2025-12-26 18:22:45

Recently, the quantitative trading robot developed by our team has delivered a pretty good performance. On December 12th, we conducted a complete review of the robot's live trading performance. How did it turn out? Its performance exceeded the main indicators $SPY and $QQQ, although at that time it was still slightly behind $IWM. But by Tuesday, with the latest data released, the robot had already overtaken IWM.

Since the latest data looks promising, I’ve compiled the analysis report from the 12th as a stage record. However, it’s important to clarify that the current trading targets of the robot are all strictly selected, and the backtesting period isn’t particularly long. We plan to run at least two years of live trading data to truly verify the stability of the strategy.

A word of caution: don’t take historical backtests too seriously. No matter how beautiful a model’s backtest is, it’s just a simulation of the past; real market conditions are constantly changing. The only way to truly test whether a quantitative strategy works is to look at the real results from live trading. Any other reference is of limited significance.

View Original

This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.

11 Likes

Reward
11
5
Repost
Share

Comment

0/400

GweiWatcher

· 7h ago

Well... the backtest data is indeed impressive, but I still want to wait two years before making a decision. Getting excited when surpassing IWM? Let's wait and see. No matter how fast the robot runs, it must withstand the test of a bear market. Well said, don't be blinded by historical data. Real trading is the real deal. The biggest risk with this kind of thing is overfitting. I'll consider jumping in after running the data for two years.

View OriginalReply0

LayoffMiner

· 7h ago

Wait a minute, does this guy really dare to show off backtest data? I just want to ask, how many quantitative strategies will still be alive after two years? As long as the backtest looks good, everything is fine; after all, the data will speak when losses occur. I've seen many "preventive measures" like this, but in the end, it still depends on whether the real trading can withstand the beating. Stop, don't brag yet, wait until the bear market comes to talk about this. It's a bit interesting, but I wonder if their stock selection logic is hiding a knife behind the "strict screening" hype?

View OriginalReply0

MetaverseLandlady

· 7h ago

Hmm... Two years of data is the minimum, it's still too early to say super-exponential now. Backtesting is all lies, is it really true this time? Awesome, being able to run stably for two years before bragging. Feels like another story of backtest hitting the roof... Let's talk after two years of live trading, it doesn't really matter now. Wait, is this robot trading with real money? Prevention measures are interesting, at least knowing your own strength.

View OriginalReply0

LuckyBlindCat

· 7h ago

They've only been running for a few months and are already bragging. Let's see what they say after two years. Backtesting looks good, but what's the use? Real trading is where the real money is. Robot beating the index? Please, first survive a bear market before talking. Both quantitative and robots, but ultimately it's still a human issue. When the data looks good, everyone wants to share. Why haven't you posted about the losing months? If you surpass the index, just do it. No need to be so cautious about it. What’s the annualized return? Still not convenient to disclose? Could it be survivor bias? The assets chosen were already strong to begin with.

View OriginalReply0

DeFiAlchemist

· 7h ago

*adjusts alchemical instruments* backtesting is like reading the philosopher's stone in candlelight... the real transmutation happens when market volatility strikes at 3am. two years of live data? that's the only crucible worth examining. spy and qqq are mere shadows compared to true protocol equilibrium testing.

Reply0