Results
The live account has been trading since 7 September 2026. Only what actually happened on it is here — no calculations, no assumptions. The five-year history and its checks are computed on the past and live on the Backtests page.
The account right now
Data as of 26.10.2026, 17:00
What the live account shows
The lines by which people check someone else’s statistics — with the same figures from the historical calculation beside them, so it is clear whether live matches computed.
Snapshot as of 26.10.2026 · updated on Saturdays
| live account | computed | |
|---|---|---|
| trading since | 07.09.2026 | 08.2021 — 08.2026 |
| starting deposit | 1 000 | 10,000 per source |
| trades | 92 | 3 887 |
| profit since launch | +5.0% | +1198% |
| average trade | +0.05% | +0.31% |
| winners | 34% | 43.0% |
| profit factor | 1.07 | 1.6 |
| average RR | 2.11 | 2.07 |
| max drawdown | 22.5% | 25.8% |
Over the first months the figures will swing either side of the computed ones, and that is normal: over the first dozen trades any ratio is chance — 57% winners out of seven trades is not a win rate, it is a coincidence. The point of the table is not to compare it today, but to see where the figures go once the account has built up a history. One more difference worth knowing: the live account risks one percent of its current balance — the stake grows with the account, which is compound interest. The calculation is made from the initial deposit, without reinvestment. So the «profit since launch» line cannot be compared head-on: over a long distance compounding lifts the live account above the computed curve, and in a drawdown drags it below.
Weekly log
Two numbers are under control here. Drawdown answers how painful it is right now; expectancy answers whether the system still works: the average result per trade over the last hundred, as a percentage of the deposit, with a −0.05% threshold. Until there are a hundred trades, an average over them is noise, so the column reads “not enough data”.
| Week | Trades | Result | Since launch | Drawdown from peak | Expectancy | Control |
|---|---|---|---|---|---|---|
| 2026-10-19 — 2026-10-25 | 11 | -28 | +6.0% | 21.6% | not enough data | normal |
Previous weeks — 6
| Week | Trades | Result | Since launch | Drawdown from peak | Expectancy | Control |
|---|---|---|---|---|---|---|
| 2026-10-12 — 2026-10-18 | 14 | -44 | +8.7% | 18.8% | not enough data | normal |
| 2026-10-05 — 2026-10-11 | 11 | -3 | +13.2% | 14.3% | not enough data | normal |
| 2026-09-28 — 2026-10-04 | 16 | -121 | +13.4% | 14.1% | not enough data | normal |
| 2026-09-21 — 2026-09-27 | 17 | +237 | +25.5% | 2.0% | not enough data | normal |
Author’s note on the weekThe week closed with a profit of $237: 17 trades, 11 of them winners. Growth since launch rose from 1.9% to 25.5%, drawdown from the peak — 2.0%. This week should not be taken as the norm. In units of risk it returned +21.5R — it earned twenty-one stakes. The five-year calculation contains 262 full weeks, and a result of this size appeared in them twice: roughly once every two and a half years. An ordinary week returns +3.2R. I count in stakes rather than in percent of the account on purpose. The live account risks one percent of its current balance — as it grows, so does the stake; that is compound interest. The calculation is made from the initial deposit, without reinvestment. In percent the two are not comparable; in stakes they are. Where the result came from is visible in the trades: three out of seventeen produced 79% of the total, and a single gold trade — 45%. The win rate came out at 65% against the calculated 38.4%. This is not a new quality of the system but a lucky coincidence of several entries; both figures will return to their averages. I show weeks like this exactly as I show losing ones: with numbers and without excitement. What matters is not how much a single week brought, but where the account is heading over the distance — that is what the control thresholds are for, and they have not changed. | ||||||
| 2026-09-14 — 2026-09-20 | 10 | -75 | +1.9% | 11.6% | not enough data | normal |
Author’s note on the weekThe week closed down 75 dollars. Cumulative account growth fell from 9.4% to 1.9%, and the drawdown from the peak is 11.6%. This is an expected figure, and here is the yardstick. The thresholds I watch were computed by shuffling the history day by day on the current sixteen-source portfolio: the median drawdown comes out at 24.6% and the 95th percentile at 36.3%. A drawdown twice as deep as the current one is the ordinary case, not a breakage. The first threshold at which I check anything at all sits at 25%. Ten trades in the week is the usual flow. The monitoring rules call for no action in this range: up to 25% is ordinary operation. | ||||||
| 2026-09-07 — 2026-09-13 | 12 | +94 | +9.4% | 4.1% | not enough data | normal |
Author’s note on the weekThe week closed with the account up 9.4%, but one number in it needs explaining — risk per trade. In the settings it is one percent of the account. On currencies and oil that is what came out: 0.89–0.95%. On gold it came out at 1.49% and 1.84%. The cause is not the advisors but the minimum lot. A stop on gold costs $15–17, and no position is smaller than 0.01 lots: when the calculation calls for 0.006, what opens is 0.01 — and the risk ends up larger than stated. The more expensive an instrument’s stop, the stronger the effect, and on a $1,000 account gold alone gives one and a half percent. I decided to change nothing. The week’s drawdown was 4.1% against a 25% attention threshold, and intervening for the sake of a formal percentage would have broken a portfolio composition tested over five years. But when I count risk on gold I now work from 1.5%, not 1% — including where the trading-halt thresholds are concerned. On accounts from $2,000 up the minimum-lot floor disappears. | ||||||
Who verifies this
The same figures are counted by three counters, and none of them is counted by me: the copy-trading platform itself, Myfxbook — an independent service connected directly to the broker account — and MQL5 with its own formulas.
The strategy card on Tickmill
Ranking position, balance, return and drawdown are calculated by the platform where copying takes place. Counted from 7 September 2026.
Open the card on the platform → counted by the platform
The account, verified by Myfxbook
These figures are taken from the account not by the strategy author but by Myfxbook, via the investor password — read-only. The curve updates itself; click to open the full statistics.
Open full statistics on Myfxbook →updates automatically
The account is live, held at Tickmill and connected to Myfxbook through an investor password — read-only, it cannot be traded on. Both of the service’s checks are passed: the track record is verified against data it takes from the broker itself, and the access rights are confirmed. Closed trades are visible in full; open positions are hidden — otherwise the strategy could simply be repeated by hand without subscribing. The count starts on 7 September 2026.
The same account on MQL5
The same trading, but the figures are calculated by MQL5 using its own formulas. Subscription there is closed — the account leverage is higher than the platform allows for copying — so the card works only as a cross-check.
How risk is watched
Oversight is run by a watchdog algorithm, not by a person in a given mood. It runs on the same server, reads the account history and compares two numbers with thresholds set in advance: the current drawdown and the expectancy over the last 100 trades. It makes no trading decisions — it only measures and reports.
The first parameter is drawdown. Thresholds: up to 24% — normal, 24–35% — attention, 35–42% — review, above 42% — trading stops. The current level is shown in the weekly feed above.
The second is expectancy over the last 100 trades: what an average trade brings in. Drawdown answers “how much it hurts right now”, expectancy answers “does the system still work”. The alert threshold is below −0.05% of the deposit per trade. Computing it before a hundred trades have accumulated makes no sense — an average over ten is noise — so while the account is young the watchdog gathers statistics and stays quiet.
Parameters are not changed during a drawdown: the settings were fitted on all available history, and tuning them to a bad month would break what took years to verify. Where the thresholds came from, how they were tested and why there are no intermediate steps — on Backtests.
If my side fails. The server can go down, the connection can drop, and I can be unreachable. What matters here is that the stop-loss of every trade sits on the broker's server, not in the expert advisor: it will fire even if everything of mine is switched off. The same goes for the target — it is placed together with the order. Only the time-based exit depends on my side: for Algorithm 1 that is the close at the end of the session, for Algorithm 3 a 20-minute hold. On a failure such a position closes by stop or target, like any other, rather than on time. Copying is executed on the broker's side, so your terminal is not needed for it. And you can close positions or switch the subscription off yourself at any moment — the money and the access are with you the whole time.
If a stop threshold is reached, nothing happens to your account on its own: new trades simply stop appearing. The money and any open positions stay under your control, while I look into what happened and report what I find.