After Earnings Season: A Review of My Chart Analysis Process

Why did the price drop when the earnings results weren't bad?

Since the latter half of July, I have been receiving this question repeatedly.
I was looking at the same thing.

The numbers were there.
Yet, the price moved in the opposite direction.

Months like this test the very way you conduct your review.

This time, I will show you the exact review procedure I always follow at the end of earnings season.

Today's Goals

・Understand why you should stop grading your reviews based on whether you were right or wrong.

・Take home a procedure for grading your decisions by breaking them down into four steps.

・Learn how to narrow down to one specific area to fix for the next season.

1. Grade reviews based on the process, not the results

In the consultations I receive, I often see reviews that end with two columns: 'this prediction was right' and 'this one was wrong'.

Since profit and loss are undeniable numbers, this seems like a clear way to grade.

However, this method has side effects.

The point is that decisions that were right are marked as successful without ever questioning why they were right.

If a trade entered with weak reasoning happens to grow, it goes into the 'right' column, while a trade where you followed the procedure and exited at the expected level falls into the 'wrong' column.

If you continue this for a few months, all you will have left is the way of doing things that relied on luck.

What I grade is not the result, but the process leading up to the decision.
If the process was followed and you lost, I mark it as a pass as an expected loss.

If you skipped the process and won, I mark it as a fail even if the amount is positive.
If you don't grade this way, you will be left with nothing the moment the market conditions change.

2. What happened in July

July was a month where the 'quality of the news' and the 'direction of prices' did not connect in a straightforward way.

First, let's confirm the facts with numbers. The closing price of the Nikkei Stock Average was 70,474.96 yen on July 1st.

From there, it continued to 68,751.51 yen on July 15th, 66,835.54 yen on July 16th, and 64,141.12 yen on July 17th, with a decline of 1,915.97 yen from the previous day on the 16th, and a decline of 2,694.42 yen on the 17th.

That is a loss of over 4,600 yen in two business days.

However, on July 21st, after the long weekend, it saw a sharp rebound to 66,232.19 yen, an increase of 2,091.07 yen from the end of the previous week.

It did not fully recover after that, and the closing price on July 28th was 62,364.92 yen, down 2,566.27 yen from the previous day (all figures are closing prices published by Nikkei Inc.).

Compared to July 1st, this is a decline of 8,110.04 yen, or 11.5%.

July 28th was also the lowest price since May 21st.

The background summarized in reports included the decline in U.S. semiconductor stocks the previous day, a significant drop in South Korean stocks, and reports that a Chinese state-owned enterprise had begun manufacturing semiconductor manufacturing equipment (lithography equipment).

Furthermore, it was noted that this was a 'decline after major companies had all announced strong earnings,' and that supply and demand had also worsened due to a series of new listings and large-scale capital increases (Kabutan Market Daily Report, July 28, 2026).

In summary, this was a phase where bad news from the outside, expectations that had already been priced in at a high level, and supply and demand that had become heavy due to capital increases and the like, all overlapped at the same time.

The point here is that the decline did not happen because the content of the earnings reports had worsened.

Note that not all first-quarter earnings reports for companies with a March fiscal year-end have been released yet.

The Tokyo Stock Exchange states that 'in principle, first-quarter earnings reports should be disclosed within 45 days after the end of each quarter,' and the deadline for companies with a fiscal quarter ending in late June is August 14th (JPX FAQ for listed companies).

This is not a summary of the earnings season, but an interim review at the point when the July data has been mostly released.

3. The four steps I use for scoring

I break down my monthly judgments into four steps: scenario, level, volume, and record, and score them one by one.

I settled on these four because the reasons I got stuck during reviews were mostly one of the following: 'I didn't write it down in the first place,' 'I didn't turn it into a number,' 'I didn't decide on the volume,' or 'I didn't keep a record.'

I don't judge based on the skill of the method, but rather on whether or not it was decided in advance.

That is the characteristic of this classification, and that is why I can see specifically where things fell apart.

Step 1: Did I write scenarios for both upward and downward directions?

I check whether I wrote down both the 'reasons to go up' and 'reasons to go down' before the earnings announcement.

If I only wrote one side, it means I was hoping rather than analyzing.

If I only prepared a one-way scenario that 'it will go up if the earnings are good,' there is no way to respond to a drop like the one on July 28th.

When I write an upward scenario, have I also written the conditions under which it would collapse?

This is the pass/fail criteria for Step 1.

Step 2: Was the exit level written down in numbers before entering?

I check if it was set as a price, not just 'if it drops a bit'.

It can be a moving average or a recent low; it doesn't matter.

The important thing is that it is written down as a number before entering.

Regarding July 28th, reports also noted that it had fallen below the 75-day moving average.

If you set such levels in advance, your decision on that day is just a matter of 'verifying' rather than 'deciding'.

If you haven't set them, you end up creating criteria while the price is falling.

Step 3: Was the amount to hold through earnings decided before holding through them?

Whether or not to hold a position through earnings should be determined by pre-established rules, not by how you feel on the day.

What I look at here is not whether I held through it, but whether I decided the amount to hold through before doing so instead.

In situations like earnings where prices jump all at once, there is no room to make decisions along the way.

Since the result is almost determined at the opening, it is the size of the position, not the accuracy of the decision, that determines the range of profit or loss.

In a situation like July where the price moves 4,600 yen in two business days, this difference becomes the difference in your account balance.

Step 4: If there is no record, you cannot grade it in the first place.

Finally, there is record-keeping.

I look at three points: what time it was, what I was looking at, and how I made my decision.

It is ideal if I also record how I was feeling at that time.

If there is no record for a month, I cannot conduct a review at all.

When looking back based on what one can remember, people tend to edit things conveniently.

I sometimes use generative AI for this.

If I have it read the decision notes I have accumulated and ask it to "extract parts where I am hesitating using the same phrasing," it returns habits I would not have noticed myself.

This is a way of using it to help with the task of reading back through records, rather than entrusting the decision itself to it.

4. How to handle "stocks falling despite good earnings" in the process

What became clear again in the July market is that earnings results and stock price direction are not on the same layer.

Earnings are on the layer of "facts."

Stock prices move on the layer of "to what extent those facts were already priced in" and "who was in a position to trade at that time."

In my review, I score these two separately.

One is the reading of the content, which is whether my reading of the earnings figures and the direction of the business was correct.

The other is the reading of the reaction, which is whether my reading of how the market would react to that content was correct.

It is not rare to have a result where the content reading was correct, but the reaction reading was wrong.

If I write them together as "I got it wrong," I will not know what to fix next time.

If the reading of the content is wrong, I fix the research procedure.

If the reading of the reaction is wrong, I fix how I view pricing-in and supply-demand.

5. Only fix one thing for the next season

When I do a review, I usually find three or four things I want to fix.

Even so, I narrow it down to just one thing each time.
This is because if I change multiple things at once, I cannot verify the results.

If I change how I write scenarios, how I set exit levels, and my position sizing all at once and my performance improves the next season, I won't know what actually worked.

That is why I only fix one step at a time and focus only on that one point for the next season.

It is a modest approach, but it is the only way to solidify a process.

6. Bridging to "Your Case"

This concludes the general theory of the review framework.

The four-step framework is the same for everyone, but the values you put inside will change based on your own time horizon and methodology.

Whether you set your exit level based on a moving average, the most recent low, or a price range. How much of your account you allocate to positions held through events. How much detail you record (if you design it to be too detailed, you won't be able to keep up with the recording itself). Whether you conduct reviews season by season or month by month.
These things depend on your holding period and the amount of decline you can tolerate, so taking someone else's numbers and using them as-is will not work.
You should have been able to take the framework home with you this time, so the next step is to fill in the content for yourself.
Summary

These factors vary depending on your holding period and acceptable drawdown, so simply adopting someone else's figures will not work.

You should have been able to take the framework home with you this time, so the next step is to fill in the content for yourself.

Summary

• Evaluate your review based on the decision-making process, not the results.

• In situations where prices fall despite good earnings, evaluating based on results is mostly just evaluating luck.

• The four steps to evaluate are: 1) Did I write scenarios for both upward and downward directions? 2) Did I set the breakdown level with a number in advance? 3) Did I decide the amount to hold through the event beforehand? 4) Is there a record of my judgment?

• Only fix one step at a time.

Earnings season is the time when hits and misses are most visible.

That is precisely why whether or not you end your review of this period based on results will change your accuracy for the following months.

In closing

The framework of breaking things down into steps for evaluation can be reused in any market environment.

On the other hand, the specific figures for setting exit levels and determining position sizes for holding through events change completely depending on your time horizon and methodology.

Therefore, for those who want to be able to structure this review process to fit their own methods, I am sending a guide via LINE on how to systematically learn how to proceduralize chart analysis and design reviews.

On LINE, I am also gradually providing review sheet items that could not fit into the article, as well as operational notes from the recent market.

Registration is free.

You can unsubscribe at any time if it does not suit you.

※This article is for informational and educational purposes only and does not constitute a recommendation to buy or sell specific financial products, nor does it guarantee profits or principal.

The stock prices and statistical values posted are based on publicly available data as of the time of writing (July 29, 2026).

Please make investment decisions based on your own circumstances.