Daily budget
$4,000
4% of the initial balance, fixed
troid's research · strategy review · tearsheet · methodology
These results are based on simulated or hypothetical performance results that have certain inherent limitations. Unlike the results shown in an actual performance record, these results do not represent actual trading. Also, because these trades have not actually been executed, these results may have under-or over-compensated for the impact, if any, of certain market factors, such as lack of liquidity. Simulated or hypothetical trading programs in general are also subject to the fact that they are designed with the benefit of hindsight. No representation is being made that any account will or is likely to achieve profits or losses similar to these being shown.
troid's own strategy shows no statistical edge. Out of sample, on 504 BTC trades from 1 January 2021 to 7 January 2026, which the parameters never saw, it measures +0.008R per trade, standard error 0.016R — a MEASURED figure, inside noise, and not a fact about the future.
Five-tranche strength ladder + regime filter + desk cap — but do not buy a challenge. The walk-forward has been run, and it came back as noise.
Of everything tested it has the lowest drawdown (−$1,058 against −$7,295 for single entry), zero failed and zero zombie accounts across 72 rolling challenge starts, median ending balance $100,409. It is also still +0.03R on the sample it was chosen on, and +0.008R on the 504 BTC trades from 1 January 2021 to 7 January 2026 it never saw (WALKFORWARD.md). That is a well-built vehicle with no engine in it.
Why not: a $999 challenge buys one noisy binary outcome. Five years of holdout just said the entry logic has no measurable edge. At a 14% industry base rate, passing once barely updates whether you have an edge — plenty of edgeless traders pass.
At the same $500 risk, a 5-tranche strength ladder holds 34% less quantity than a single entry (0.109 BTC vs 0.167 BTC; $8,524 vs $13,000 notional). Because the stop stays anchored at T1's structure, each later tranche sits further from the stop and therefore gets less size.
The ladder's benefit is conditionality, not size: losers fill 1.66 tranches, winners fill 4.26. You're small when wrong and large when right. If what you actually want is a bigger position, that comes from raising risk per trade, and the ruin math says don't.
income_math.py ran at 30 trades/month. The strategy produces
10.1/month (78 trades in 7.7 months). Every encouraging number in the
$5k/month analysis — the 4-months-to-target row, $614 at the real 9/mo frequency (was $1,862 at an assumed 30/mo)/account — assumed a frequency
this strategy does not reach.
Corrected: at 10.1 trades/month and 0.5% risk, $6,250/month gross needs +1.24R per trade. Not credible at any skill level. Either the timeframe drops below 4h to raise frequency, or the income target needs more accounts than the scaling table implied.
It was hand-written and drifted from the engine after every change. It is now
generated by gen_results.py from the live engine on the frozen sample, and
verify_claims.py re-derives the published statistics from
journal.csv and fails if either drifts. The table below is that output.
bitfunded.md presents it as broadly dangerous. Measured: it needs
both uncapped risk ≥4% per position and a policy that holds
losers through the reset. Under desk sizing plus swing_safe it is dead twice over —
by cap and by policy. Largest floating loss actually carried into a reset was $3,422
against the $4,000 line, at 4% naive risk.
On TP3 it fired 4 times in 8 months, because only ~8% of trades reach 3R. On the stop-out it fires 24 times in the 78-trade sample, which is a sample rather than an anecdote — but it is still one asset over one regime, and the holdout (1 January 2021 to 7 January 2026) says the whole configuration is inside noise.
All of it is in kunjancollective/troid, and troid's ledger runs daily from
.github/workflows/shadow.yml: fresh bars, replay, journal, ledger, commit.
The journal's logged_utc column is the timestamp nothing can backdate.
income_math.py's withdraw=False arm reports $0 by
construction — it counts only withdrawals as income, so the "bank profits to grow the
buffer" comparison can't be sized. And risk.py accepts
--json only before the subcommand.
| Configuration | n | exp R | PF | max DD | fills W/L |
|---|---|---|---|---|---|
| v1 pullback, no filter | 89 | −0.05 | 0.91 | −$8,598 | — |
| + regime filter | 75 | +0.03 | 1.04 | −$7,332 | — |
| + breakout, single entry | 76 | +0.01 | 1.02 | −$7,295 | — |
| 3-tranche weakness ladder | 43 | −0.22 | 0.47 | −$5,829 | 1.53 / 2.62 |
| 5-tranche weakness ladder | 33 | −0.31 | 0.14 | −$4,973 | 1.22 / 3.42 |
| 3-tranche strength ladder | 54 | −0.03 | 0.83 | −$3,147 | 3.00 / 1.54 |
| 5-tranche strength ladder | 57 | +0.00 | 1.01 | −$1,931 | 4.26 / 1.66 |
| 5-strength + reverse on TP3 | 60 | +0.02 | 1.14 | −$2,044 | 4.10 / 1.64 |
| 5-strength + reverse on stop | 78 | +0.03 | 1.29 | −$1,058 | 4.48 / 1.64 |
| 5-strength + reverse on any | 88 | −0.01 | 0.90 | −$1,488 | 4.24 / 1.63 |
Ladder direction is the single largest structural effect found: same signals, same stops, only the side the tranches sit on. Weakness at 5 tranches is the worst profit factor in the project (0.14); strength at 5 with the reverse on stop is the best drawdown (−$1,058).
The budget cap is not conservatism — it is the only thing that makes the account survivable. At +0.35R, a genuinely excellent system, naive sizing at 1% blows the account 68% of the time within a year, and at 2% it's 98%. Every one of those cells is 0.0% under the cap.
RESULTS.md with corrected numbers; revise
bitfunded.md's rollover section to the measured condition.HANDOFF.md.
Done for the baseline: inside noise. Next is troid-shadow-2.Sample figures from 1,539 4h bars, Jan 8 – Sep 21 2026, api.binance.us — one
asset, one regime. The holdout, 1 January 2021 to 7 January 2026 (backtest/WALKFORWARD.md) came back
inside noise on BTC and ETH.
troid's desk · troid's compare · troid's ledger · troid's research · tearsheet · ask troid · faq · terms · source · x · reddit
troid is a free informational tool, not financial or investment advice. troid does not place trades or publish trade signals. Its shadow account is a simulated strategy, shown only after trades close. Prop-firm rules change without notice — verify every number with the firm before trading. troid is an independent affiliate of the firms it compares and earns a commission on purchases through its links; this does not affect the calculations or comparisons. Crypto Fund Trader does not own or operate this website and is not responsible for its contents. It is owned by troid, an independent marketing affiliate.
© 2026 Kunjan Collective LLC