume ≥$50,000, ≥100 trades, ≥55% buys, buys ≥ sells in the last 5 minutes). Up to five sampled for 30 minutes with batched DEX Screener prices. Entry: data gate over 3 minutes (same pool), liquidity ≥$20,000 and stable or rising, price at least 0.5% above its 3-minute high, and the social screen still passing within the last 10 minutes. Round-trip cost 0–3%.

Data gate (all entries): at least 4 observations, an observation 3–4 minutes old, no gap over 150 seconds, latest at most 90 seconds old. Tokens failing the gate are only watched.

Same as Group 3: $20 entry budget, three accounts with identical entries, maximum three open positions per account, one new entry per cycle, no re-entry. Exits: −5% loss trigger for all plans; take10 / take20 exit at +10% / +20% net or after 15 minutes; split40 sells half at +40% then trails 10% from its peak, 30-minute limit. Assumed 0.00005 SOL per transaction plus 0.00204 SOL setup at entry. Quote-implied prices include swap fees and price impact for a ~$20 buy and are comparable only to each other.

2026-10-27 03:38:17 UTC → 2026-10-27 03:40:29 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 4.4% lost to costs (market moved -4.35%, worst plan -8.75%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price kept falling after the exit, so the entry was wrong, not the exit (-26.31% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.

2026-10-27 03:37:17 UTC → 2026-10-27 03:38:59 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 6.4% lost to costs (market moved 0.97%, worst plan -5.41%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Stopped out quickly by a small price move (exited after 102 s on a 0.97% market move).
  Better: Give new entries room: no stop in the first 2–3 minutes unless the loss passes ~10%, or set the stop from recent volatility instead of a fixed −5%.
• Price kept falling after the exit, so the entry was wrong, not the exit (-7.5% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.

2026-10-27 02:59:17 UTC → 2026-10-27 03:00:14 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 51.4% lost to costs (market moved 37.25%, worst plan -14.18%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price came back above the entry price after the exit (reached 104.04% vs entry within 15 minutes of the exit).
  Better: The stop was too tight for this token's volatility; size the stop to the typical 1-minute move (e.g. 2×) or confirm the stop on two samples.
• Price kept falling after the exit, so the entry was wrong, not the exit (-29.4% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.
• The trade was in profit before it turned into a loss (peaked at +37.25% while held).
  Better: Protect gains: a trailing stop once the trade is up ~5%, or a lower first target.
• The entry "breakout" was a tiny move above the recent high, within normal noise (entry price only -0.54% above the prior 5-minute high).
  Better: Require a larger breakout (≥1–2%) or a new high on two consecutive samples.

2026-10-27 02:03:17 UTC → 2026-10-27 02:03:29 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-27 01:55:16 UTC → 2026-10-27 01:56:14 UTC · exits: loss trigger · final

• Price came back above the entry price after the exit (reached 9.28% vs entry within 15 minutes of the exit).
  Better: The stop was too tight for this token's volatility; size the stop to the typical 1-minute move (e.g. 2×) or confirm the stop on two samples.

2026-10-27 01:43:17 UTC → 2026-10-27 01:53:44 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 15.6% lost to costs (market moved 9.87%, worst plan -5.69%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• The trade was in profit before it turned into a loss (peaked at +9.87% while held).
  Better: Protect gains: a trailing stop once the trade is up ~5%, or a lower first target.

2026-10-27 01:07:17 UTC → 2026-10-27 01:07:29 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-27 00:15:19 UTC → 2026-10-27 00:20:29 UTC · exits: 10% net target, loss trigger · final

• Trading costs took a large share of the position before the market moved (about 32.0% lost to costs (market moved 17.73%, worst plan -14.25%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• The trade was in profit before it turned into a loss (peaked at +17.73% while held).
  Better: Protect gains: a trailing stop once the trade is up ~5%, or a lower first target.

2026-10-27 00:10:19 UTC → 2026-10-27 00:10:44 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 21:05:16 UTC → 2026-10-26 21:05:25 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 20:44:49 UTC → 2026-10-26 20:45:12 UTC · exits: loss trigger · final

• Not enough saved prices around the trade to diagnose it (no price observation near the entry or during the hold).
  Better: Keep sampling a traded token until at least 15 minutes after its exit.

2026-10-26 20:19:01 UTC → 2026-10-26 20:21:55 UTC · exits: loss trigger · final

• Trading costs took a large share of the position before the market moved (about 14.8% lost to costs (market moved 5.87%, worst plan -8.98%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price kept falling after the exit, so the entry was wrong, not the exit (-36.04% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.
• The trade was in profit before it turned into a loss (peaked at +5.87% while held).
  Better: Protect gains: a trailing stop once the trade is up ~5%, or a lower first target.

v1: candidates are the leading copy of an active narrative burst (≥6 launches of the same ticker from ≥4 creators within 5 minutes; leader = most curve progress, ≥8%, at least a minute old). Up to five sampled for 15 minutes with ~$20 Jupiter quote prices; burst copies get a curve check every cycle. Entry: data gate over 2 minutes (≥3 prices), quote-implied price ≥1% above its 2-minute high, curve progress steady or rising, still the burst leader. Round-trip cost 0–5%. Relaxed exits replace the shared ones, measured from the sell-back value at entry: stop −15%; take10 → +30%, take20 → +60%, split40 → half at +30% then a 15% trailing decline; time limits 12 / 12 / 15 minutes.

Data gate (all entries): at least 3 observations, an observation 2–3 minutes old, no gap over 150 seconds, latest at most 90 seconds old. Tokens failing the gate are only watched.

2026-10-26 23:16:20 UTC → 2026-10-26 23:23:14 UTC · exits: −15% stop · final

• Trading costs took a large share of the position before the market moved (about 22.6% lost to costs (market moved 0.13%, worst plan -22.42%)).
  Better: Only enter when the expected move clearly exceeds the round-trip cost: require a breakout at least 2× the quoted round-trip cost, or skip tokens costing more than ~2% to trade.
• Price kept falling after the exit, so the entry was wrong, not the exit (-18.34% vs entry 15 minutes after the exit).
  Better: Add entry confirmation: a second higher high, rising volume, or rising liquidity before buying.

v1: candidates are Solana projects that passed the weekly CoinGecko use-case screen and the 5-minute market check (liquidity ≥$250k, pool ≥90 days, 24h volume ≥1% of market cap, 24h change below +10%). Entry: still passing, no mint authority (safety screen within 7 days), round-trip cost 0–2%. No price-spike trigger. $100 per entry, up to 5 holdings per account. Exits from the full cost: stop −35%; take10 → all at +100%, take20 → all at +200%, split40 → half at +100% then a 30% trailing decline; 180-day holding limit.

Data gate (all entries): at least 2 observations, an observation 1–1 minutes old, no gap over 150 seconds, latest at most 90 seconds old. Tokens failing the gate are only watched.

v1: a fixed universe of 15 liquid Solana tokens (JUP, BONK, WIF, JTO, PYTH, RAY, RENDER, HNT, ORCA, KMNO, W, POPCAT, PENGU, TRUMP, MEW), all sampled all the time with DEX Screener prices. Daily candles from GeckoTerminal for the pool the live price comes from, refreshed every 6 hours. Entry: 50 completed daily closes up to yesterday; live price above the 20-day average, 20-day average above the 50-day average, yesterday's close above the 20-day average, price at most 15% above the 20-day average (no chasing), liquidity ≥$100k, round-trip cost 0–2%. $150 per entry, up to 6 holdings per account, no re-entry under v1. Exits from the full cost: stop −15%; trend break when the live price is 3% below the 20-day average; take10 → trend break only, take20 → all at +30%, split40 → half at +30% and the rest on the trend break; 120-day holding limit.

Data gate (all entries): at least 2 observations, an observation 1–1 minutes old, no gap over 150 seconds, latest at most 90 seconds old. Tokens failing the gate are only watched.

v1: candidates pass the Group 9 survivor screen in the last 30 minutes (token 3–14 days old by its earliest DEX Screener pool, liquidity ≥$40k, market cap $0.3M–$50M, 24h volume ≥$100k, ≥500 trades, 6h above −20% and 24h above −30%, buys ≥90% of sells), found among GeckoTerminal's 100 busiest and 60 trending Solana pools and our own pump.fun migrations. Up to five sampled for 60 minutes with batched DEX Screener prices. Entry: data gate over 10 minutes (≥6 prices), liquidity ≥$40k and stable or rising, safety screen within 24 h with no mint or freeze authority, a calm dip (1h change −15% to −3% while 6h change ≥0) and the latest price above the previous sample. Round-trip cost 0–3%. $50 per entry, up to 5 holdings per account, no re-entry. Exits from the full cost: stop −20%; take10 → +25%, take20 → +50%, split40 → half at +40% then a 20% trailing decline; time limits 48 / 48 / 72 hours.

Data gate (all entries): at least 6 observations, an observation 10–15 minutes old, no gap over 150 seconds, latest at most 90 seconds old. Tokens failing the gate are only watched.