Method

Thesis-Driven Investing

The method Telumin is built around: every position rests on a written, falsifiable argument — and the conditions that would prove it wrong become rules a machine can watch.

The core idea

You don’t buy a stock because it “looks cheap” or “has a good story.” You state, in writing: why the market is wrong, what must happen for the investment to work, and — most importantly — what would prove you wrong. It treats a position like a scientific hypothesis instead of a feeling.

That last part is where most investors actually lose money: not for lack of a fair-value estimate, but by ignoring a dying thesis, holding on out of hope, or missing the filing that broke the story. A written thesis with explicit kill criteria makes those failures visible early.

The six pillars of a strong thesis

1Variant perception

Why is the market wrong or incomplete about this company right now? The most important pillar — and the one most often missing.

2Business quality

Competitive advantages, unit economics, reinvestment runway, management and capital allocation.

3Valuation

How undervalued is it? This is where DCF, multiples, or your model of choice does its work.

4Catalysts

The specific events or trends that should close the gap between price and value — with rough timing.

5Risks & kill criteria

What would invalidate the thesis? Written as concrete break-conditions — these become your monitoring rules.

6Expected return & horizon

The return you are underwriting, over what period, across upside / base / downside.

Common thesis-driven strategies

StrategyCore focusHorizonThesis style
Classic valueUndervalued quality businesses3–7 years“The market undervalues durable free cash flow.”
Quality compoundingHigh-ROIC businesses that reinvest5–10+ years“This company can compound capital at high rates for a long time.”
GARPGrowth at a reasonable price2–5 years“Growth is durable and not fully priced in.”
Special situationsSpin-offs, turnarounds, events6–24 months“This event unlocks value the market ignores.”
Contrarian / deep valueBeaten-down or hated names2–5 years“Temporary problems are over-discounted.”
Thematic with rigorStructural trends, bottom-up picks3–7 years“This secular trend benefits these specific companies disproportionately.”

The strategy differs; the discipline is the same. Every one of these produces a thesis with break-conditions — which is exactly the input Telumin monitors.

Where valuation fits — and where Telumin does

A DCF answers “what is this business worth today?” Telumin answers a different question: “are the reasons I own this still true?” They are not substitutes — they are stages of the same process.

AspectValuation (e.g. DCF)Telumin
QuestionIs this cheap or expensive vs. future cash flows?Is anything breaking the reasons I hold it?
TimingBefore you buy; periodic re-valuationAfter you own it — continuously
InputsCash-flow forecasts, margins, WACC, terminal valueYour written reasons + break-conditions
OutputIntrinsic value and margin of safetyEvidenced alerts with priced exposure and a decision window
Telumin does not value stocks — deliberately. Valuation helps you enter with a margin of safety; Telumin helps you stay in, or get out, based on whether your original reasons still hold. Use both, in that order.

The process, end to end

  1. Generate ideas — screens, reading, industry knowledge.
  2. Write the thesis down (the template on the next page).
  3. Stress-test it — pre-mortem, devil’s advocate, scenarios.
  4. Quantify with valuation and set a margin of safety.
  5. Define concrete break-conditions — measurable, dated where possible.
  6. Monitor continuously — this is the stage Telumin automates.
  7. Update or exit when the thesis changes; log the decision and the reason.

In Telumin, step 5 is not paperwork: when you add a position you write the reason in one line, the engine structures it into break-conditions, and the thesis drift sweep re-checks them around the clock. Accepts and rejects land in the decisions journal with your reasoning attached — so step 7 compounds into institutional memory.

Keep reading