The scoreboard
How calls are scored
Every rule below was published before a single call came due. That is the entire point of the page.
Published August 5, 2026. First call scores on July 14, 2027.
Anyone can publish a price target. The part almost nobody does is come back a year later and mark it, in public, including the ones that went badly. A target that is never marked to market is not research, it is marketing.
The problem with promising to score your own calls is that the rules can be written afterward, once you know which ones worked. A generous benchmark here, a longer horizon there, and any record can be made to look good. So the rules are set out here in advance, with dates attached, while every call on the board is still open and the outcomes are unknown. If these rules ever change, the change will be dated and the reason given, and calls already open will continue to be scored under the rules in force when they were made.
The horizon
Every call is scored twelve months after publication, on the first US trading day on or after the anniversary. The horizon is fixed at publication and does not move. A call is not scored early because it worked, and it is not given extra time because it did not.
Calls stay on the coverage board permanently, before and after scoring. Nothing is removed.
What gets measured
Two separate things, because they answer different questions.
1. Direction, against a benchmark
The security's total return (price change plus dividends paid over the window) is measured from the reference price stated in the report, and compared against the S&P 500 total return over exactly the same window. The benchmark is named here, now, and applies to every call.
Measuring against a benchmark rather than against zero matters. A buy that returns 6% in a year the market returned 22% was not a good call, and scoring it as a win because the number is positive would be dishonest.
| Rating | Counts as correct when | Reasoning |
|---|---|---|
| BUY | Excess return is positive | A buy claims the security is worth owning in preference to the index. Beating the index is the test. |
| HOLD | Excess return is below +5 percentage points | A hold claims there is no edge at this price. It is wrong if the security went on to meaningfully beat the market, and right if it matched or lagged. |
| SELL | Excess return is negative | A sell claims the security will lag the index. Nothing else is being claimed, and short outcomes are not modeled. |
2. Fair value error, which is a different question
Getting the direction right while badly missing the magnitude is luck wearing a suit. So alongside the hit or miss, each scored call records how far the fair value estimate sat from the actual price on the scoring date, as a percentage of the fair value. Direction tells you whether the call was useful. Fair value error tells you whether the analysis was calibrated. Both get published.
Edge cases, decided now
- Stock splits and stock dividends. All prices are split-adjusted. Reference prices are restated on the board with the adjustment noted, because unadjusted per-share figures are the most common way a track record quietly becomes wrong.
- Acquisitions that complete. Scored at the consideration received on closing, and marked as a special situation. These are reported but excluded from the headline hit rate, because a takeover tests deal flow rather than the thesis.
- Delisting or bankruptcy. Scored at the last available traded price, or at a total loss if the equity is canceled. No exclusions.
- A rating changed before its scoring date. The original call is still scored at its original twelve-month date, at the reference price it was published with. The new rating starts a fresh twelve-month clock as a separate entry. Both stay on the board. Changing your mind is allowed; deleting the earlier opinion is not.
- A report withdrawn. The entry remains on the board, marked withdrawn, with the date and the reason stated. It is not scored, and it is not deleted. This has already happened once, on the Money side, and it will be handled the same way if it ever happens to a research call.
- Trading halts on the scoring date. The next available close is used, and the substitution noted.
How results get published
On the Monday following each scoring date, the coverage board entry changes from open to scored and shows the outcome, the total return, the excess return against the benchmark, and the fair value error. The result is published whether it is good or bad, in the same place and the same format either way.
Nothing is edited silently. If a figure on this site turns out to be wrong, the page is corrected and the correction noted openly rather than quietly changed. Errors can be reported to contact@twocommainvestor.com and corrections take priority over everything else in that inbox.
The open calls and their dates
These are committed. Each will be marked on the date shown, under the rules above.
| Ticker | Published | Rating | Reference | Fair value | Scores on |
|---|---|---|---|---|---|
| BKNG | July 14, 2026 | BUY | $175.52 | $240.00 | July 14, 2027 |
| QCOM | July 20, 2026 | BUY | $171.78 | $205.00 | July 20, 2027 |
| COST | July 27, 2026 | HOLD | $935.03 | $900.00 | July 27, 2027 |
| TSM | August 3, 2026 | BUY | $404.25 | $485.00 | August 3, 2027 |
| TPL | August 10, 2026 | HOLD | $340.65 | $321.00 | August 10, 2027 |
| PLTR | August 15, 2026 | HOLD | $174.04 | $163.00 | August 15, 2027 |
| UBER | August 18, 2026 | BUY | $74.99 | $91.00 | August 18, 2027 |
| PGR | August 22, 2026 | HOLD | $219.28 | $208.00 | August 22, 2027 |
| NKE | August 31, 2026 | HOLD | $38.44 | $32.00 | August 31, 2027 |
What the scoreboard will not tell you
This part matters as much as the rules, and it cuts against the site's own interest, which is why it is here.
A handful of calls is noise, not evidence. With nine open positions, and even with fifteen, a hit rate says almost nothing about skill. Randomly selected stocks beat the index roughly half the time. A short run of wins is the single easiest thing in this business to mistake for ability, and the temptation to advertise one is exactly why the rules above are fixed now.
Treat the record as meaningful only once it covers a few dozen calls across a full market cycle, including a bad one. Until then it is a discipline, not a claim. The scoreboard exists so that the analysis can be checked, and so that being wrong carries a cost that is visible to everyone. It is not a performance advertisement, and no particular outcome is promised.
It is also not a portfolio. Ratings are opinions on individual securities at a stated price on a stated date. They are not weighted, they assume no position sizing, and no return shown anywhere on this site is a return anyone actually earned.