CryptoPainter
CryptoPainter
An old friend calls me a "painter", technical/data analysis and quantitative trading, providing various tricky angles to see the market, and using time to leverage. The real account is an agent account, a self-evolving strategy system is being tested, please do not copy!
32Following
3Kfollowers
Feed
Feed
ethereum:0xb1110919016846972056ab995054d65560d5f05e has finally dropped…
Not sure if it's time to hedge and short?
Positions haven't continuously decreased with the price drop, and the funding rate remains stable without significant deviation, indicating that when shorts entered, there was an opposing side in the market. Who is the counterparty?
The long-short ratio has risen sharply, meaning more people are going long, so retail traders going long might have increased…
Conclusion: Market makers started closing long positions yesterday afternoon at 16:00 after exploiting the new high to trigger a short squeeze, then took a break. Currently, the market decline is caused by retail traders competing without market makers supporting the bottom. Among them, short positions have large capital but fewer people, while long positions have smaller capital but more people.
The market makers' task is basically done. Using a unique chip supply-demand relationship, they pushed the price up nearly 10 times with a small amount of zero-cost capital, and now they are finally starting to withdraw funds…
In the coming period, the price will maintain a volatile range at a high level, with retail traders battling each other and gradually closing remaining long positions at the bottom.
The only uncertainty is what these market makers will do with the funds they have gathered. If they continue to play before unlocking, hedging in this range still carries risks.
If the current manipulation ends and it's time to distribute profits, then hedging can be done with peace of mind…
In other words, whether Bill will gradually decline before unlocking on 10-31 or have a second short squeeze entirely depends on the market makers. They have the money and choices; we have to hold positions for 6 months with no options.
This is a completely asymmetric game. Except for those who know the market makers' specific strategies, there is no solution…
Therefore, I think it's best to keep watching. Later, we can pay some attention to technical signals. Once a high-level range consolidation forms, a second short squeeze is possible. Only a continuous slow decline can indicate that market makers have completed their task and exited…
Although the price might be even lower than now at that time, my requirement isn't high—above 0.06 is fine…
Still choosing to keep watching, will check again in a week!"
Apple Glasses positioning has actually been attempted by many companies before. Personally, I believe the key should be the real-time connection with a personal AI assistant or local Agent.
Using image, voice, and gesture inputs to upgrade interaction with AI might be the main theme for the next decade.
After integrating the state machine, the ASR trend was successfully combined with the swing logic, and the current optimization effect is significant!
The returns have slightly decreased, but the maximum drawdown has decreased even more. Simply put, this means the trend strategy no longer rigidly holds from start to finish, but instead continuously performs T trades during the holding period, then reconnects...
Will it miss some selling opportunities? The answer is yes, but it's important to understand that the drawdown of trend strategies often comes from the wear caused by choppy markets. With this logic, even if the trend strategy doesn't catch the trend, the mid-term take-profit or T trades can just offset some of the stop-loss drawdown, making the overall equity curve a bit smoother!
Finally, there's the issue of overfitting. This optimization process inevitably involves fitting, but whether it is overfitting or not will be the key point to investigate next...
CryptoPainter
While researching new strategies for the Agent and configuring a pure algorithmic state machine, I suddenly realized this mechanism could also be applied to the previous ASR strategy. So I hurriedly started modifying the code, and then I was moved to tears...
An old strategy that hadn’t been successfully optimized forward for a whole year suddenly came back to life!
As for the specific changes, it’s just using the state machine to record market volatility in real time, then fine-tuning the volatility into the original strategy’s parameters. Altogether, it’s less than 20 lines of code, but it made the original ASR channel differ in many subtle details...
The overall return of the 5-year BTC strategy improved by over 75%, while the maximum drawdown decreased by 14%!!!
Previously, when researching pure algorithmic quantification, I looked down on state machines. Only when it truly produced positive optimization effects did I realize how great it is...
I guess an updated version can be released soon!
It’s really been so tough...
Wow wow wow wow wow! Agent has turned a loss into a profit in one go!
This performance really feels like a gambler’s move. The 22 short positions on altcoins have already been gradually taken profit on by him, because he detected a huge volume spike in a short time, so even if the take profit wasn’t hit, he still exited…
It ran for a whole week, with big losses and small gains, accumulating nearly a 10% loss, but tonight Trump made a sudden comeback, recovering everything and even earning 2 points…
Sigh, just two hours ago I was still hoping for a big drop on Friday…
CryptoPainter
Spent 2 hours buying a domain and made a Dashboard on the website with all Agent's trading data...
Today is a bit interesting, the strategy positions are all short, let's see if there's hope to dump some altcoins on Friday. Currently, Agent's overall account has already lost 8%, sigh...
If there's anyone to blame, it's me for constantly changing features and fixing bugs every day. Previously, I optimized strategies using backtest data, but this time I want to try pre-testing with live trading data...
The former has perfect backtest data but performs poorly in live trading; the latter performs poorly in live trading but with continuous optimization, the strategy is getting better and better...
I won't disclose the website domain for now, I'm a bit worried about security measures. I'll wait until Claude audits it for me and the strategy starts making profits before saying more...
Spent 2 hours buying a domain and made a Dashboard on the website with all Agent's trading data...
Today is a bit interesting, the strategy positions are all short, let's see if there's hope to dump some altcoins on Friday. Currently, Agent's overall account has already lost 8%, sigh...
If there's anyone to blame, it's me for constantly changing features and fixing bugs every day. Previously, I optimized strategies using backtest data, but this time I want to try pre-testing with live trading data...
The former has perfect backtest data but performs poorly in live trading; the latter performs poorly in live trading but with continuous optimization, the strategy is getting better and better...
I won't disclose the website domain for now, I'm a bit worried about security measures. I'll wait until Claude audits it for me and the strategy starts making profits before saying more...
G... Genetic algorithms really aren't suitable for running locally, the alien laptop broke down...
It was originally used as a development environment for Lobster, didn't expect this...
Luckily, all projects are synced to GitHub, will see how to migrate them to the server to run later...
So exhausting... Now there's really no way to update...
CryptoPainter
Someone asked me why I haven't updated that "genetic algorithm" engine?
I actually want to update it!
After setting up the framework and perfecting the features, the next step is full-scale training. But the problem is, look at the log below: training one tree requires 80 iterations, nearly 1 hour. There are 6 trees in one Forward window, and the entire sample set is divided into 6 windows for Walk Forward...
1h x 6 x 6 = 36h, and this is only the first layer of volume-price data training. After that, it moves on to the macro data layer, and finally the futures, on-chain, and other Alt-Data layers. Running all of this probably takes about 3 days...
And the most, most, most critical issue is that, when luck is bad, the factors randomly injected into the factor library fail to hybridize into excellent expressions, resulting in the current screen full of "Failed" statuses. This means the trained factors perform well in-sample but fail out-of-sample...
These are basically overfitted factors and get directly PASSED...
For a whole week, I've been stuck in this tedious loop, feeling like my alien laptop is going to be wrecked sooner or later...
So, it's really not that I don't want to update, it's just that I honestly have nothing to update...
