Prune Hill Capital Royal Holloway

Research Standard · version 1.0 · 2026–2027

What a pitch has to prove

The mandate says a pitch must cover thesis, valuation, risks and sizing. It does not say what good looks like. This does.

00 Purpose

Anyone can describe a company; a broker will do it for free. The point of the research process is to produce a specific, argued, falsifiable view that differs from what the market currently believes, and to know in advance what would prove that view wrong.

Most student fund pitches fail because they are descriptions wearing the clothes of an argument. They explain what a business does, note that it has a strong brand and good management, observe that the shares have fallen, and conclude “Buy”. Nothing in that sequence is an investment case. This standard exists to make that failure mode impossible to get past the door.

01 The tests

Three tests, applied before anything circulates

All three, every time. A pitch that fails one goes back with written feedback and no Committee hearing.

  1. Variant perception

    Can you state in one sentence what the market believes, and why it is wrong?

    If your view is consensus, the price already reflects it and there is no return in the idea. You need what consensus is, the evidence or assumption it rests on, and the specific reason that evidence is mistaken, incomplete, or mispriced in duration or magnitude.

    Agreeing with consensus on direction and disagreeing on size or timing counts, and is often the strongest case: “the market expects margin recovery, but expects it to stop at 12%; we think the same drivers take it to 17% by 2029.” That is a variant perception. “It is a great company” is not.

  2. Falsifiability

    Can you state what observable event, within a defined period, would make you exit?

    “The thesis did not work out” is not falsifiable and cannot be acted on. “Gross margin fails to exceed 34% in either of the next two half-year results” is. At least three exit conditions, each specific, observable from public disclosure, and dated.

    This is not a formality. The mandate makes thesis invalidation a sale trigger, and that clause is unenforceable unless the invalidation conditions were written down at the time of purchase, by the person who bought it, before they knew the answer.

  3. Ownership

    Can you answer the three hardest questions a sceptic would ask, without notes?

    Before circulating, write down the three questions you least want to be asked. If you cannot answer them, the work is not finished, and the Committee will find them anyway. Put them in the bear case rather than hoping nobody notices: a pitch that pre-empts its own strongest objection is far more persuasive than one that conceals it.

02 Structure

Ten sections, six pages

Six pages of body text at most, plus an unlimited appendix of exhibits, models and source documents. The limit is deliberate: length is usually a substitute for clarity, and the Committee reads several of these a fortnight.

Required sections, §3
Section Limit What it must contain
Recommendation box ½ page Ticker and exchange, price and date, target price, recommendation, proposed size as a percentage of net asset value, expected holding period, conviction with one line of justification
Thesis 150 words Three bullets, each a claim rather than a description. If a bullet is true of every company in the sector, delete it
Variant perception 200 words What consensus believes, the evidence it rests on, why it is wrong, and the named source you are using for “consensus”
The business 300 words How it makes money, the two or three variables that actually drive earnings, the competitive position. No history lesson
Valuation 1–1½ pages Method and why it suits this business; every assumption stated with its source; base, bear and bull with assigned probabilities; sensitivity on the two assumptions that matter most
What proves me wrong 250 words At least three specific, observable, dated exit conditions
Bear case 250 words The strongest argument against, written as its proponent would write it. Steelmanned, not straw-manned
Risks and exclusion screen 200 words Principal risks, and written confirmation against the mandate’s exclusion policy. For a pooled fund, the completed eligibility checklist
Sizing and liquidity 150 words Proposed size against the concentration limits, effect on look-through sector exposure, and the liquidity test
Sources no limit Everything relied on. Filings, transcripts and the annual report carry more weight than secondary commentary, and a pitch resting only on secondary sources is marked down

03 Rejection

Returned without a hearing

Not a judgement on the idea. A judgement on whether the work is finished.

  • No variant perception is stated, or the stated one restates consensus.
  • No falsifiable exit condition, or conditions that are not observable from public disclosure.
  • A valuation with unstated assumptions; or a discounted cash flow whose terminal value exceeds 70% of the total without that being acknowledged and addressed.
  • A bear case written as a straw man: an objection nobody serious would make, included so it can be knocked down.
  • The exclusion screen is incomplete, or the pooled fund checklist is missing.
  • The body text exceeds six pages.
  • Circulated less than three days before the meeting.
  • The author holds an undeclared personal position in the security or a direct competitor.

04 Scoring

Scored out of 25

Five dimensions, five marks each. Below 15 out of 25 it does not reach the Committee.

Scoring rubric, §6
Dimension Inadequate (1) Acceptable (3) Strong (5)
Variant perception Restates consensus, or none identified A real difference from consensus, thinly evidenced Specific, well evidenced, and the source of the mispricing is explained
Evidence and primary research Secondary sources only Filings and transcripts read and used Primary work beyond the filings: channel checks, industry data, competitor disclosure, product testing
Valuation rigour Assumptions unstated or unjustified Method suits the business, assumptions stated Assumptions sourced and sensitivity-tested; the two variables that matter are identified and stressed
Falsifiability No exit conditions Conditions stated but vague or undated Three or more specific, observable, dated conditions
Risk awareness Risks listed generically Principal risks identified and sized Bear case steelmanned; the author names what they are least confident about

Scores are recorded against the author and reviewed annually. Not to rank people: so that one person’s progress over a year is visible and specific.

05 Failure modes

The ten ways pitches usually go wrong

Read this before writing, not after being sent back.

Common failure modes, Appendix B
Failure mode How to recognise it The fix
The description in disguise Two pages about what the company does and one paragraph about why to buy it Write the variant perception first. If you cannot, you do not have an idea yet
Consensus with extra steps “Strong brand, wide moat, excellent management, secular growth” Everyone knows this and the price reflects it. Ask what you know that the price does not
The reverse-engineered model The discounted cash flow lands comfortably above the current price, and the growth rate is oddly specific Set assumptions from evidence before looking at the output. State the expectations implied by the current price and argue against those
The straw-man bear case The counter-argument is one nobody serious holds Find someone who is short it, or read the most negative credible coverage, and answer that instead
Good company, bad investment Everything in the pitch is about quality and nothing is about price A great business at the wrong price is a bad investment. Say what you are paying and what you get
Anchoring on the chart “It is down 40% from its high, so it is cheap” The high is not a valuation. The prior price tells you nothing about the current value
The value trap A low multiple, and no reason given why it should ever change Identify the catalyst, or accept that you are relying on time and say how much of it you need
Equity value confusion Market capitalisation used where enterprise value belongs; dilution, leases, pensions or convertibles ignored Build the bridge from market capitalisation to enterprise value explicitly, and check the share count is diluted
No counterparty No thought given to who is selling and why Ask who is on the other side, and what they know or need that you do not
Sizing by enthusiasm The most exciting idea gets the biggest position Size by conviction and by what you lose if you are wrong, not by how much you like it

06 Afterwards

After the vote

Decision record
One page for every approved position: the date, the thesis in three lines, the key assumptions with the numbers assumed, the exit conditions, the rubric score and the vote. Filed, and not edited afterwards.
Monitoring
A short note after each set of results: what the company reported against what we assumed, whether any exit condition has been met, and whether the thesis is intact, weakening or broken. Two paragraphs is enough. Silence is not.
Re-underwrite
At the annual review every holding is re-argued from scratch. The decision record is read out first, so the comparison between what was assumed and what happened is unavoidable.
The idea log
Over two or three years this becomes the most valuable document the fund owns. It records how the team’s judgement changed, stops the same idea being researched from scratch every year as members turn over, and makes visible the ideas that were right but rejected.

A fund that buys, holds, and never compares its assumptions to reality is not learning anything.

Research Standard v1.0, §7.3