← Building the Desk

Building the Desk

Two wins, three losses, and still ahead on paper

The paper scorecard is two wins and three losses, and still slightly ahead, because every loss stopped at one unit of risk.

Our paper scorecard now holds five scored results. Two are wins and three are losses. That is a 40 percent win rate, and the paper book is still up $171.67. This piece walks through why both of those numbers can be true at once, and why we think the second number matters more than the first.

Everything here is paper research. No real money was at risk in any of these plans, and our live scorecard still shows zero scored trades and $0. The paper numbers come from the Desk scorecard as of Thursday morning, Nov 12. The results come from a paper observation model with no fees, liquidity or real fills.

What the five results are

The two wins came first. A Nebius paper long reached its target for a paper gain of $490, which the Desk scores as 2.0R. A Tesla paper long was closed by hand for a paper gain of $329.20, about 1.37R.

The three losses came after. A second Nebius paper long filled and hit its stop on Nov 2 for a paper loss of $248.50. A Microsoft paper long filled on Nov 5 and hit its stop the same morning for a paper loss of $242. An Nvidia paper long filled on Nov 9 and hit its stop the same morning for a paper loss of $157.03. Each of the three was scored as exactly one unit of risk lost, which we write as -1R.

Add it up and the two wins total $819.20 on paper, the three losses total $647.53, and the difference is $171.67. The average result across all five is about 0.07R.

What R means and why we track it

R is the amount we planned to risk on a trade if the stop is hit. A result of 2.0R means the trade made twice what we were prepared to lose. A result of -1R means the stop did its job and we lost what we planned to lose, and no more.

That is the whole reason the book can lose more often than it wins and still sit slightly ahead. The wins were larger than one unit of risk, and every loss stopped at one unit. None of the three losing plans was allowed to drift into a bigger loss.

What we are not claiming

A 40 percent win rate with an average near 0.07R is a thin edge on five results. Five is far too few to call a track record. The last three scored fills were all stops, so the recent run is three losses in a row, and we are not going to dress that up. What the scorecard shows is that the risk rules held on every losing plan. That is a process result, not a profit claim.

A small build fix that unblocked the Desk

One build note from the same stretch. On the evening of Nov 9, the Desk refused to save a new paper plan. The server returned an "Invalid level label" error because one of the chart labels on that plan ran to 127 characters, and the limit is 60. Four labels on that draft were over the limit. The save was not retried.

The next morning, every chart label was kept at 60 characters or fewer. The longest was 53. Both paper plans saved on the first attempt. The takeaway is a plain one. When a save fails, we read the error, fix the cause and try again on the next run, instead of forcing the same request through.

The takeaway

When you look at any scorecard, ours included, look past the win rate. Ask how big the wins were compared with the losses, and whether every loss stopped where the plan said it would. On our paper book so far, the answer to the second question is yes on all three losses. Whether the edge is real will take many more results to know.

DePappa Capital trades only its own capital. Nothing here is investment advice or a recommendation to buy or sell any security. Paper results are simulated, exclude fees and real fills, and do not predict future results.

AI-assisted editorial preparation. DePappa Capital is responsible for its published work. This journal is not a live signal service, an investment recommendation, or a verified track record.

More from the journal →