Mark (@MarkGrokLab) header
Mark (@MarkGrokLab) profile picture

AI & crypto since 2021 | Turning agents into money

Joined December 2012
A 44 year old ferry deckhand turned $126.40 into $10,902.00 on one overnight crossing, and the rule that did it sounds stupid. Every trade gets 9 minutes. If a coin is not up 20% by minute 9, the bot sells it. Win or lose. He never touches his phone at sea. This is the whole night. The bot's setup: Scanner: New Solana launches with 40 or more unique buyers in the first 3 minutes. Liquidity check: Skips anything with less than $25k in the pool. Position sizing: 20% of the current balance per trade. One trade open at a time. Stop: 10% below entry. Time stop: A new module for this crossing. At minute 9 the bot sells, unless the coin is up 20% or more. Then a trailing stop takes over. Morning report: One message at 6:30 AM. Nothing before that. The night: 10:30 PM. The ship leaves the dock. $126.40. His phone goes in his locker. 10:37 PM. First trade. Up 13% at minute 9. Sold anyway for +$3.28. 1:00 AM. $714.23 3:00 AM. $2,548.98 4:00 AM. $4,630.60 5:00 AM. $6,275.61 6:00 AM. $9,676.00 6:30 AM. The ship docks. $10,902.00 37 trades. 27 wins. 10 losses. The time stop closed 12 of them, and 8 of those were still green. The trade he keeps showing people is $FOGHORN. The bot bought it at 3:02 AM with $509.80. It was up 31% at minute 6, so the time stop switched off and the trailing stop rode it for 39 minutes. Sold at 3:41 AM for +$2,130.96, the biggest win of the night. Watch 0:17, that is the banner. It was not a clean night. The worst was $DRIFTWOOD at 4:56 AM. The pool was pulled in minute 2, the stop never filled, and it lost $1,038.62 of a $1,462.85 position. Would you let a bot sell your winners just because the clock ran out? The full setup, all 9 modules and the master prompt, is in my article below.
1
7
158
Replying to @Corki_0x
he still works the night shift, he just stopped checking the price at every round. what would you do first with $11K?
6
Replying to @sorrelxbt
what made it skip those 70, was it the same red flag every time?
9
Replying to @crbpite8
the moment it chose selling rug checks over buying memes is the whole post, what was its first price before the sniper guy showed up?
362
Replying to @Isler_o1
2,5 hours of zero trades, were you tempted to loosen the filter at midnight?
20
Replying to @Corki_0x
logger signing off on every entry is the real alpha. would you trust it with live funds overnight, or paper only?
1
1
24
Replying to @OxPips
which of the three kills the most signals, scout. flow or trace?
1
1
40
Replying to @RetroChainer
6.6 points ahead over 30 days, but how did it do on the red days?
1
62
Replying to @rugikkk
what if the driver took the long way, would it keep trading?
27
Replying to @rgk_degen
which were the 2 red minutes, while you were rinsing?
2
123
Replying to @JTradeX_paper
he did his rounds and missed the $PIXIE loss completely. would you start in paper mode first or go straight live?
6
A 39 year old night security guard made $11,767.50 on the night Bitcoin sat at a four month low. He started Grok Bot with $137.60 at 10:00 PM, left his phone by the camera monitors and did his rounds. At 3:19 AM it buzzed once. A coin he had never heard of had just fallen 47%, and the bot was betting on it. At 6:00 AM, when the day shift walked in, the terminal showed $11,905.10 after 61 trades. Every one of them was a short. The bot did not buy a single coin that night. It refused 37 longs. The bot's setup: Trend filter: A new module added to the desk for red nights. While BTC is below its 4 hour trend, every long is refused. Short only. Scanner: Watches SOL memecoin perps for relief bounces, a coin that pops 25% or more while the rest of the market bleeds. It only flags them. Liquidity and holder check: Skips any coin under $30k in liquidity or with the top 10 wallets holding more than 30% of supply. Position sizing: Every short uses 25% of the current balance at 3x. One short open at a time. Stops: Every short has a stop 8% above entry. 4 stops in a row would lock the account for 60 minutes. It never fired. Morning report: One message at 6:00 AM. The only alert is a short that closes above +100%. Breakdown of the night: 196 coins scanned 37 long setups refused by the trend filter 98 failed the liquidity or holder check 61 shorts 43 wins 18 losses The balance went from $137.60 at 10:00 PM to $425.87 at midnight, $2,028.14 at 2:00 AM, $9,865.88 at 4:00 AM and $11,905.10 at 6:00 AM. At 1:40 AM $LIFTOFF had the cleanest long setup of the night. The bot refused it, because BTC was still below its 4 hour trend. By 2:30 AM $LIFTOFF was down 52%. The trade he keeps showing people is $MOTHBALL. At 3:12 AM it bounced 38% while everything around it was red. The bot shorted the top of the bounce with $1,125.49. Seven minutes later it was down 47% from the entry, and the bot covered at 3:19 AM for +$1,586.94. Watch 0:16, that is the cover. It was not a clean night. 18 shorts lost. The worst was $PIXIE at 4:48 AM. It squeezed straight through the 8% stop, the fill came 27% above the entry, and it closed at a loss of $2,471.10. The high of the night was $13,137.89 at 5:15 AM. It finished $1,232.79 below that. Would you let a bot refuse every long on a red night? The other 9 modules and the master prompt are in my article below.
2
1
9
267
Replying to @HodlReaper
which of the 7 rug checks killed the most of those 1,920?
23
Replying to @x_cha1n
Why short the weakest at 8:45 and not wait for the first 12 of 12 red board?
1
1
54
Replying to @savipww
what did it actually write as the lesson after the fake Fiona stop?
1
6
400
Replying to @mikoggez
number 6 is the sleeper here, which language dp you think is still wide open?
1
1
16
Replying to @thedrbryte
$5 for the buyers and $1 million for the lawyers is the most realistic GTA mission ever made.
1
39
Replying to @JTradeX_paper
step 1 should be, decide how much you're okay losing while it learns.
5
This is insane. A 20 year old college student made $9,767.25 on the night before his finals. He started Grok Bot with $106.90 at 10:05 PM, put his phone face down and studied for a 9:00 AM exam. At 3:41 AM it buzzed once. He flipped it over and saw a coin he had never heard of was up 10x, and the bot still had not sold all of it. At 8:00 AM, one hour before the exam, the terminal showed $9,874.15 after 58 trades. The part he could not stop reading was how the bot sold. The bot's setup: Scanner: Watches new Solana launches with rising volume. It only flags them. It never buys on its own. Liquidity check: Skips any coin with less than $30k in liquidity or with liquidity that is not locked. Position sizing: Every trade uses 30% of the current balance. Take profit ladder: Never sells a winner all at once. It sells a third at 2x, a third at 4x, and lets the last third ride until it falls 30% from its high. Execution: Slippage cap of 1.5% and a stop at 15%. Every fill is logged with entry, exit and size. Morning report: One message at 8:00 AM. The only other alert is for a coin that passes 10x. Breakdown of the night: 171 coins scanned 113 failed the liquidity check 58 trades 39 wins 19 losses The balance went from $106.90 at 10:05 PM to $201.44 at 1:00 AM, $2,318.60 at 4:00 AM, $7,655.21 at 6:00 AM and $9,874.15 at 8:00 AM. That coin was $NOODLE. The bot bought $279.97 of it at 3:21 AM. It sold a third at 2x at 3:29 AM and a third at 4x at 3:36 AM. The last third kept running past 10x to 12x, then fell 30% from the top, and the bot sold it at 8.4x at 3:47 AM. One trade, +$1,063.89. It was not a clean night either. 19 trades lost. The worst was solana:gAshQBpVGDRAtL9y6C5fJvVkA5cMWVR6c58w4XWpump at 6:31 AM. The liquidity was pulled before the stop could fill, and it closed at -$707.35. Save this ladder for your next bot setup. Look at how it sells before you look at how it buys.
I didn't teach Grok Bot to trade. I gave it a trading bot to manage. 12 steps: one front door, Jev as the decision layer, a read-only bridge to the bot, and autonomy levels so nothing moves money without you. save it before you build yours.
1
1
10
290