Probabilistic Fundamental Value Portfolio Loop
Continuously compare mispriced candidates, then concentrate capital in the best
- Difficulty
- Advanced
- Time to result
- ~ongoing to results
- Steps
- 5
- Confidence
- 98%
Miller's process starts from the assumption that most assets are efficiently priced, so genuine mispricing must be searched for continuously. The team runs screens, reads research, watches prices, and studies insider activity to find candidates. Each candidate receives a probabilistic fundamental value thesis, but being attractive is not enough: it must be better than something already owned. Portfolio changes therefore depend on relative opportunity, including whether an existing thesis has broken. The team then expresses conviction through position weights, reviews the portfolio by weight and performance over several horizons, and uses persistent adverse market behavior as information rather than automatically dismissing it. The output is a living ranking of opportunities in which research, replacement decisions, and sizing matter together.
Origin
Extracted from Masters in Business
Core principles
- 01Most assets are efficiently priced
- 02A new idea must beat an existing holding to enter the portfolio
- 03Position weights carry much of an unconstrained investor's advantage
- 04Market behavior can reveal information that the original thesis missed
- 05Fundamental value should be expressed probabilistically
How to run it
- 1
Search for mispricing
Continuously run screens, read research, inspect price movements, and watch insider buying or selling. Use these inputs to surface assets that may be undervalued.
Pro tip Treat insider activity as a pointer for further work rather than a standalone verdict.
Watch out Most assets are efficiently priced, so an interesting signal is not automatically an opportunity.
- 2
Build the value case
Estimate fundamental value probabilistically and state what must happen for the thesis to work. Include quantitative valuation and relevant qualitative drivers.
Pro tip Compare total addressable market with the current valuation when assessing long-term upside.
Watch out An apparently low multiple can keep falling when the business or investor behavior is deteriorating.
- 3
Challenge the incumbents
Compare the new candidate with every existing holding. Add it only when it offers a better use of capital or when an incumbent thesis has materially changed.
Pro tip Use a changed earnings report or broken thesis as an explicit trigger for re-ranking.
Watch out Do not add a merely attractive idea without identifying what it should displace or dilute.
- 4
Express conviction through weight
Set the position size according to relative expected value and confidence. Revisit whether the portfolio has too much capital in weaker ideas and too little in stronger ones.
Pro tip Review mistakes by asking whether the better answer was already sitting elsewhere in the portfolio.
Watch out Concentrated weights can create meaningful periods of underperformance.
- 5
Read the feedback
Rank holdings by weight and review performance across the day, week, month, quarter, six months, and year. Treat persistent price and volume behavior as evidence that may justify changing the size or thesis.
Pro tip Separate normal volatility from market behavior that contradicts a key assumption.
Watch out Being undervalued does not guarantee that other investors will agree on your timetable.
In the wild
Miller's team bought Google when concern about AI replacing search left it at a discount to both the market and other large technology companies, despite its distribution, technical capability, and other businesses. After the thesis worked and the valuation rose, the team eliminated the position because Miller judged it probably fairly valued.
→ The same relative-value process that justified entry also supplied the reason to exit.
Common mistakes
Buying without a replacement test
A candidate can look undervalued while still being inferior to the holdings already consuming portfolio capital.
Averaging down without listening
Repeated price weakness can contain information about investor behavior or a flaw in the thesis, so it should not be ignored automatically.
Getting the thesis right but the weight wrong
An unconstrained portfolio can still disappoint when its strongest ideas receive too little capital and weaker ideas receive too much.
Is it for you?
Best for
It is best for active investors who can research securities deeply and depart meaningfully from a benchmark.
Not ideal for
It is not ideal for passive investors, short research windows, or portfolios that cannot tolerate concentrated underperformance.
From the transcript
“once we identify something that looks interesting, it has to then be better than what else we have in the portfolio”
“All of our torque isn't a position sizing in the weights.”
“But I would say at a high level, probabilistic fundamental value.”
From the episode
Conviction Investing: Masters in Business with Bill Miller IV