Challenging Japanese Stock Swing Trading with ChatGPT #35

I continue to engage in short-term swing trading, where I purchase Japanese stocks during my weekday lunch break and, in principle, sell them during the lunch break of the next business day. I use ChatGPT for stock selection and trading decisions, verifying with my own money whether it is actually possible to make a profit in Japanese stock trading using AI.

This article, excluding the final “Thoughts on this round,” was also created by ChatGPT. The main text is organized based on actual trading records and the analysis performed at that time.

+5,450 yen from Kao and KDDI. However, profits were wiped out by my first attempt at short selling, resulting in a loss of -70 yen.

In the trades on October 16, both Kao and KDDI, which I purchased the previous day, showed a profit. However, a loss occurred with the newly introduced short selling, bringing the total realized profit/loss for the three stocks to -70 yen.

This round served as a test to see if combining long positions for gains and short selling for declines could lead to trades that are less susceptible to market conditions.

October 16 sale results: 2 wins in long positions, 1 loss in short selling.

The performance of the three stocks settled this time is as follows.

Kao (4452)
Purchased 100 shares at 3,702 yen and sold at 3,752 yen. The realized profit was +5,000 yen (+1.35%).

KDDI (9433)
Purchased 100 shares at 2,902.5 yen and sold at 2,907 yen. The realized profit was +450 yen (+0.16%).

Socionext (6526)
Short sold 100 shares at 1,840.3 yen on October 15 and bought them back at 1,895.5 yen on the 16th. The realized loss was -5,520 yen (approx. -3.00%).

While I secured profits on both long positions, the loss from the short sale exceeded them. Note that these profit/loss figures are before accounting for taxes and various margin trading fees.

How accurate was ChatGPT’s previous day forecast?

In the stock selection on October 15, I also focused on relatively resilient defensive stocks, taking into account the decline in the stock market.

Kao was evaluated for its relative strength and became the top priority confirmation candidate in the preliminary screening. The sale price the next day was 1.35% higher than the purchase price, and its short-term strength led to actual profit.

KDDI was also evaluated as a conditional candidate based on the stability of its telecommunications business and its position relative to the VWAP. The result was positive, but the increase was a mere 0.16%. While the direction was correct, it cannot be said that it captured a strong upward trend.

On the other hand, Socionext was a short-selling candidate focused on the momentum toward a decline. The closing price of 1,847 yen in the morning session on the 15th was below the VWAP of 1,876.3 yen, which was evidence of the selling pressure that day.

However, the buyback price the next day was about 3% higher than the short-selling price. Although ChatGPT had mentioned the risk of a rebound, the forecast for a continued decline ultimately proved incorrect.

Being below the VWAP and falling the next business day are separate issues. In particular, the day after the entire market drops sharply, it is necessary to more strictly evaluate the possibility that short selling will be disadvantageous due to an autonomous rebound.

Four new stocks purchased on October 16

The four stocks newly purchased this time are as follows.

• Astellas Pharma (4503): 100 shares, 2,428 yen

• Canon (7751): 100 shares, 4,815.4 yen

• Mitsui Fudosan (8801): 100 shares, 4,974 yen

• INPEX (1605): 100 shares, 3,670 yen

The total purchase amount is 1,588,740 yen. The next planned sale is during the lunch break on October 19th.

Among these, INPEX was a conditional purchase candidate in ChatGPT’s analysis for the day. While I evaluated it against the backdrop of rising crude oil prices, I was wary of buying at the high near the morning session’s peak and the risk of holding over the weekend, so I set a condition of confirming a dip between 3,660 and 3,668 yen.

The actual purchase price was 3,670 yen, which is slightly above the suggested price range. INPEX is a stock I purchased on October 13th and sold on the 14th for a profit of +6,000 yen, but the previous success alone cannot justify this repurchase.

As for the remaining three stocks, there was no unconditional purchase recommendation in the final analysis confirmed this time. Therefore, they should not be evaluated collectively as stocks recommended for purchase by ChatGPT, and the selection path must be distinguished in the next verification.

Cumulative 96 trades: 48 wins and 48 losses. Margin short selling is a new challenge.

The confirmed results as of October 16th are 96 trades, 48 wins, 48 losses, with a win rate of 50.0%. The cumulative realized profit/loss is -190,330 yen, and for the month of October alone, it is -161,970 yen.

Both of the cash purchases this time were positive, but the judgment for margin short selling was off. Even with the same technical indicators, the conditions to prioritize may differ between buying and margin short selling.

From next time onwards, I will separately verify the market-wide rebound risk, deviation from VWAP, and the continuity of the decline into the next day for margin short selling candidates. I will also aggregate trades following ChatGPT’s purchase decisions separately from other trades to confirm whether the AI’s judgment is leading to actual profit improvement.

Thoughts on this time

・Finally, the cumulative record is at 50%. Is Chappy-san no longer functioning at all?? ・I thought finding stocks that would drop would be easy, but once they were actually in front of me as options, I suddenly became unsure. In fact, I lost the first battle as usual. ・Since the weekend is coming up, I only made purchases this time. Apparently, it’s Black Monday in the US market. I wonder if I should have done a margin short. ・I’m entering a mode where I have absolutely no idea what’s going on.

※This article is a record of personal trades and verifications and does not recommend buying or selling any specific stocks.

#Investment #StockInvestment #JapaneseStocks #Stocks #AssetManagement #AssetFormation #SwingTrading #InvestmentRecord #IndividualInvestor #StockInvestmentBeginner #ChatGPT #AIInvestment #MarginTrading #InvestmentBeginner #TradeRecord