Algorithm system

Hard core analysisLin Yuan loaded the entrances of multiple stock forums into the system, hoping that the system could train a short-term AI model for sniping Good Group.

Lin Yuan did not expect to become a stock market god with the system; it wasn’t that he didn’t want to, but rather that it didn’t align with the system’s logic.

Even if Lin Yuan hadn’t tried, he was certain that such an operation would not conform to the system’s settings.

His judgment was entirely derived from the perspective of AI training.

The two primary points for AI training are: sufficient sampling parameters and ample sampling data samples.

This is an objective law determined by the fundamental principles of current AI training.

Unless these fundamental principles can be broken, even the Trisolarans would have to abide by this law.

Since Lin Yuan acquired the system, he has fully realized that his system is a super tool within the realm of scientific laws; there are no ’one-click instant’ supernatural abilities.

Therefore, starting from the two major fundamental elements of AI training,

expecting the computing power system to train a universal short-term operation AI model applicable to all stocks is unfeasible. The reason is simple: it gets stuck on the point of ’sufficient sampling data’.

This is essentially the same reason why Good Group couldn’t develop a food delivery algorithm; in both cases, the sampled data parameters did not sufficiently cover the problem itself.

Simple example: too many factors influence a stock’s price.

Regional economic environment, global economic situation, economic cycle fluctuations, climate factors, cultural factors, religious and political factors, and even unexpected black swan events.

Even these overtly visible factors are already complex enough, so complex that Lin Yuan doesn’t know how to guide the system to complete the modeling.

Not to mention the invisible factors; who knows how those unseen big hands behind the scenes will manipulate stock price fluctuations like kneading playdough.

So, given that, why did Lin Yuan still want to use the computing power system to snipe Good Group?

Because if you only want to know part of the answer to a problem, you don’t need to understand all the unknown parameters in the problem when solving it.

To explain this point, mathematical formulas are far more convincing than words.

Suppose: F(x) = f1(x) + f2(x);

Given 1: F(0) = f1(0);

Given 2: F(1) = f1(1) + 2;

Given 3: F(2) = f1(2) + 4;

Then, through forced fitting, it’s easy to analyze: f2(x) = 2x;

Even if you don’t know what f1(x) is, you can still fit the formula for f2(x).

The three given conditions above are equivalent to:

1, f2(0) = 0;

2, f2(1) = 2;

3, f2(2) = 4;

Looking at it this way, it’s very simple. So why mix f2 and f1 together?

Yes, why?

Once the formula is laid out, everyone will ask this question—why?

But when the formula is hidden beneath the mists of reality, people cannot glimpse its existence.

Just like the situation before Lin Yuan’s eyes—f1(x) represents the original trend of Good Group’s stock, the trend without Lin Yuan’s little butterfly flapping its wings and interfering. And f2(x) is the variable caused by Lin Yuan, this little butterfly.

For f1, it is impossible for current human technology to derive it. Because f1 represents the regular trend of a stock.

However, f2 can be derived.

Although f2 is still too complex for an ordinary person, for the computing power system, it is as simple as a linear equation like f(x) = 2x.

To explain this point, we still need to return to the two major elements of AI training: sufficient sampling parameters and ample sampling data samples.

The inability to derive f1 is precisely because it falls short on the point of insufficient sampling parameters.

But f2 has no such problem at all.

Because f2 represents the interfering factor caused by Lin Yuan, this little butterfly.

This interfering factor is naturally Good Group taking advantage of the opportunity to manipulate the stock price.

Such cases are abundant in human stock history, providing plenty of data for the computing power system to reference.

Therefore, the computing power system can definitely calculate f2.

But is this enough? Lin Yuan knew very well that it certainly wasn’t.

Having only historical data of stock manipulation can only allow the system to analyze the general routine of such manipulation, but it cannot map this routine to the current stock price of Good Group.

The root cause is the inability to accurately separate the factors influencing the stock price.

Simply put, it’s impossible to accurately divide F(x) into: f1(x) and f2(x).

Because, how do you know which stage of the stock price is a natural evolution and which stage is subject to manipulation?

Only after a wave of harvesting is completed for each stock, and then analyzing it in hindsight, can one roughly discern the traces of manipulation. When the manipulation is happening, people cannot perceive it. Otherwise, why would the market maker bother to manipulate; they would surely choose to leave quietly.

Therefore, if you cannot clearly know the current cycle of a stock, then even if the AI model trained by the computing power system is powerful, it will be greatly discounted.

Because you simply don’t know when f2(x) should take effect.

In layman’s terms: you don’t know when the current stock started being operated, you don’t even know which This data to feed into the AI model. That is to say, you don’t even know what value to fill for the parameter x.

But Lin Yuan not only knew Good Group’s true intention but could also analyze Good Group’s approximate timing of action from the food delivery algorithm’s test iteration records.

The reasoning behind this is simple: Good Group’s move must have been made after confirming the effectiveness of Lin Yuan’s food delivery algorithm; otherwise, why would they first suppress and then raise, self-slandering would turn into truly digging their own grave.

And Lin Yuan was fully involved in the food delivery testing work; he knew all the time nodes like the back of his hand.

Then, combining these time nodes with a comparison of Good Group’s stock price trend, everything was pretty much spot on.

So, the system could get to work.

Even with a powerful computing power system, to use it well, one still needs a solid theoretical foundation.

Lin Yuan once again achieved flexible use of the system.

So, after having the system load the entrances of various major stock forums on the internet, he began to let the system train.

Then, his world fast-forwarded another two hours...

These two hours were still because Xia Ye discovered him in time after returning from class; otherwise, who knows how long Lin Yuan would have been unconscious.

This made Lin Yuan feel that the system seemed to find being used for stock trading incredibly difficult to understand, more so than being used for mining.

At least the essence of mining was still blockchain computation, still somewhat related to cutting-edge technology. But this time, it was purely brute-force fitting training of stock market data. It was truly a case of ’when the map is unrolled, the dagger is revealed,’ no pretense at all.

Xia Ye was frightened by Lin Yuan’s fainting. She forced Lin Yuan to lie down on the bed. She believed this was because Lin Yuan hadn’t recovered from the recent ordeal of being in custody.

Lin Yuan leaned against the headboard, listening to Xia Ye’s busy sounds in the kitchen, and silently complained to the system with lip-syncing.

“System, you made me stew in there for over a week.”

“Even if you don’t compensate me, you should at least admire my astonishingly brilliant moves. Even if you are the sharpest sword, you still need a swordsman who knows how to wield you.”

“And this swordsman now needs some travel expenses to roam the jianghu. Please help out!”

The system remained silent, and Lin Yuan didn’t expect it to respond. This was just his frustrated rambling.

“How can I make you comply?”

“Come, drink it while it’s hot.”

A fragrant chicken soup aroma permeated the air. Hands that once never touched spring water now didn’t even get manicures.

“I’ll go cook for you,” Lin Yuan said, about to get out of bed.

“You lie down quickly, you scared me to death just now. Are you really not going to the hospital for a check-up?”

“No need.” Lin Yuan was really afraid that a CT scan would reveal a small black block in his head.

Lin Yuan took the chicken soup, drank two sips, and only then did Xia Ye return to the kitchen to continue her work.

“System, if you don’t help, I’m really going to get an X-ray. Don’t force me to expose you.”

But the system still didn’t respond.

“You’re so fearless.”

The delicious chicken soup, filled with love, warmed Lin Yuan’s chest and abdomen. He suddenly saw the long-lost task list on the system panel—publish an SCI paper.

Lin Yuan gave a wicked smile, ’Heh, you forced my hand!’