HOW PARRA WORKS IT OUT · TRACK RECORDS

How track records are kept

Every thesis and every call is scored against the market from the day it was made, with the price of that day. Nothing is backfilled, and misses stay on the record.

Method as of 25 Oct 2026 · What changed

Measures
Whether a thesis or a call did better than the market over the same days
Who makes it
Code, from daily closing prices. No AI decides whether something was right.
Updated
After every trading day’s close

Your theses

How a thesis is scored

  1. Day 0

    You open it. Entry is that day’s close.

  2. Day 5

    First verdict: ahead or behind the S&P 500

  3. You close it

    Played out if it beat the S&P 500, else missed

BenchmarkThe S&P 500 (SPY), over exactly the same days
Entry and exitThe close on the day you open the thesis and the close on the day you close it, or today while open
First verdictAfter 5 trading closes. A day’s move is noise.
OutcomeOpen: on track or behind. Closed: played out or missed. A tie counts as behind.
Left outWatchlist ideas, which are not positions, and theses with no price for their dates (counted and shown)

Voices you follow

Parra reads posts that say a company will rise or fall. A call counts once: repeating the same view is not a new call; only a change of mind is.

EntryThe close on the day of the post
Read at30, 90, 180 and 365 days
A hitThe stock beat the S&P 500 for a rising call, or trailed it for a falling one
TrustHits at 90 days, pulled toward 50% while there are few calls: (hits + 1) ÷ (calls + 2). One lucky call can’t make a voice look perfect.

Parra’s own analyses

Parra keeps the same score on itself. Each analysis is read against the company’s sector fund, and its stories are checked once a year has passed.

Read at3, 6 and 12 months, against the sector ETF
A hitA buy rating when the stock beat its sector ETF, a sell rating when it trailed it, a hold when it stayed within 10% of it
StoriesAfter a year: which story the company landed nearest, whether it stayed between bear and bull, and a Brier score for the likelihoods
CalibrationStories given about 30% should land about 30% of the time. The record shows how often they did.
Worked example: a thesis on Northwind Semis

Opened on 3 March at a close of $80. The S&P 500 closed at 500 that day.

Closed on 1 September at $92 (+15%). The S&P 500 was at 540 (+8%). The thesis played out, 7 points ahead.

A made-up company, to show the method. Not a view on any stock.