Amber安伯

Onchain Hunter,AI、量化、DeFi,跑回测、看盘面、记踩坑。活着才有下一轮
Open Trade
BNB Holder
BNB Holder
Occasional Trader
2.4 Months
0 Following
0 Followers
0 Liked
Posts
Portfolio
·
--
Binance has launched perpetuals on U.S. stocks. The most hassle-free strategy seems to be to enter at the open and exit at the close—no holding overnight. I backtested the daily spot charts for 6 of them from November 2024 to this November. Doing that, all 6 lost money. I worked out the fees: 0.2% per round trip. It looks like very little, but do that every day for around 250 trading days a year, and fees alone eat up half your position. The wildest part is that with the same stocks, trading less and holding them longer instead actually made over 60% in two years for some of them. Dodged the overnight risk, but not the fees 😂 $BTC $BNB #美股永续 #回测
Binance has launched perpetuals on U.S. stocks. The most hassle-free strategy seems to be to enter at the open and exit at the close—no holding overnight.

I backtested the daily spot charts for 6 of them from November 2024 to this November. Doing that, all 6 lost money.

I worked out the fees: 0.2% per round trip. It looks like very little, but do that every day for around 250 trading days a year, and fees alone eat up half your position.

The wildest part is that with the same stocks, trading less and holding them longer instead actually made over 60% in two years for some of them.

Dodged the overnight risk, but not the fees 😂

$BTC $BNB #美股永续 #回测
Whenever a big-name influencer posts a tweet, a bunch of coins with the same name pop up immediately. Everyone wants to be first. I looked through the past 60 days of data on four.meme: more than 5,000 coins were tied to trending topics like Musk, CZ, Trump, and SpaceX. In the end, only 11 made it through the bonding curve and successfully graduated. Put another way, only 2 out of every 1,000 survived. Most of the people who launched the other 998 coins only lost around 0.1U in gas. The craziest example was the day Musk changed his name: 157 related coins popped up in a single day, and not one graduated. They got the name, but not the buyers 😂 $BNB $BTC $ETH #Meme #OnChainData
Whenever a big-name influencer posts a tweet, a bunch of coins with the same name pop up immediately. Everyone wants to be first.

I looked through the past 60 days of data on four.meme: more than 5,000 coins were tied to trending topics like Musk, CZ, Trump, and SpaceX.

In the end, only 11 made it through the bonding curve and successfully graduated.

Put another way, only 2 out of every 1,000 survived. Most of the people who launched the other 998 coins only lost around 0.1U in gas.

The craziest example was the day Musk changed his name: 157 related coins popped up in a single day, and not one graduated.

They got the name, but not the buyers 😂

$BNB $BTC $ETH
#Meme #OnChainData
A backtest curve that looks great might be the cheapest thing in crypto. I came across a framework that put it plainly: train on old data first, then test it on the two years it hasn’t seen. Getting every homework question right doesn’t count—the exam is what matters. Then, before going live, put it through a few more checks: If it only makes money in one bull run, don’t use it; If a slight parameter tweak makes it fall apart, don’t use it; If all the profits come from just a few days, don’t use it; If fees and slippage aren’t accounted for, don’t use it either. After filtering them out one by one, many curves that looked like a ticket to financial freedom are left with nothing but screenshot-worthy value for your social feed 😂 #QuantitativeTrading #AI#btc
A backtest curve that looks great might be the cheapest thing in crypto.

I came across a framework that put it plainly: train on old data first, then test it on the two years it hasn’t seen. Getting every homework question right doesn’t count—the exam is what matters.

Then, before going live, put it through a few more checks:
If it only makes money in one bull run, don’t use it;
If a slight parameter tweak makes it fall apart, don’t use it;
If all the profits come from just a few days, don’t use it;
If fees and slippage aren’t accounted for, don’t use it either.

After filtering them out one by one, many curves that looked like a ticket to financial freedom are left with nothing but screenshot-worthy value for your social feed 😂

#QuantitativeTrading #AI#btc
The most widely practiced skill in crypto is called flipping through the calendar. The halving happens, wait 12 to 18 months, the market tops, get out. More people have memorized this mantra than have memorized their multiplication tables. Today I came across a summary of Fu Peng’s views. The gist was: this time window still bears some resemblance to previous cycles, but the people driving prices have changed. Who used to drive prices? The halving meant fewer new coins, and with a smaller market, a few whales could move prices by buying and selling. Who drives prices now? Institutions, ETFs, and money from the broader macro environment. The market is bigger, so whales can’t move it as easily, and volatility has narrowed too. It’s all pretty abstract. The calendar is still the same calendar, the mantra is still the same mantra, the timing of the top is still roughly the same, but the people holding the calendar aren’t the same crowd as before. I’ve been thinking about it, and the trickiest scenario isn’t that the calendar stops working. If it does, people will naturally change their approach. The trickiest thing is if it happens to be right again this cycle, making everyone believe in it even more, while the money driving prices next cycle doesn’t look at the calendar at all. When that happens, what exactly are you holding in your hands—a pattern, or a stopped clock that just happens to have stopped at the right time
The most widely practiced skill in crypto is called flipping through the calendar.

The halving happens, wait 12 to 18 months, the market tops, get out. More people have memorized this mantra than have memorized their multiplication tables.

Today I came across a summary of Fu Peng’s views. The gist was: this time window still bears some resemblance to previous cycles, but the people driving prices have changed.

Who used to drive prices? The halving meant fewer new coins, and with a smaller market, a few whales could move prices by buying and selling.
Who drives prices now? Institutions, ETFs, and money from the broader macro environment. The market is bigger, so whales can’t move it as easily, and volatility has narrowed too.

It’s all pretty abstract.
The calendar is still the same calendar,
the mantra is still the same mantra,
the timing of the top is still roughly the same,
but the people holding the calendar aren’t the same crowd as before.

I’ve been thinking about it, and the trickiest scenario isn’t that the calendar stops working. If it does, people will naturally change their approach.
The trickiest thing is if it happens to be right again this cycle, making everyone believe in it even more, while the money driving prices next cycle doesn’t look at the calendar at all.

When that happens, what exactly are you holding in your hands—a pattern, or a stopped clock that just happens to have stopped at the right time
Gave the strategy its own macro weather bureau, and the annual backtest results got pretty surreal Press-release version: Smaller drawdowns, smoother equity curve Real version: More than a third less profit The most memorable moment was April 6, 2025: the weather bureau warned that the skies were about to change, so we liquidated everything with one click. Over the next 90 days, $BTC rose by nearly 29%, while the version using it actually lost over 3%. It didn't rain, but the umbrella had already cost us money Even more absurdly, ditch the macro stuff and switch to always being 50% invested—the drawdown is just as small Giving up a third of the profit for peace of mind sounds like a pretty good deal Until you realize that being 50% invested gives me the same peace of mind, and it's free 😂 #BTC #量化交易
Gave the strategy its own macro weather bureau, and the annual backtest results got pretty surreal

Press-release version: Smaller drawdowns, smoother equity curve
Real version: More than a third less profit

The most memorable moment was April 6, 2025: the weather bureau warned that the skies were about to change, so we liquidated everything with one click. Over the next 90 days, $BTC rose by nearly 29%, while the version using it actually lost over 3%. It didn't rain, but the umbrella had already cost us money

Even more absurdly, ditch the macro stuff and switch to always being 50% invested—the drawdown is just as small

Giving up a third of the profit for peace of mind sounds like a pretty good deal
Until you realize that being 50% invested gives me the same peace of mind, and it's free 😂

#BTC #量化交易
Over the past couple of days, $BTC has swung more than $5,000 a day, while $ETH and $BNB have also dropped over 3%. Watching it makes me itch to buy the dip, but I figured I’d sit down and think it through a bit more. My trend-following strategy, to put it bluntly, is a clueless rookie that only knows how to read candlestick charts. It has no idea whether the Fed is pumping money into the system or pulling it out. It chases prices when they rise and runs when they fall, completely disconnected from the world outside. So I want to give it a weather watcher. When there’s plenty of money around, it can open more positions; when there’s less, it opens fewer. If something really serious happens, it pulls the plug and closes everything. The strategy still makes the buy-and-sell decisions—I just control how much it bets each time. Halfway through writing the rules, the funniest thing happened. I’d been thinking of adding a buy-the-dip switch, so it could charge in when the Fed makes an emergency rate cut or starts pumping money into the system again. But after combing through the Fed’s website, I found that from 2018 to now, this has only happened once: during the pandemic in March 2020. You can’t build a whole rule around a single example. No matter how pretty the backtest looks, it’s still just a guess. In the end, I made one rule absolutely clear: if it hasn’t shown up in new data, you can’t use it. There’s another pretty ridiculous trap, too. Most of the time, this weather watcher tells me to open fewer positions. If the backtest then shows smaller drawdowns, is it actually good at reading the weather, or is that simply because my positions were smaller? So I added two dumb benchmarks for comparison: always invest 70%, or always invest 50%. If it can’t beat either of those, it’s just putting on a magic show. I also ran the current conditions through these rules. The reading came out cautious: I should only invest 50%. Neither the emergency shutdown nor the buy-the-dip rule was triggered. But I haven’t run the backtest yet, and I haven’t changed anything in my portfolio—just watching for now. After spending so long in crypto, I’ve come to realize that I used to think the more rules you had, the smarter you were. Now I increasingly feel that most rules are just stories we tell ourselves. It’s like buying an insurance policy that only pays out once in ten years. You can’t tell whether the insurance is really any good, or whether you’ve just been lucky.
Over the past couple of days, $BTC has swung more than $5,000 a day, while $ETH and $BNB have also dropped over 3%. Watching it makes me itch to buy the dip, but I figured I’d sit down and think it through a bit more.

My trend-following strategy, to put it bluntly, is a clueless rookie that only knows how to read candlestick charts. It has no idea whether the Fed is pumping money into the system or pulling it out. It chases prices when they rise and runs when they fall, completely disconnected from the world outside.

So I want to give it a weather watcher. When there’s plenty of money around, it can open more positions; when there’s less, it opens fewer. If something really serious happens, it pulls the plug and closes everything. The strategy still makes the buy-and-sell decisions—I just control how much it bets each time.

Halfway through writing the rules, the funniest thing happened. I’d been thinking of adding a buy-the-dip switch, so it could charge in when the Fed makes an emergency rate cut or starts pumping money into the system again. But after combing through the Fed’s website, I found that from 2018 to now, this has only happened once: during the pandemic in March 2020.

You can’t build a whole rule around a single example. No matter how pretty the backtest looks, it’s still just a guess. In the end, I made one rule absolutely clear: if it hasn’t shown up in new data, you can’t use it.

There’s another pretty ridiculous trap, too. Most of the time, this weather watcher tells me to open fewer positions. If the backtest then shows smaller drawdowns, is it actually good at reading the weather, or is that simply because my positions were smaller? So I added two dumb benchmarks for comparison: always invest 70%, or always invest 50%. If it can’t beat either of those, it’s just putting on a magic show.

I also ran the current conditions through these rules. The reading came out cautious: I should only invest 50%. Neither the emergency shutdown nor the buy-the-dip rule was triggered. But I haven’t run the backtest yet, and I haven’t changed anything in my portfolio—just watching for now.

After spending so long in crypto, I’ve come to realize that I used to think the more rules you had, the smarter you were. Now I increasingly feel that most rules are just stories we tell ourselves.

It’s like buying an insurance policy that only pays out once in ten years. You can’t tell whether the insurance is really any good, or whether you’ve just been lucky.
·
--
Bearish
That guy from 48Club lost over 90 million U on a 30-day contract, and the extra 13.5 million he added to his position went down the drain too 😅 Over the past two days, $BTC fell more than 5% in a day, to around 64,000. $ETH and $BNB also dropped more than 6%. I have just one rule: as soon as the trend weakens, I move everything into U. No buying the dip, no adding to my position. My simulated portfolio is now all cash across all three tiers, and I haven't opened a single trade. Backtesting shows returns of only about 0.7% a month—not much, admittedly—but the maximum drawdown is just over 10%, so I can sleep at night. Holding onto losing positions and adding to them is like borrowing money at a gambling table to win back your losses. Win once and you think you're a gambling god; lose once and you're out of the game 😂
That guy from 48Club lost over 90 million U on a 30-day contract, and the extra 13.5 million he added to his position went down the drain too 😅

Over the past two days, $BTC fell more than 5% in a day, to around 64,000. $ETH and $BNB also dropped more than 6%. I have just one rule: as soon as the trend weakens, I move everything into U. No buying the dip, no adding to my position. My simulated portfolio is now all cash across all three tiers, and I haven't opened a single trade.

Backtesting shows returns of only about 0.7% a month—not much, admittedly—but the maximum drawdown is just over 10%, so I can sleep at night.

Holding onto losing positions and adding to them is like borrowing money at a gambling table to win back your losses. Win once and you think you're a gambling god; lose once and you're out of the game 😂
Log in to explore more content
Join global crypto users on Binance Square
⚡️ Get latest and useful information about crypto.
💬 Trusted by the world’s largest crypto exchange.
👍 Discover real insights from verified creators.
Email / Phone number
Sitemap
Cookie Preferences
Platform T&Cs