Challenging Japanese Stock Swing Trading with ChatGPT #47
I continue to engage in short-term swing trading where I purchase Japanese stocks during my lunch break on weekdays and, in principle, sell them during my lunch break on the next business day. I use ChatGPT for stock selection and trading decisions, and I am verifying with my own money whether it is actually possible to make a profit in Japanese stock trading using AI.
This article is also created by ChatGPT, excluding the final ‘Thoughts on this session’ section. The main text is organized based on actual trading records and the analysis performed at that time.
A loss of -29,300 yen due to a sharp drop in NGK. Trading exceeding ChatGPT’s purchase limit leads to a significant loss.
On Wednesday, November 4th, I sold the two stocks I purchased on Monday, November 2nd. Since November 3rd was a public holiday, the holding period was two days.
The result was 1 win and 1 loss, with a realized profit/loss of -29,300 yen. While I secured a profit with Japan Tobacco, the significant decline in NGK worsened the balance.
A feature of this session is that neither stock met ChatGPT’s pre-purchase conditions.
1. Japan Tobacco saw a small profit, while NGK saw a decline of over 6%
① Japan Tobacco (2914)
I purchased 100 shares at 7,340 yen on November 2nd and sold them at 7,366 yen on November 4th. The realized profit was +2,600 yen.
Before the purchase, an upward revision of earnings forecasts was attracting attention as a positive factor. However, the stock price weakened toward the end of the morning session and was below the VWAP, so ChatGPT’s final judgment was ‘NO BUY’.
As a result, the stock price rose by about 0.35% from the purchase price, which was the opposite of the ‘wait-and-see’ judgment. However, the rise was small, and it cannot be evaluated that a strong upward trend backed by the material continued.
② NGK (5333)
I purchased 100 shares at 5,108 yen on November 2nd and sold them at 4,789 yen on November 4th. The realized loss was -31,900 yen.
NGK was attracting attention due to an upward revision of earnings forecasts and a dividend increase, and ChatGPT also evaluated it as a conditional purchase candidate.
However, the suggested purchase limit was 5,005 yen, and the actual purchase price was 5,108 yen, which is 103 yen higher. Furthermore, at the time of purchase, I was cautious about conditions such as falling below 4,975 yen as a reason to wait.
The selling price fell by about 6.25% from the purchase price. Even with positive news, there is no guarantee that the rise will continue until the next business day, and the result was that short-term downside risk could not be sufficiently avoided.
2. Verifying ChatGPT’s judgment from the previous day
While Japan Tobacco made a profit despite the ‘wait-and-see’ judgment, NGK resulted in a significant loss from a conditional purchase candidate.
In the case of NGK in particular, by exceeding the purchase limit by 103 yen, I started the trade from a price that was disadvantageous by 10,300 yen per 100 shares.
Admittedly, there is no guarantee that the order would have been executed even if I had placed it at the limit price. While I cannot prove the effectiveness of price limits based on this result alone, it has once again highlighted that stock selection and purchase price management need to be verified separately.
3. Purchased JR Central, short-sold Kioxia
On November 4th, I opened positions in two new stocks.
① Central Japan Railway Company (9022)
I purchased 100 shares at 4,098 yen in a cash transaction. The purchase amount is 409,800 yen.
During the morning session, trading volume increased to about twice the normal level, confirming relative strength against the overall market and the land transportation sector. On the other hand, there was a sharp drop during the 11 o’clock hour, so ChatGPT did not recommend an aggressive purchase, limiting its judgment to a conditional recommendation with a reference limit price of 4,078 yen.
The actual purchase price is 20 yen higher than the reference limit price. Whether the volume-backed rise continues until the next business day will be the subject of verification.
② Kioxia Holdings (285A)
I opened a new short position for 100 shares at 9,750 yen. The short sale amount is 975,000 yen.
A short sale is a transaction where you can profit if the stock price falls.
On that day, Kioxia fell significantly in the morning session, and at the time of confirmation, it was down about 11.55% from the previous day and about 3.37% against the VWAP.
However, there was a rebound from the morning low, so ChatGPT was wary of a sharp rebound due to buybacks after the decline. It did not recommend a new short sale, limiting it to a monitoring target conditional on breaking the morning low.
In reality, I opened a short position, so whether the sale after the sharp drop was effective will be the next verification point.
4. Cumulative loss exceeds 150,000 yen. Next time, I will verify compliance with trading conditions
The cumulative performance reflecting this transaction is 126 trades, 68 wins, 58 losses, win rate 54.0%, cumulative realized profit/loss -156,680 yen. The cumulative realized profit/loss for November has also fallen to -27,860 yen.
Even if the win rate exceeds 50%, the situation where profits are lost due to a single large loss continues.
In the future, it is important to compare the results when following the purchase limits and wait-and-see conditions specified by ChatGPT with the actual trading results.
Next time, I will check if JR Central’s rise continues and if the short sale of Kioxia leads to profit. In particular, the focus is on whether short-selling stocks that have plummeted is a strategy worth the risk of a rebound on the next business day.
My thoughts this time
・I’ve been working overtime lately, so I’ve been slacking on keeping records. ・It seems I made a decent loss on NGK, but the flow after that was so intense that I have no memory of it at all. Still, a minus 30,000 yen. That’s pretty big, honestly (゜-゜)
※This article is a personal trading and verification record and does not recommend buying or selling specific stocks.
#Investment #StockInvestment #JapaneseStocks #Stocks #AssetManagement #AssetBuilding #SwingTrading #InvestmentRecord #IndividualInvestor #StockInvestmentBeginner #ChatGPT #AIInvestment #MarginTrading #InvestmentStrategy #TradeRecord