Live · real capital
Live Track Record
Monthly returns on real capital, net of all commissions, fees, and slippage. Updated on a fixed date each month, append-only: prior months are never restated or removed, and every tracked strategy is shown, including its losing months.
Shown in percent (the dollar capital base is intentionally not published). Cumulative curves are indexed to 100 at each strategy's start.
Starting the clock: July 1, 2026
As of July 1, 2026, I began tracking a live, real-capital record for the three strategies below. I fixed this exact strategy set in a public, timestamped commit on June 15, 2026 — sixteen days before any capital went live — so it was locked on the record before a single result existed. (The earliest a completed July number can appear is early August.) You can check that pre-registration on the Verify page; nothing here is backdated.
July 1 is the first day I deployed real capital in fresh, dedicated accounts that trade only these three portfolios— one account per strategy, nothing commingled. I built and tested earlier versions of these strategies before then, live and on paper, during development. That development period is not part of this record and I'm not presenting it as one; July 1 simply marks the clean, purpose-built start this record measures. I'm telling you it existed rather than pretending the strategies appeared from nowhere.
From here, the rules I hold myself to on this page:
- All three, every month, including losers. No strategy is ever dropped and no bad month is ever skipped.
- Append-only, never restated.Once a month is posted it's permanent. A script blocks any commit that would edit or delete prior history.
- Fixed cadence. Posted on a set date each month regardless of the outcome.
- Real, net, in percent. Realized returns on my own capital, net of all commissions, fees, and slippage. Percentages only; the dollar base stays private.
- Auditable. The data and its full history are version-controlled, so the record — and the timestamp on this very note — can be independently verified.
You don't have to take my word for the timestamp on this note — see Verify for how to check it independently. This is a transparency exercise about my own trading. It is not an offer, not a solicitation, and not investment advice. See Disclosures.
Expected versus actual
I care about my simulations matching reality. I don't really care whether they make or lose money in any given month, because that is not what the statistics expect. I care that they did what I expected them to do. At the end of the day it is just statistics, and I need them to follow the statistics.
So the column worth judging me on below is not what the account did. It is the gap.
| July 2026 | Model said | Account did | Gap |
|---|---|---|---|
| Core Long Term | -7.9% | -10.8% | -2.3% |
| Intraday Futures Breakout | -7.5% | -7.0% | matched by luck |
| Intraday Volatility Breakout | tracked | +2.9% | clean |
| QQQ (Nasdaq-100), same month | · | -6.6% | for context |
Model figures are the same code run over the same July window. July was a down month for the index, and two of these three are meant to lose money in a month like that.
That last row is the point. The market fell 6.6% in July. Two of my strategies were modelled to lose money in a month like that, and they did. The losses are not the story. The gaps are.
Core Long Term
I failed to close all positions before re-initializing, and 80 positions were left open and untracked. Whoops. So the strategy returned about 2% worse for the month than it should have. I have cleaned that up, and its daily returns have matched simulation since. This one is correlated to the index by design, being a long term strategy, so the loss itself was expected. Straying from simulation early in the month is what annoyed me. Hopefully in future months it matches reality cleanly. That matters more to me than the number.
Intraday Futures Breakout
This one actually lost less money than it was supposed to, which sounds good and is not. It only took 4 of the 11 trades it should have. I have been battling warmup issues, where data does not arrive as expected and it misses trading days, and margin was being miscalculated, so orders I expected to fill never happened. Matching the model by accident is not matching.
Its win/loss ratio is 3.7 and its win rate is 32%. It does not win often, but the payoff is worth it when it does. In practice that feels like a slow bleed and then a large gain, which is not the most fun thing to watch. It is the more volatile of the two intraday strategies. I just have to make sure it is doing what I expect.
Both faults are rebuilt and tested. The sizer now works from margin measured against the broker rather than assumed, and over a six and a half year test it takes zero rejected orders across 639 entries, against 18 before. The warmup now tries several data sources and retries every morning until it arms, so a failed start costs one session instead of three weeks. Between July 15 and August 2 it placed no orders at all while the volatility estimate rebuilt itself the slow way. It armed on its own and took its first trade in three weeks on August 3. I am watching it fill correctly before I trust it again.
Intraday Volatility Breakout
+2.9% in July, in a month the Nasdaq-100 fell 6.6%. No rejected orders, no downtime, no divergence from the model. It has run on real capital since June 27 and is up about 9% over that stretch.
What makes it work is unglamorous: a 54% win rate and a 1.06 win/loss ratio. It wins slightly more often than it loses and slightly bigger than it loses. That is the whole edge, and it is kind of funny that something that ordinary turns into anything interesting. Its measured beta to the Nasdaq-100 is negative 0.29, so it is not a leveraged version of the index. It is a separate return stream, which is exactly why it can make money in a month the market falls.
What a normal year is supposed to look like
A losing month only means something if you know what the strategy is supposed to do in a bad stretch. So here is what the simulations say to expect in a typical year. These are drawdown episodes, measured peak to trough and counted until the equity curve makes a new high.
| Per year, expect | 5% or worse | 10% or worse | 20% or worse | Worst seen |
|---|---|---|---|---|
| Core Long Term | 5.5 | 2.7 | 0.9 | 31.7% |
| Intraday Futures Breakout | 1.7 | 0.6 | 0 | 17.5% |
| Intraday Volatility Breakout | 3.1 | 0.7 | 0.2 | 27.9% |
From backtests: Core Long Term 2020 to 2026, Intraday Futures Breakout 2020 to 2026 on the live configuration, Intraday Volatility Breakout 2018 to 2026. Simulated, not realised, and the future is free to be worse.
Read that against July. Core Long Term is supposed to see a 10% or worse drawdown roughly two or three times a year, with a median depth around 16%. So a 10.8% month is not an anomaly, it is Tuesday. Intraday Volatility Breakout is supposed to spend a fifth of its years inside a 20% hole. If I panic-sold the first time one of these had a bad month, I would be trading a completely different strategy than the one I tested.
That is the whole reason I write the expected numbers down before the month starts. Otherwise a normal drawdown feels like a broken system, and a broken system feels fine as long as it happens to be up.
How I am judging this
I am not so emotional about any of this. I am young. Figuring out what actually works is all that matters, and then I can allocate more of my money where I feel comfortable. What I expect is that I will enjoy putting money into the long term strategy that is correlated to the broader market, while the intraday strategies give me uncorrelation and a way to make money in other market regimes.
The numbers can be hard to wrap your brain around. So I have to make sure the statistics and the simulations are working out as expected. That is the whole job right now.
Core Long Term
View strategy →︎Cumulative · indexed to 100
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Year |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 | · | · | · | · | · | · | -10.8% | · | · | · | · | · | -10.8% |
Intraday Futures Breakout
View strategy →︎Cumulative · indexed to 100
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Year |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 | · | · | · | · | · | · | -7.0% | · | · | · | · | · | -7.0% |
Intraday Volatility Breakout
View strategy →︎Cumulative · indexed to 100
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Year |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 | · | · | · | · | · | · | +2.9% | · | · | · | · | · | +2.9% |
Important disclosures
This site presents my own personal, proprietary trading results. It is for informational purposes only and is not an offer to sell, or a solicitation of an offer to buy, any security or interest in any fund. I am not asking anyone for money.
Past performance is not necessarily indicative of future results. All results shown are on my own capital. Live results are shown net of commissions, fees, and slippage; backtests are models built in QuantConnect and are hypothetical.
Trading futures involves substantial risk of loss and is not suitable for everyone. Commodity trading involves substantial risk of loss.Any futures results shown are proprietary (my own trading) and are not combined with anyone else's.
Backtested or hypothetical results shown elsewhere on this site have inherent limitations, do not represent actual trading, and can differ materially from results subsequently achieved. They are clearly labeled as such and are kept separate from the live, real-capital track record.