Sabitlenmiş Tweet
Audio Off
5.9K posts
Audio Off
@AudioOfff
Just a dude navigating life. Following me to get updates on a book I am in the slow process of writing. Comment your thoughts.
Katılım Kasım 2024
285 Takip Edilen149 Takipçiler
@polybacktest The one thing that shocked me the most from all my back testing and live testing is how efficient the markets actually are. There is edge out there though.
English
We found a strategy on Polymarket that wins 98.9% of the time.
6,140 BTC 5-minute markets. Real orderbook data.
The strategy: wait until minute 4 of a 5-minute market. Check which way BTC is moving. Bet that direction. That's it.
The results:
Minute 0:30 -- 55.8% win rate
Minute 1:00 -- 59.4%
Minute 2:00 -- 65.0%
Minute 3:00 -- 71.2%
Minute 4:00 -- 77.1%
Minute 4:30 -- 80.8%
Filter for BTC moves greater than 0.2% at minute 4? Win rate: 98.9%. Out of 804 trades, 795 won.
Free money. Right?
Now look at the token price when you're trying to buy in.
At 0:30 the winning token costs $0.60
At 1:00 it costs $0.65
At 2:00 it costs $0.72
At 3:00 it costs $0.80
At 4:00 it costs $0.87
At 4:30 it costs $0.90
You win more often. But you pay more every time. Your profit per win shrinks from $0.40 to $0.10.
The expected value per trade at every single checkpoint?
Negative. Every one of them.
$-0.027 at 30 seconds.
$-0.034 at 2 minutes.
$-0.033 at 4 minutes.
$-0.028 at 4:30.
A 98.9% win rate strategy that loses money. The market already priced in the information before you could act on it.
That's not a broken strategy. That's an efficient market. The more certain the outcome becomes, the less profit is left on the table.
Next time someone shows you a win rate, ask them one question: what was the entry price?
Historical data, one BTC regime. Markets evolve.
The data's all there - polybacktest gives you historical full orderbook snapshots to backtest your strategies on.
English
@polybacktest Love this handle but guys stop haha I figured this out friday and you guys keep giving my edge away
English
We tested whether order book imbalance predicts the winner in BTC 5-minute markets on Polymarket.
It does. But not the way you think.
1,023 markets. 3 million snapshots analyzed.
Core thesis: When one side's order book is stacked with bids, that side should win more often. Market makers know something.
It's the opposite. The order book is lying to you.
When one side has extreme buying pressure (imbalance >1.5), that side wins only 25.3% of the time.
The stronger the signal, the worse it performs:
Mild imbalance (0.1–0.2): 49.0% win rate
Moderate (0.3–0.5): 47.3% win rate
Strong (0.7–1.0): 42.3% win rate
Extreme (1.5+): 25.3% win rate
It gets worse over time. After 4 minutes, the heavy-bid side wins only 31.5%. The bids pile up. The side still loses.
Why? Stale orders. Bids sitting on the losing side aren't smart money they're unfilled limits from when the price was different. Informed flow hits the ask, not the bid. Smart money lifts offers. They don't sit in the book waiting.
The only thing that actually predicts the winner? Price itself. Above 70c: 74% win rate. Above 90c: 94.4%.
Volume confirmation barely helps. Even when the imbalanced side has 5x more volume, win rate only reaches 59.4%.
Bottom line: if you're reading the order book to pick sides, you're trading against yourself. The book shows where liquidity is trapped, not where smart money is going.
Want to run your own backtests? polybacktest has sub-second historical market data and you can test it for free.
English
@Polymarket So this is the new Meta? Pun intended. Who's launching the 20% layoff memecoin lol
English
BREAKING: Meta stock surges following reports they’re laying off 20% of the company due to AI.
English
I've spent about 3 weeks and over $2,000 in Claude tokens setting up multiple agents to trade on Polymarket
It was paper trading for a week and averaged about $80 per day
It's been live for 2 days & completely broken
All of these other people gotta lying on the timeline 😂😂
English
The spread is visible to everyone. But capturing it requires private RPC nodes, Jito bundles, custom transaction routing, and sub-400ms.
Most people see this thread and think ‘I’ll build my own bot this weekend before realizing the hard part was never the strategy.
We spent 4 months and 50,000+ lines of code building this infrastructure so you don’t have to.
We already built it and it’s free to use.
0xMarioNawfal@RoundtableSpace
A BOT TURNED $2,050 INTO $178,000 IN ONE MONTH BY ARBITRAGING 5-MINUTE BITCOIN MARKETS ON POLYMARKET. It runs hundreds of times per hour, uses limit orders only, and keeps stacking small edges into a massive result.
English
@zerohedge I see this as a way to keep the industry real still and use Ai as a tool not a way to replace actors and the creative work that goes into making films.
English
Ben Affleck Once Criticized AI, Now Netflix Is Buying His AI Startup For $600 Million zerohedge.com/markets/ben-af…
English
@zerohedge No one uses btc for currency they would use usdc or usdt. Btc has no use case. Its novelty.
English
Bitcoin Has A Golden Opportunity With AI Agents, It's Time To Build zerohedge.com/crypto/bitcoin…
English
Can I get a GM?
If you reply, I’m following you.
Don't be a bot, GM 🥰
English
@Gustafssonkotte Nice article but I have a hard time taking of this with a grain of salt when it is in Python and not Rust. Also back testing is apples and oranges from live testing. 62% back tested is easily a losing strategy live.
English
@MasonVersluis People who think ripple will move or do anything a comical. Things down on this news lol
English
MASTERCARD REPLYING DIRECTLY TO RIPPLE, BOOM! $XRP
Mastercard@Mastercard
@Ripple Fueling the future of the digital payment world 🌐
English
@polybacktest @marlowxbt Everyone trades or traded Binance latency and there is nothing there. There is always a 0-15 gap difference in price sometimes as high as 40 + but poly price moves with that not CL price.
English
I fed 0x8dxd's entire trade history into DeepSeek. 27,961 positions. Every entry price. Every timestamp. Every payout.
Asked one question: What is this wallet doing that I can't see?
DeepSeek took 47 seconds. Longest I've ever waited for a response. Then returned something that made me screenshot it immediately.
This wallet does not trade Bitcoin. It trades Binance's latency against the platform's update speed. Entries cluster within 8-14 seconds after Binance price movements exceeding 0.12%.
The wallet has no directional bias. It buys whichever side is stale. Estimated edge per trade: 11.3 cents.
0x8dxd. $2,056,408 profit. 27,961 predictions. Joined December 2025.
→ Wallet: polymarket.com/profile/0x63ce…
I asked DeepSeek: Can you calculate the exact delay the wallet exploits?
Based on entry timestamps versus Binance tick data, the average delay is 23 seconds. The wallet enters between second 8 and second 14 of each lag window. Never earlier. Never later. This is not a human clicking. This is a script with a hard-coded wait function.
27,961 trades. 275 per day. $2.05 million in three months.
I asked DeepSeek to estimate the code length.
Based on the execution pattern: one WebSocket listener, one comparison function, one buy trigger with a sleep timer. Approximately 20-30 lines. Any language. Python most likely.
20 lines of code. $2 million.
→ Copy on PolyGun: t.me/PolyGunSniperB…
I asked the question everyone wants answered: Can this be replicated?
DeepSeek: The strategy is replicable. The edge is shrinking. In December the average delay was 34 seconds. Now it's 23 seconds. At current rate of compression the window closes entirely in approximately 4 months. This wallet has extracted $2 million from a gap that is actively disappearing.
1.2 million people watch this wallet. The window is 23 seconds. Four months ago it was 34. The clock is ticking.
I asked one final question: If you were this bot, would you still be running?
DeepSeek: Yes. Even at a 15 second window the math remains positive. But the profit per trade drops from 11 cents to 4 cents. The bot would need to triple its volume to maintain current income. The golden period is ending.
$2,056,408. 27,961 trades. A 23 second window that shrinks every week.
DeepSeek didn't find a strategy. It found an expiration date.
The bot is still running. The window is still open. But DeepSeek can see the exact day it closes.
I saved the conversation. Some things are worth more than the answer.
English
@polybacktest One day ill try and put something together lol but you guys have been very helpful with your data. Much appreciated 👏
English
@AudioOfff Eventually there won’t be enough liquidity to support your fills tho
English
I did a quick 200-market backtest on this, and the results were surprisingly profitable.
The strategy:
- Before every market open, buy one side for $0.50, alternating between up and down.
- Start with $10 and double your bet size after every loss.
- Because you double, you recover all previous losses and still make $10 profit when you eventually win.
Results:
- +$520 PNL
- 55% win rate
- Max drawdown: $80
- Max losing streak: 4
Nonetheless, I still wouldn’t recommend this strategy. It’s profitable—until it isn’t.
If you want to run your own backtest, I use PolyBackTest. It offers sub-second historical market data going back a month with the full order book.
Pablo 👨🏽💻@psbarrales
@polybacktest Could you try Martingale: before the start, place UP, then the next DOWN. What is the max drawdown?
English
@polybacktest Idk what do you think max liquidity threshold is 8 losses in a row is grand. I guess the question is what is the most conservative up or downs have they're been in a row. Or sit out markets until price stabilized to get the chop again.
English
@polybacktest I mean if you have enough money this will always be profitable. It is not capped like a casino table game. NO? Lol
English
@polybacktest Yes backtesting things usually are. Live they never are. Big difference
English
I gave these screenshots to an AI and asked it to build a strategy and backtest it over a million market snapshots. To my surprise, it was actually profitable.
I've seen this formula and these images posted here several times and thought it wouldn't work, but the results speak for themselves.
- 461 trades
- $0.4 average entry
- 50.5% win rate
With a $10 bet size, total profit was $1,086.14.
I even added a 2-second latency delay to trades to ensure replicability, and it remained profitable.
I'll test it much more to confirm, but ultimately I'm very optimistic about this strategy. Most strategies I find on X don't hold up when I backtest them, so I was genuinely surprised when this one came back profitable.
You can run your own backtests on sub-second historical data with PolyBacktest. It has data going back a month, and that's what I used for this test.
What is the Bayesian formula?
The Bayesian formula is a mathematical way to update your up/down guess as new clues appear:
Initial guess: It's a coin flip (50% UP).
The clue: Bitcoin jumps $100 on Binance.
The update: The formula calculates that the new chance of winning is now 80%.
The profit: If the Polymarket price is still 60¢, the bot buys the gap between the market's slow guess and the math's fast truth.
Simple version: It snipes the difference between what the market is paying and what the math says it should be paying.
may.crypto {🦅}@xmayeth
A friend quit his quant fund job and sent me 2 pages. “These are the key formulas I used to make money on Polymarket. $400K a year. If you can apply them, you’ll get rich.” I didn’t believe him. I dropped both pages into OpenClaw and sent one prompt: “build a bot for Polymarket.” Then I left for the gym. When I came back, there was a Telegram message from the agent: “MVP is ready.” Now it’s been making me $150 a day for 4 days straight. I attached both documents. Drop them into your AI agent and tell me what it builds.
English
he told me to never share the rest
i'm sharing it anyway
the 2 pages were just the intro
47 more just landed in my inbox
full quant fund playbook
the models, the formulas, the filters
everything that made him $400K/year
my bot is already running on it
don’t you dare miss this, i’ll upload the rest of the files soon
may.crypto {🦅}@xmayeth
A friend quit his quant fund job and sent me 2 pages. “These are the key formulas I used to make money on Polymarket. $400K a year. If you can apply them, you’ll get rich.” I didn’t believe him. I dropped both pages into OpenClaw and sent one prompt: “build a bot for Polymarket.” Then I left for the gym. When I came back, there was a Telegram message from the agent: “MVP is ready.” Now it’s been making me $150 a day for 4 days straight. I attached both documents. Drop them into your AI agent and tell me what it builds.
English
Audio Off retweetledi
Neither of these men are married or have kids.
Both are simply obsessed with their own personal perfection and optimization.
There is nothing impressive about a single man with no kids sleeping well and being fit.
Show me a man with young children, a full time job, disrupted sleep, who works out regularly, eats healthy, trains Jui Jitsu, with a muscular body…
THIS is impressive. THIS requires extreme discipline.
Camus@newstart_2024
Chris Williamson just shared his "nuclear" sleep stack that's quietly changing his life—and Andrew Huberman breaks down exactly why it works: If you're lying in bed at 2 a.m. scrolling or staring at the ceiling, this 4-minute protocol combo might be the fastest way to shut your brain off without pills. The two killer techniques Williamson swears by: 1. The Mind Walk (visualization on steroids) - Imagine walking a route you know perfectly (your house → front door → street) - Do it with insane detail: feel the shoehorn, hear the key turn, feel the door handle, pressure of the pavement - It's like reading fiction for your nervous system—engages the brain just enough to stop problem-solving loops, but not enough to keep you awake 2. Resonance breathing with the Ohm stone lamp - Bedside lamp with induction-charging stone that has a built-in FDA-cleared HRV sensor - Hold the stone → 3/6/9/12-minute guided sessions with silent tactile vibration (no sound, no light, partner-safe) - Guides you into true resonance frequency (max vagal tone) → the stone knows when you hit it - Williamson calls it “the sickest” sleep tool he’s ever used—currently in stealth (ohmhealth, not widely available yet) Huberman adds the neuroscience: Looking down + eyelids lowering activates parasympathetic circuits and deactivates wakefulness-promoting brainstem nuclei. It’s literally pedaling the sleep pedal while shutting off the alertness arm. Williamson: “Some days you need the adventure story (mind walk), some days you need the physiological hammer (resonance breathing). Stack them and I’m cross-eyed into sleep.” Already trying one of these? Or is your nighttime routine still a war zone?
English
Audio Off retweetledi
She literally tells how to create a frequency so magnetic your desires chase you.
English
Audio Off retweetledi
Someone just built an AI system that runs 60+ AI agents simultaneously and they all learn from each other.
It's called Claude-Flow and it's ranked #1 in agent-based frameworks on GitHub.
One agent plans. Another codes. Another tests. Another reviews security.
All running in parallel. All sharing memory. All getting smarter every run.
The wildest part?
It cuts Claude API costs by 75% using smart routing, simple tasks go to a free WebAssembly layer, complex ones to the right model.
Your Claude subscription just became 2.5x more powerful.
14,100+ developers already starred it.
100% Opensource.
English