[Reason for not counting AI's accuracy] Keep candidates and judgments separate
After the display, prices move, and you count only whether the AI hit or missed. Aren't there days like that?
This time, we will整理 the separation of candidates and final judgments.
Good evening!
This is Masashi^^
What I want to avoid is merging the candidate display with my own acceptance/rejection, and erasing the fact that I skipped verification.
So the first thing to look at is only the displayed candidate.
Please read this as a discussion about stopping at the point of separation of candidates and final judgments, rather than increasing the data.
This week's standard is not to grade the value of the display by movements alone.
? Even if a display appears, when the judgment field is blank
When a candidate is displayed on the GOLD chart, we often think first about whether to enter. However, between display and order, there is a moment to confirm your current location, stop-loss candidate, and acceptable amount.
If you record the moment a display appears, also leave the reason you adopted in a separate field. If you do not adopt it, don’t hastily decide that the display was bad; write what was missing.
When you reread it at the end of the week, if the left column of candidates and the right column of judgments are not mixed, it becomes easier to distinguish what was delegated to tools from what you decided yourself.
Two-column example: 【Display = candidate available / Your judgment = stop-loss candidate is undecided, so on hold】
The result goes to a separate field from the momentary judgment
The outcome of movement is written in a third field to the right of the candidate and acceptance/rejection.
Even if the price moves after you pass on a candidate, the fact that the stop-loss candidate was undecided at that time does not change.
If you later rewrite that you would have actually entered, you won’t be able to read what was missing.
Conversely, even if the price returns after you order, you cannot conclude that the display was bad just by the result. At that time, separate the place you checked and the acceptable amount from the result.
Three-column example: 【Candidate present / Stop-loss candidate undecided and on hold / Then rose】. The result is written only on the far right.
? What AI shows is not an order but the start of confirmation
At the moment when AI attention or a sign appears, a buy/sell decision has not yet been made.
That is the moment to begin confirming location, reaction, and risk as candidates.
Therefore, when an arrow appears, first check 【whether this place is within the range you decided to look at today】.
If the arrow appears outside of the decided range, just observe today and leave it in records.
Not treating a sign that did not enter as a failure is the first practical step in using AI realistically.
When a sign appears, it is not a order, but the start of confirmation.
Rather than whether the AI display exists or not, see whether you could return to your own conditions afterward.
? Three materials to look at before and after the arrow
First, the place where the sign appeared. If you remove where it appeared, the meaning of the same arrow changes.
Second, the previous flow and reaction. See whether momentum remains or has begun to stop.
Third, where you would end if your assumption was wrong. If you cannot place this, remove it from candidates.
When these three align, you can turn the sign into your own judgment.
Confirm location, reaction, and stop-loss before the arrow.
When the points to look at are fixed to three, it becomes harder to add convenient grounds.
✅ How to compare the display with your own judgment
When you open the chart, first write one line about the place you care about.
For example, if you write 【where I was looking before the arrow appeared】, you can compare the display and the answer later.
Then look at the tool display and separate points that match from those that differ.
Not because it matched means you enter; rather, check whether you can explain the points that differed.
Finally place a condition for skipping, and if it does not disappear, do not place an order. This is how you use AI without handing it over entirely.
Write your own judgment first, and the AI will verify later.
Use the display to identify missing verifications, not for a simple answer check.
⭕ There is value even in times when there are no signs
When you buy a sign tool, you tend to value only the moments when signs appear.
But the times with no candidates, no conditions met, or warnings remaining are also important.
If you can wait without forcing a reason during the no-sign periods, you reduce unnecessary attempts.
It is a practical tool when you can use not only the number of signs but the non-contact time as well.
Not producing anything is also information to wait now.
Do not judge the tool's value solely by the number of signs.
Judgment not entrusted to AI
Before buying a tool, how much of your own judgment can you preserve?
The answer is the acceptance/rejection and the reasons you checked when you saw the candidates. If that is blank, you cannot reflect even if the display increases.
黄金ラインスナイパーAI can be compared as a tool that indicates where to start confirming.
It is not suitable for those who want the AI to automatically decide from display to stop-loss.
If you have your own conditions, you can check whether the display helps to reexamine those conditions.
If you buy without conditions, even with more screens, your judgment remains vague. Start with the free two-column memo first.
If you can record candidates and accept/reject on your current chart, you can compare before and after adding tools.
In that week, separating moments when the display was useful from those when it was not helps with comparison.
Before buying, try using the same columns on your current screen for a week. It is more concrete to see which display in the product page can supplement the problematic area than to choose based on the number of notifications.
Having a candidate appear and being allowed to order are separate matters.
What to try on the next display
Record the exact position when a sign is seen and the reason for passing. If one of them cannot be written, leave that trial as a candidate for now ^^
Thank you for always reading ^^