Last time,

“Just because the rejection rate is 70% doesn’t mean you should enter right away”

was what we talked about.

Past statistics are important.

However, statistics do not guarantee the future.

At ACE Gold, we value not only looking at past statistics, but also combining “past statistics” with “current market conditions” to make judgments.

So, how much data should be used for those “past statistics”?

Is it 10,000 bars?

10万本でしょうか。

Or is it 1,000,000, 2,000,000 bars? Is more data always better?

At ACE Gold, for statistical analysis we search about 300,000 candlesticks of the past.

This time, I will explain why we chose this number of “about 300,000 bars” and the thinking behind it.

---

Is more data always better?

When you hear “statistics,”

the impression is that the more data, the more reliable it is

might be.

Indeed, having too little data can be a problem.

For example, suppose there were 10 past instances of a channel similar to the current one.

Among them, it rebounded 7 times.

Even then, numerically,

the rebound rate is 70%

.

Now suppose there were 1,000 similar cases in the past, and 700 rebounded.

This is also

a 70% rebound rate

.

Same 70%, but

7 out of 10 vs 700 out of 1000

are not received the same way.

Therefore, in ACE Gold we also place importance on securing a substantial amount of data.

So,

Is it from 100,000 to 300,000 bars?
From 300,000 to 1,000,000 bars?
From 1,000,000 to 10,000,000 bars?

Should we just keep increasing it as much as possible?

In fact, I don’t think it’s that simple.

Because there is another issue here,

“markets change over time”

.

---