From pre-trade analysis to a trading journal
Pre-trade analysis asks a forward-looking question before you place an order: what execution costs and risks does the proposed basket present? A trading journal asks a different question afterward: what did you actually do, and what can you learn from the record? PreTrAIde Memory connects those two steps without placing trades for you.
You build a basket of stocks, each with a buy or sell side, quantity and execution conditions. PreTrAIde analyzes the basket before any order is placed and returns a GO, NO-GO or MODIFY verdict for the basket and each line. Each verdict includes a model-estimated success probability, a risk assessment, a suggested execution strategy and plain-language reasoning. The probability is not a guarantee or a forecast of returns.
You decide what, if anything, to trade through your own broker. PreTrAIde is analysis software: it does not receive, transmit or execute orders. Memory begins with the distinction between that proposed basket and what you later log as having happened.
How the closed loop works
- Analyze a proposed basket. Review the verdict and its reasoning before placing any order.
- Record what you did. Log a trade as executed, partially executed or skipped, and enter real fill prices where applicable. A proposed order and a completed trade are not automatically the same thing.
- Review the measured outcome. PreTrAIde measures outcomes from market data and builds a memory for your account based on logged trades.
- Use that history in later analyses. The memory personalizes the AI layer’s reasoning. It does not change the deterministic quantitative scores.
This loop depends on what you log. If an order was only partly filled, recording it as fully executed would erase an important distinction. If you skipped a trade, logging it as skipped preserves the difference between a decision you considered and a position you actually took.
What changes—and what does not
PreTrAIde calculates 0–100 liquidity, volatility and execution scores for each stock and for the basket. Its quantitative layer also estimates measures such as spread, participation in daily volume, market impact and 1-day 95% value at risk. That layer is deterministic and reproducible: Memory does not alter its quant scores.
Personalization applies to the AI layer, which reasons about individual stocks and the basket as a whole while anchored on the quantitative scores, news and analyst ratings. A later explanation can therefore draw on your account’s logged history without presenting a different liquidity score simply because it belongs to you.
The same distinction matters for verdicts. Memory is context for analysis, not permission to override data-based safeguards. PreTrAIde’s guardrails correct AI output that contradicts the data; for example, a suspended stock is always NO-GO. Guardrails move verdicts only in the conservative direction.
| Part of an analysis | Effect of PreTrAIde Memory |
|---|---|
| Quantitative liquidity, volatility and execution scores | Unchanged |
| AI reasoning about a stock or basket | Can be personalized using the account’s logged-trade memory |
| Trade execution | Still handled by you through your own broker |
Memory levels and your controls
Personalization levels unlock at 1, 5, 15 and 40 logged trades. These are milestones for the per-account memory, not claims that an analysis becomes more accurate by a particular amount. A larger log offers more recorded history for the AI layer to consider, but it does not guarantee a better decision or outcome.
You can pause or erase PreTrAIde Memory at any time. That control is worth considering alongside the ordinary discipline of keeping a trading journal: decide what you want to record, distinguish intended trades from actual executions, and use the resulting history as context rather than as a verdict on your investing ability.
A hypothetical journal entry
Suppose you analyze a basket containing a proposed purchase of 100 shares. You later buy 40 shares at an actual fill price of US$25 per share and do not execute the remaining 60. The useful record is partially executed, with the fill price for the shares bought—not a claim that the full 100-share order was filled. In another case, you might analyze a basket and place no order at all; logging it as skipped keeps that decision distinct from a trade.
These are hypothetical bookkeeping examples, not predictions. Market-data outcomes can provide a point of reference for later analysis, but they cannot establish that the original trade was right or wrong for every investor. Your objectives, constraints and reasons for acting may extend beyond what a market-data outcome shows.
When this feature is useful
Memory gives structure to a habit many investors try to maintain manually: write down the plan, record the action, and revisit the result. Its specific role is to carry logged-trade history into later AI reasoning while leaving the underlying quantitative scores intact.
That makes it different from paper trading. PreTrAIde’s trade simulator lets you follow how a basket would have performed without real money; Memory records what you say you actually executed, partially executed or skipped. Neither replaces your broker’s execution record or turns a pre-trade verdict into investment advice. The practical starting point is straightforward: log actual actions and fill prices carefully, then read personalized reasoning alongside the same quantitative measures you would see without Memory.