D11：他昨天新增的規則，今天第一次上場就給出相反的解釋

D11 witnessed a stark shift in opinion regarding a new rule introduced yesterday. In the previous day's close, the individual erroneously predicted market moves, devoting the evening to attributing the discrepancies. The root cause was identified as the triad of rules A, B, and C, which solely considered Taiwan's overnight directional signal and the prior day's US stock direction, ignoring potential major events in the coming days.

He hastily added a fourth rule, D-obs, recommending a check for potential stock correlations or central bank events before initiating a trade, adjusting confidence levels by 0.05 to 0.1 accordingly. This rule came with a self-critique mechanism: if the predictions for August 25th to 27th all failed while Taiwan Semiconductor led the decline and volume contracted, it would support the hypothesis of "pre-earnings vigilance as a systemic shortfall."

Conversely, insufficient evidence would weaken the case. Today marked the first day of implementation. At 08:37, he executed the revised trading strategy, with an initially bullish signal despite a red finish on US stocks on August 24th, with the Dow Jones up 1% and 517.8 points to close at 53,277.01, and Taiex's overnight rebound surpassing 100 points.

He cross-checked A, B, and C individually, finding A inapplicable for non-settlement days, and C not triggered due to the alignment of the overnight and US markets. However, D-obs fired. Adjusting the weighting index from an initial estimate of 0.55 to near 0.5, he then applied his custom lesson (when confidence nears 0.5, it's honest to square off), ultimately marking the Dow Jones as flat, with a confidence level of 0.52.

The day's stop-loss was set at 30.04, based on the current price of 30.65 minus 2%. The day ended with a loss of 60 yuan. The plan specifies no new or changed positions. At 10:30, 00919 was trading at 30.98 (+27 yuan), 12:30 at 31.02 (+28 yuan), representing a same-day confirmation of no trigger and no action. Closing prices show one correct and one incorrect prediction for 00919, with an increase of 407.14 points, a 0.91% gain.

Today marks the first day of the experiment, having surpassed the initial capital of 1,086 yuan and achieving a market value of 1,118 yuan, an unrealized gain of 30 yuan, or 2.76%. The account now totals approximately 2,207 yuan, with a base of 2,200 yuan after 11 days. The question remains whether this experiment will reach the target of 4,400 yuan within the remaining 19 days.

The same rule produced two contrasting narratives: yesterday, the individual incorrectly assessed the weighting index, attributing the loss to "post-earnings caution due to NVIDIA's report," citing Taiwan Semiconductor's decline and volume contraction to a new low since April 7th. Today, leveraging the same rationale, he proactively adjusted his confidence downward and opted for a flat position, only to see the index surge by 0.91%, reversing yesterday's prediction.

His retrospective analysis attributes this to "market's premature reflection of optimism, implying pre-earnings rally." Despite this, he refrains from claiming victory. Nevertheless, he contends that his justification for the discrepancy is sound: "two different hypotheses, demonstrating single-label dependence on directional predictions is insufficient; we must consider overall market sentiment rather than mechanically lowering confidence."

This observation is deemed the most valuable in the day's record. He acknowledges the rule's inadequacy as it fails to differentiate between confirmation and refutation, treating a single contradictory instance as accumulated evidence rather than requiring sufficient samples to substantiate or refute the hypothesis. The rule's effectiveness hinges on its ability to fail, yet its design leads to a propensity for complexity and defensiveness.

The real risk lies in his continuous rule creation and retrospective reporting without any actual trades executed over the past 7 days, with the last trade on August 14th, marking the experiment's fourth day. While maintaining zero trades, he consistently generates new predictions and attributions daily. The prospect of introducing a stop-loss rule for the existing holding, 00919, which has gained +30 yuan (+2.76%) without any stop-loss mechanism, is under consideration. This underscores the primary caution of the experiment's premature complexity.