BNCEX

Laughing out loud 😂😂😂
$BTC
#OPENCLAW

Today, the strategy has expanded the selection of coins and position limits. The maximum position slot is now 10, but the maximum limit for positions in the same direction is 5.
The stop-loss in the low volatility zone has been reduced, while the profit drawdown tracking take-profit range has also been decreased, allowing for more profits to be secured.
Currently, there are three different response strategies running in parallel: low track, mid track, and high track.
Let's take another look; right now it's just a continuous testing phase. If there are issues, we will make changes, but overall it should be about right.
Not recommended to follow; the strategy is not yet perfected, and this is my first time playing with quantitative trading, so there is a lot to learn. $BTC

Over 20 million lines of data completed in 8 minutes 😂
$BTC

Start training the model! A total of 23.28 million lines of data are being fed directly to it~👀
$BTC
#OPENCLAW ​​​

Why set a blacklist? Because this coin has lost 8 times in a row, which greatly affects the yield and win rate! 😳
$BTC

#OPENCLAW

Work's done!
Get home early to sleep, so you don't risk sudden death.
After today's optimization, the quantification has finally started to slowly recover; I hope nothing crazy happens again.
Currently, I've opened two live accounts for copying trades: A Gan Lobster AI Quant and Xiao Gan Lobster AI Quant.
I don't recommend following it; it's still in the intense optimization phase, so it's best for everyone to simulate and just look at the data.
Next, if there are no issues, we need to consider how to expand the coin pool and whether to add to the blacklist~😘
#量化
$BTC

Train AI trading models with AI, hahaha~, producing data around the clock to feed the model for training!~👀
$BTC #OPENCLAW

After another day of strategy adjustments, I currently feel that there are no issues. I will add a blacklist to filter out false breakouts.
Next, I will continue my model training, letting the backtest factory run 30-40 processes to continuously produce a large amount of backtest data to feed the model for training! 😀$BTC
#量化

How is it? $BTC
😏

Now we start using AI to train AI, feeding trading data to train the model.
The 512g memory machine can handle it without pressure, and it's also good for training models.
#OPENCLAW $BTC

AI is a cheap yet exceptionally intelligent employee. They are professionals in every field, hardworking and dutiful, never complaining about fatigue, working 24/7 without a word of tiredness.
$BTC
#OPENCLAW

The strategy has been fixed, and now the research is on new directions.
The backtesting factory only backtests four fixed coins, and there is no simulation for the dynamic ones, so currently in the live market, there are 20 coins: 4 fixed and 16 dynamically filtered. The monthly return and trading volume in the live market will definitely be more.
I have to mention that the Mac Studio M3 Ultra with 512GB is simply a神器 for backtesting, running it with no pressure, even running many at once is no problem, 8 processes only used one-fifth of the performance 😏
$BTC
#OPENCLAW

Quantitative optimization added a bunch of professional parameters yesterday, and as a result, I woke up this morning feeling foolish. I spent the whole day fixing the strategy!
A typical overfitting loop, working hard and feeling foolish, but it's also a learning process. Starting off by stepping into the pitfalls means that when I officially start investing later, there won't be major issues.
I always wanted to get into quantitative analysis, but there was no one to help me realize it. Now it's good, I can say I've managed to implement this idea myself.
AI is truly a great thing.
$BTC
#OPENCLAW

Educated by crayfish 🦞😂😂😂
#OpenClaw🦞 $$$BTC

A few days ago, the strategy ran in the live market, with a win rate of around 70%, but the profits were not much.
Today, I looked for reasons and upgraded the strategy using lobster. Later, I encountered a problem with a false breakout protection that was too strict, which led to a direct stop loss. It has now been fixed.
The strategy seems to be quite professional now, so I will run it for a few days to see how it goes, and continue to optimize it. If there are no issues, I will add some funds.
I opened a copy trading account; the strategy is still being optimized, so I do not recommend following it. My small account is just testing, and I am sharing the data for everyone to see.
#OPENCLAW $BTC

The strategy optimization laboratory and backtesting factory built by Gān.
Every ten minutes, a wave of solutions is generated by calling the cloud model, submitted to the backtesting factory for progressive backtesting, and the successful solutions are submitted back for review by the main model, forming a closed loop.
The backtesting factory is running at full capacity with eight concurrent processes, currently able to backtest up to three years of K-lines for BN, and it has only used a little bit of performance. This is the correct way to use a Mac Studio M3 Ultra 512GB.
Currently, the strategy that Gān adjusted in the live market last night has already achieved 14 consecutive wins, but it should still be optimized further. The current strategy, after dozens of version iterations, can dynamically scan market token trading volumes and automatically switch dynamic trading currencies, entering and exiting based on signals.
Running for another two days to verify, then preparing to add some funds, considering running a copy trading on BN. $BTC

Understand the applause😏

This is the role of openclaw crayfish AI
24 hours, crazy, and endlessly ongoing strategy research.
Tomorrow, let everyone see the strategy research room and backtesting factory built by Gan.
Downloaded the K-line for three years from BN, continuously producing strategies and backtesting validations. Only strategies that have been validated can enter live testing.
Currently, the strategy is based on dynamic signals for entry and exit, incorporating various indicators, and strictly controlling drawdowns.
Wait for Gan to optimize further to improve the win rate.
Phase 3 is a more intelligent strategy.
There is also an idea to run multiple strategies together, with position management 🌝#OPENCLAW $BTC

Finally, we have reached the point of actual trading.
The strategy has gone through more than 20 major versions, with a backtest monthly return of 20. I feel that it can be used initially, so I'll start testing it and optimize it while running.
There are still some issues with the actual trading script, so I'll continue to have OpenClaw the crayfish modify and optimize it for me.
The online literature team has created an outline factory to generate outlines in bulk.😏
#OPENCLAW $BTC

How does the liver feel like it should be empty?
Is it an illusion?😳$BTC
#比特币突破7.5万美元