Last updated: September 24, 2026. Version 2026-09-24.
1. What these numbers are
Every return, ranking and S&P 500 comparison in an ECSTI briefing is a simulated result.
ECSTI agents trade paper portfolios on ECSTI's own trading simulator. No real money is invested, no order reaches a real broker, and no one's brokerage account is involved. When an ECSTI message says an agent placed a paper trade, it means the agent placed a simulated order in its paper portfolio.
- Each paper portfolio starts with $10,000 of simulated cash by default.
- Simulated orders fill against 15-minute market data, one data cycle after the agent places them.
- Paper portfolios are long-only cash accounts: a simulated order can't borrow money or sell short.
2. How simulated returns are calculated
A simulated return is the percentage change in the value of an agent's paper portfolio, meaning its simulated cash plus its holdings at market prices, over a period.
An agent's value at a close is its paper portfolio's value as recorded shortly after that close. A period's simulated return runs from the close of the previous period's last trading day to the close of the period's own last trading day: the later value divided by the earlier one, minus one. Paper portfolios have no deposits or withdrawals, because the starting balance is fixed when an agent is created, so this ratio is the whole return. The S&P 500's return over the same period is measured between the same two closes.
Simulated returns are measured over calendar periods on the US market clock (America/New_York): a trading day, a week, a month, a quarter or a year. They are never measured over rolling windows, so "September 2026" means the same dates for every agent. A period closes at the NYSE close on its last trading day, including half days.
3. What simulated results leave out
Simulated results are not what real trading would have produced. The simulator does not include:
- commissions or other trading fees;
- bid-ask spreads;
- slippage, partial fills, or the effect of an order on the market price;
- taxes; or
- dividends, stock splits and other corporate actions, which are a known limitation of the simulator.
Dividends. A paper portfolio doesn't receive dividends. A holding is valued at its market price, which falls by roughly the dividend on the ex-dividend date, so an agent holding dividend-paying stocks shows a slightly lower simulated return than the same positions would have produced in reality. Because briefings report favorable milestones only, this understates rather than overstates: it can cause a milestone to be missed, not invented.
Real trading could produce materially different results from these simulated ones, including worse ones. A simulation also can't reproduce what it is like to trade with real money.
4. The S&P 500 comparison
- Comparisons of an agent's simulated return with the S&P 500 use the S&P 500 price index (symbol
^GSPC), over the same calendar period. - The price index leaves out dividends, so its return is lower than a total-return version of the index would be. Paper portfolios don't receive dividends either (section 3), so dividends are excluded from both sides of the comparison.
- An index is unmanaged, has no trading costs and can't be invested in directly. A paper portfolio holds far fewer stocks than the index, so an agent's simulated return can move much more than the S&P 500's.
5. Rankings
A ranking line reads, for example, "#3 of 412 ECSTI agents by simulated return, September 2026".
- The criterion. Agents are ranked by simulated return over the calendar period.
- Who is ranked. Every active ECSTI agent is ranked together, whoever owns it. An agent is ranked only if it was active for the whole period, meaning it already had a recorded value at the previous period's close and is still active when the period closes, and a simulated-return ranking is reported only when at least 20 agents were ranked.
- Surviving agents only. A simulated-return ranking counts only agents that still exist when the period closes. Agents deleted during the period drop out, so a simulated-return rank compares an agent with the survivors, not with every agent that ran.
- Ties. Rankings use unrounded simulated returns, so two agents a briefing shows with the same return can hold different ranks. Agents with exactly the same simulated return are ranked by age, the older agent first.
- Who computes it. ECSTI computes these simulated-return rankings. They are not an independent rating, and they are not an endorsement of any agent.
6. Milestones, and what they don't tell you
A briefing reports milestones that an agent's simulated results reached when a period closed:
| Milestone | Periods | Reported when |
|---|---|---|
| Simulated-return ranking | Month, quarter, year | The agent finished first, or in the top 10, by simulated return among at least 20 ranked agents, and was active for the whole period. |
| Simulated return above the S&P 500 | Month, quarter, year | The agent's simulated return was positive and above the S&P 500's return over the same period, and the agent was active for the whole period. |
| Simulated weekly streak | Week | The agent's simulated weekly return was above the S&P 500's for 3, 5, 10, 20 or 52 consecutive trading weeks. |
| Highest simulated return since launch | Day, week | The agent's simulated return for the day or week was its highest since launch, and positive. A day needs at least 20 trading days of history, and a week at least 8 completed weeks. |
Returns are compared as a briefing shows them. A briefing prints returns to one decimal place. When a milestone depends on a return being positive, above the S&P 500's, or above an earlier return, that comparison uses the one-decimal values, so the difference is always visible in the message. Rankings are the exception: they use unrounded returns (section 5).
Briefings report favorable milestones only. ECSTI doesn't send a message when an agent's simulated return falls below the S&P 500's, when a simulated streak ends, or when an agent drops out of a simulated-return ranking. A briefing is not a full record of an agent's simulated performance. Where it's available, a briefing shows each agent's simulated return since launch next to the S&P 500's over the same time, and the agent's page in ECSTI shows its record.
7. AI output
ECSTI agents run on AI models. Their simulated trading decisions can be wrong, the same agent can decide differently on different runs, and an agent can behave differently over time as its instructions or the underlying models change. Section 4 of the Terms of Use explains the limits of AI-generated output.
8. Not investment advice
Nothing on this page or in any ECSTI message is investment, financial, legal or tax advice, or a recommendation to buy or sell any security. A briefing describes what agents in your workspace did in their paper portfolios, not what you or anyone else should do.
ECSTI is not a broker, a broker-dealer or a registered investment adviser, and using ECSTI creates no advisory or fiduciary relationship.
9. No guarantee
Simulated results don't predict future results, simulated or real. Investing real money involves risk, including the risk of losing everything you invest.
10. Market data and corrections
Simulated results are computed from third-party market data, which can be delayed, incomplete or wrong.
ECSTI doesn't withdraw or reissue a briefing once it has been sent. A briefing describes what the data showed when the period closed, and it isn't recalculated if that data is later corrected — so a milestone a briefing reported can rest on figures that have since changed. Figures shown in ECSTI itself are drawn from current market data, so a briefing and the app can disagree about the same period.
11. Changes to this page
| Version | Date | Change |
|---|---|---|
| 2026-09-16 | September 16, 2026 | First version. |
| 2026-09-22 | September 22, 2026 | Confirmed that paper portfolios do not receive dividends, and said so in section 3 with the direction of the resulting error. Noted in section 4 that dividends are therefore excluded from both sides of the S&P 500 comparison. Stated in section 10 that a briefing is never withdrawn or reissued after sending, and may therefore disagree with the app once market data is corrected. |
| 2026-09-24 | September 24, 2026 | Stated how simulated returns and the S&P 500 comparison are measured (section 2), who is ranked, that rankings use unrounded returns and how exact ties are ordered (section 5), and that other milestones compare returns as a briefing shows them (section 6), as the achievements job computes them. |
12. Contact
Questions about how ECSTI calculates simulated performance: support@ecsti.io.