1.1The question
Some decisions can be tried and taken back. Others leave a mark that stays whatever is learned afterwards. A trader who sends a real order into a real market cannot recall the fill; a desk that adds to a losing position cannot later un-add except by closing at the price then available. In both cases the person deciding wants to know how the step will turn out, and in both cases the best evidence about how it turns out is the step itself.
This article asks a narrow question of two decision records written for that kind of step. Where the step cannot be undone, do the records divide it into stages, as the economic theory of the value of waiting would seem to advise, and if they do, what do they treat as the information that justifies the next stage? The question is put to two records that WSTiger, a studio in the RISCON family that connects economic history and market signals to the question of what risk to take now, publishes for study. Both records are synthetic. They are modern summaries written for the platform, not archival documents, and the firms, desks and traders in them are invented. What they offer is a decision written down before its outcome, which is the moment the question concerns.
1.2The contribution
The contribution is to connect three bodies of argument that are usually read apart, through a fourth, older text.
The three are the option-value argument for waiting (Arrow and Fisher 1974; Dixit and Pindyck 1994), Collingridge's dilemma of control (Collingridge 1980), and Kelly's arithmetic of sizing (Kelly 1956), each set out in Section 2. The first gives value to keeping an option open before an irreversible step because information will arrive later. The second holds that the effects of a new technology cannot be known until it is in use, by which time it is hard to change, and argues for decisions kept corrigible. The third sets the stake as a fraction set by the edge, and holds that staking more raises the risk of ruin.
The fourth text is Section IV of David Hume's An Enquiry Concerning Human Understanding (1748). It is not about markets or technology. It is about how anyone comes to know what a thing will do. Its answer is that the effects of an object are learned only from experience of that object, never by reasoning from its appearance, and that experience informs only about the objects and the period it covered.
Put side by side, these texts meet at one point. The option-value argument values waiting because information will arrive later. Collingridge's dilemma says that for a new technology the decisive information arrives only once the thing is in use. Hume's section explains why that should be so: the effects of a new object cannot be read off its appearance, and an object that merely resembles it, however closely, informs only about itself. The connection this article draws is that the information whose arrival gives waiting its value is, for some steps, information that only the step itself can produce. Where that holds, waiting yields nothing, and the choice is not between committing and waiting but between committing in full and committing in part. That is the reason, on this article's argument, why the simulation record divides the step: not to wait for information, but to buy it with a small irreversible exposure. Kelly's arithmetic then marks the limit of the purchase, since a stake beyond the fraction set by the edge raises the risk of ruin, and the edge itself is something only experience can estimate.
The records add a document-level test of this connection, in a limited and clearly labelled form. They were not written to test Hume or the option-value literature, and nothing suggests their authors had either in mind. They show what a decision written before its outcome says about a step that cannot be undone, and what it leaves out.
1.3The debate joined
The disagreement this article joins has two sides with named authors and a third position that cuts across them.
On one side stand Arrow and Fisher (1974) and Dixit and Pindyck (1994). As their reading notes give it, Arrow and Fisher hold that irreversibility and uncertain future benefits give value to keeping the option to preserve, and Dixit and Pindyck hold that an investment that cannot be undone and can be delayed carries an option value of waiting.
On the other side stands Collingridge (1980), whose dilemma holds that while a technology is young its effects cannot be known, and once they are known it is entrenched and hard to change. His remedy is not delay but corrigibility: decisions that can be monitored and corrected, and options kept open.
Across both stands Lindblom (1959), who holds that administrators in practice decide by successive limited comparisons of a few options that differ little from present policy, and defends this incremental method as the one that works. Lindblom's note says nothing of irreversibility, but the method it describes decides in steps.
Hume (1748) enters beneath the debate. His Section IV does not take a side on waiting. It says something about where knowledge of effects comes from, and that claim bears on what either side can expect from delay.
The placement of the secondary works in this debate, and their application to Hume's text and the two records, are this article's own. Each work is credited only with what its reading note in the parent study records.
1.4The plan of the article
Section 2 sets the debate out side by side, with the arithmetic of sizing and the judging of decisions by outcomes. Section 3 describes the primary text, the records and the parent study's ledger. Sections 4 to 6 read Hume's Section IV beside the record on moving from simulation to live trading, the record on adding to a losing position, and the two records' revisit lines. Section 7 discusses what the reading can and cannot carry, and Section 8 concludes.