Principles and thinking tools for investment decisions
Start with why you hold a view, not just where you expect prices to go. Separate facts from assumptions, consider the loss if you are wrong, and name the evidence that would change your mind. These principles and thinking tools offer a way to examine your own decisions under uncertainty.
On this page
Start with one decision on your mind
Two people can read the same news and reasonably choose different actions. One needs the money soon; the other can wait. Their existing holdings and capacity for loss may differ too. Turning a market view into a personal decision begins with those conditions, not the forecast alone.
You do not need to use every tool at once. Bring one decision to mind, look for a missing principle, then choose a tool that helps you examine it. A sentence or two in response to the closing questions creates a record you can revisit later. This guide does not tell you which asset to buy or sell.
Principles for investment decisions
Before acting, examine the purpose and the risk. After acting, watch for evidence that should change the view. Clear grounds and manageable exposure matter more than the strength of a conviction.
Define the purpose, horizon, and constraints
Money needed soon is not the same as capital that can stay invested for years. Before comparing assets, write down when the money may be needed, which funds are unavailable, and the costs, taxes, or obligations that matter. Compare returns within those limits.
- The purpose is invented after the price moves, or money needed soon is treated like capital that can remain invested for years.
- What is this decision meant to achieve, when will it be judged, and which constraints cannot be violated?
Leave room for the next decision
A large loss shrinks more than an account balance. It can remove the option to wait or change course. Before considering the upside, examine the money, liquidity, and emotional capacity that would remain if the view proved wrong.
- Attention stays on maximum upside while a single adverse outcome is allowed to break the entire plan.
- Under a severe but plausible outcome, would there still be enough capacity to make the next decision?
Separate facts, interpretations, and assumptions
A release showing slower inflation is a fact. The view that pressure for tighter policy has eased is an interpretation. A forecast of lower rates is an assumption. Separating these layers shows where the evidence ends and your reasoning begins.
- An opinion is treated as confirmed fact, and later evidence is reshaped to preserve the original narrative.
- Which statements are observed facts, which are interpretations, and which remain untested assumptions?
Use scenarios instead of one forecast
Even the most plausible forecast has alternatives. Sketch base, upside, and downside paths, with the conditions and rough likelihood of each. The useful part is not an impressively precise number. It is knowing which signals would make you give a path more weight.
- Only the most vivid or desirable outcome is considered, while a low-probability but damaging path is ignored.
- What are the main paths, their early signals, and the range of outcomes attached to each?
Distinguish a good outlook from a good price
Good news does not always lead to a higher price. If an even better result was expected, the same release may disappoint. Consider both the outlook and what you believe the price already reflects, then identify the difference between them.
- Positive news is collected without considering that other market participants may already expect the same result.
- What base case appears to be reflected in the price, and exactly where does this view differ?
Compare likelihood with the size of the outcome
Several small gains can be outweighed by one large loss. Expected value considers the likelihood of each outcome alongside its size. Put potential gains, losses, and costs in the same comparison, and allow for the possibility that the estimates themselves are wrong.
- The chance of being right is emphasized while upside is precise and the size of failure remains vague.
- Have likelihood and payoff been compared together, and how would the decision change if those estimates were wrong?
Write down what would change your mind
A thesis explains why you believe an opportunity exists. Invalidation conditions name the evidence that would make that explanation hard to sustain. Record an adverse price move separately from a change in the central reasoning: they are not necessarily the same thing.
- The entry case is detailed, but no evidence can change the view, so every new fact is interpreted as support.
- What observation would show that the thesis is wrong or no longer relevant?
Size from a risk budget, not conviction
A risk budget is the amount of loss you plan to tolerate across your capital. Use it to examine position size, volatility, and risks shared with other holdings. It is a planning boundary, not a guarantee that actual losses will stop at that amount.
- Size rises with recent performance or emotion, while the same hidden risk across several positions is counted separately.
- If the thesis fails, how much of the total risk budget is exposed, including correlated positions?
Look at what happens when risks overlap
Imagine a sharp price move while trading becomes thin and other holdings also come under pressure. Leverage magnifies the change in profit or loss, while poor liquidity makes it harder to reduce exposure at the intended price. This is why leverage, liquidity, and changing correlations belong in the same review.
- Margin or recent volatility is treated as the full measure of risk, excluding calls for more capital, slippage, and simultaneous declines.
- Can the exposure still be managed if prices gap, liquidity falls, and correlations rise at the same time?
Review the process with the outcome hidden
A profit does not prove that every step was sound. A loss does not prove that the entire process was wrong. Review the information available at the time and the rules set in advance. That helps distinguish a lucky or unlucky result from a process worth keeping or changing.
- A profit is taken as proof that every step was sound, while a loss causes useful rules to be abandoned immediately.
- Would this still look reasonable if the outcome were hidden, and what is the single best improvement for next time?
Useful thinking tools
Choose the question that fits the difficulty. If the assumptions are unclear, try first principles. If failure paths are missing, try inversion. The aim is to reveal a blind spot, not to use every tool on every decision.
First-principles thinking
Pause when an explanation amounts to “that is how it works.” Break the claim into smaller parts: cash flow, supply and demand, contract terms, or funding. Which of those parts must hold for the larger claim to make sense?
- Trying to reinvent every practice from scratch or ignoring real institutional and historical constraints.
- What basic facts and causal links must be true for this claim to hold?
Base rates
Before asking why this time is different, ask what usually happened in similar cases. That frequency is the base rate. Once the comparison group is clear, examine whether the differences in the current case justify a different expectation.
- Choosing a comparison group that favors the conclusion or applying history after the underlying structure has changed.
- What normally happens in the best available reference class, and what concrete evidence makes this case different?
Inversion
If the route to success is unclear, ask what would make failure likely. An unaffordable position or a shortage of needed funds may reveal more than another optimistic forecast. Work backward from those conditions to the risks that can be reduced now.
- Generating an anxious list of negatives without translating them into avoidable or manageable conditions.
- If this ended in a major failure, what would be the most likely cause, and which part can be reduced now?
Second-order effects
The first effect of a change may alter what someone does next. Higher costs, for example, raise questions about pricing decisions and demand. Follow that next step, while naming the conditions and time needed for it to occur.
- Extending an endless chain of speculation without identifying testable links or a useful time horizon.
- Who changes behavior after the direct effect, when might that happen, and where could the effect travel next?
Probabilistic thinking
There is a wide space between “certain” and “impossible.” Describe how much confidence the evidence supports and what range of outcomes remains plausible. If you use a number, keep visible that it is an estimate rather than a known fact.
- Attaching precise numbers without evidence or treating a low probability as if it meant impossible.
- How much confidence does the available evidence support, and which unlikely outcome still requires preparation?
Bayesian updating
When new evidence arrives, compare how naturally it fits the current thesis and an alternative explanation. Evidence that distinguishes them is a reason to shift the weight of a view. A surprising headline alone is not a reason to reverse the conclusion.
- Reversing the conclusion after every headline, or fixing the initial probability so firmly that no evidence can move it.
- Which explanation made this evidence more likely, and how much should it change the current view?
Falsification
Supporting material is often easy to find. Falsification asks the question in the other direction: what would you expect to observe if the explanation were wrong? Name that evidence first, then look for it in the available information.
- Changing the thesis so that it explains every result, or assuming that any adverse price move disproves every underlying claim.
- If the explanation were wrong, what evidence would probably appear, and has that evidence been sought?
Opportunity cost
Choosing one option leaves less money and time for another. Put the current choice beside another asset, a simpler approach, or doing nothing for now. Include the attention and ongoing work it requires, not just the capital committed.
- Considering only the benefits of the current option while ignoring the next-best alternative and the cost of complexity.
- What is the best alternative use of the same capital and attention, and why is this choice better?
Systems thinking
A price can be a cause as well as an outcome. A move may affect collateral or funding; the resulting actions may then feed back into prices. Look for the loops that amplify or soften the move, and the delays between each step.
- Building an elaborate map without separating decisive variables from decorative links and time delays.
- Which feedback loops could amplify or dampen the move, and what delays stand between cause and effect?
Incentive analysis
When reading what a person or institution says, also ask which choices are rewarded. Governments, firms, and financial institutions face different rules and funding pressures. Those conditions can help explain why stated goals and actual behavior may diverge.
- Reducing every action to immediate financial gain while ignoring rules, reputation, and long-term relationships.
- How is each participant rewarded or constrained, and where might that make actions differ from statements?
Margin of safety
Estimates can be wrong. A margin of safety is the room left in price, size, time, or liquidity to account for that possibility. Identify which estimate could miss, by how much, and whether the plan would still hold. A buffer does not remove risk.
- Using the phrase as a synonym for vaguely cheap or treating it as a device that removes risk altogether.
- Where is the buffer, and would it remain adequate if the central estimate were reasonably wrong?
Premortem
Before acting, imagine that the decision has already failed. Write down why, alone or independently of others. For each possible cause, identify an early signal and a response. This turns a vague concern into a condition that can be checked.
- Producing a ceremonial risk list without assigning early signals, responsibility, or a response.
- What signal would reveal failure earliest, and who should take which action when it appears?
A short record before and after the decision
An unanswered question is useful information too. Rather than filling every blank, note what still needs checking. Later, compare the outcome with what you actually believed at the time, not the story that became clear afterward.
When defining the problem
- State the purpose, evaluation horizon, and when the capital may be needed.
- Separate observed facts, interpretations, and assumptions.
- Identify the main expectations already reflected in price.
- Choose an appropriate base rate and comparison group.
Before deciding
- Describe base, upside, and downside scenarios with outcome ranges.
- Explain the thesis and invalidation condition in one sentence each.
- Compare the opportunity cost and the option to do nothing.
- Review size against the loss boundary and total risk budget.
When testing vulnerabilities
- Check leverage, possible calls for more capital, and severe plausible loss.
- Consider reduced liquidity, slippage, and interrupted trading.
- Look for hidden overlap with other exposures.
- Use a premortem to define failure signals and response conditions.
When reviewing the outcome
- Judge the process using only information available at the time.
- Separate rule-following from a lucky or unlucky result.
- Record how unexpected evidence should update the view.
- Choose one practice to retain and one to change next time.
Worked examples and a blank template
Open a question that interests you. Each example separates facts from interpretation using the same eight fields. Then copy the blank template to work through your own question.
Retrospective educational examples
These examples were written after the official releases. They are not forecasts recorded beforehand, actual trading records or evidence of investment performance. The English versions were prepared on September 25, 2026; source and observation dates are shown separately.
The suggested next checks illustrate how to keep a record. They do not mean those follow-up checks have been completed.
Does a crude inventory draw mean demand is recovering?
- The question
- Can a decline in US crude inventories establish that oil demand is recovering?
- Verified facts
- In the week ended September 4, 2026, US commercial crude inventories, excluding the Strategic Petroleum Reserve, fell by 0.4 million barrels. Gasoline stocks rose by 1.3 million barrels and distillate stocks by 2.1 million. The four-week average of total products supplied was 3.7% below the same period a year earlier.
- Working interpretation
- The crude draw alone is not enough to establish a recovery in consumption. Rising product stocks and lower products supplied leave that interpretation without sufficient support.
- Another explanation
- A refinery processing more crude, or a decline in crude imports, can also reduce crude inventories. These are possible explanations, not a claim that imports fell in this report. The cause of an inventory change and the change in final consumption are separate questions.
- What is still unknown
- This example does not verify the market consensus before the release or the immediate WTI price reaction. It therefore cannot establish whether the report beat expectations or explain how prices responded.
- What to check next
- First compare refinery crude inputs and imports for the same period. Then check subsequent releases for agreement between product inventories and the four-week average of products supplied.
- What would change my mind
- Falling product stocks alongside improving four-week products supplied would justify revisiting the demand-recovery explanation. Continued stock builds and weak products supplied would call for more caution.
- The decision for now
- Withhold the conclusion that a crude draw proves stronger demand. This report alone does not determine the direction of WTI prices or whether to make a trade.
Products supplied is a proxy for consumption, not a direct measure of final purchases. Inventory changes here follow the report narrative; subtracting rounded inventory levels can give slightly different results.
Is a rate cut, by itself, good news for stocks?
- The question
- Does a lower policy rate alone establish better conditions for US large-cap stocks?
- Verified facts
- On December 18, 2024, the Fed cut its target range by 0.25 percentage points to 4.25–4.50%. In the projections published that day, the median policy-rate projection for the end of 2025 rose from 3.4% in September to 3.9%. The 2025 PCE inflation projection rose from 2.1% to 2.5%.
- Working interpretation
- The decision today and the projected path after it carry different information. One cut does not establish that the future discount-rate burden will ease faster.
- Another explanation
- The 2025 real GDP growth projection also rose from 2.0% to 2.1%. Stronger growth and earnings could offset some pressure from rates. A higher GDP projection, however, does not verify an increase in any particular company’s earnings.
- What is still unknown
- This record does not verify the market’s prior expectations, stock or Treasury yield reactions, or earnings forecasts. It does not judge whether the announcement was more hawkish than expected or explain any stock-price move.
- What to check next
- Examine the PCE releases following that decision, Treasury yields and corporate earnings expectations in turn. Separately check whether improvement extends across industries and companies.
- What would change my mind
- Reassess conditions if inflation eases, the rate burden falls and corporate results improve together. If inflation pressure persists or earnings expectations weaken, a cut alone is not enough for optimism.
- The decision for now
- Do not infer a direction for large-cap stocks from the cut alone. Keep the current decision, future rate path and corporate earnings separate until each has been checked.
The policy-rate median summarizes participants’ projected year-end target-range midpoints; it is not a promised decision. Inflation and growth projections measure fourth-quarter changes from a year earlier. These were forecasts at the time, not subsequent outcomes.
- Federal Reserve · December 18, 2024 statement · Opens in a new tab
- Federal Reserve · December 2024 projections, Table 1 · Opens in a new tab
- Federal Reserve · September 2024 projections, Table 1 · Opens in a new tab
Must the 10-year yield fall when the Fed cuts?
- The question
- Does a cut in the Fed’s policy rate mean the 10-year Treasury yield must fall too?
- Verified facts
- On December 18, 2024, the Fed cut its target range by 0.25 percentage points to 4.25–4.50%. Treasury’s daily 10-year constant-maturity yield was 4.40% on December 17, 4.50% on December 18 and 4.57% on December 19. It rose by 17 basis points, or 0.17 percentage points, over two trading days.
- Working interpretation
- Over this interval, the policy decision and the 10-year yield moved in opposite directions. A change in today’s policy rate does not determine the direction of a long-term yield.
- Another explanation
- A different expected path for short-term rates, or different compensation for the uncertainty of holding a long-term bond, can also move the 10-year yield. These are possible explanations; this example does not estimate their contributions.
- What is still unknown
- The record does not verify prior market expectations, minute-by-minute trading, or contributions from real yields and the term premium. Daily observations cannot identify a single cause or establish the immediate announcement reaction.
- What to check next
- Compare 2-year and 30-year yields, real yields and subsequent daily 10-year observations. To evaluate a particular bond or futures contract, check its own prices and terms separately.
- What would change my mind
- If later observations fall below the starting yield, revisit whether this rise persisted. In either direction, do not turn three observations into a rule for every rate cut.
- The decision for now
- Distinguish the policy rate from long-term yields. Do not choose a bond trade solely because the Fed cut, or calculate an investment return from this table.
Constant-maturity yields are curve estimates based on indicative quotations. They are not executed prices for a particular bond, or the return on a Treasury futures position. A rise of 17 basis points is not a 17% investment return.
A blank template for your own decision
Copy the template into your own notes. Write “unverified” where you do not yet have an answer. This page does not save or send your record.
Reference material
This guide was independently written in the language of Futures Compass after consulting the following public investor-education resources and decision-research references.