This is our actual research desk — every strategy, every backtest, the wins and the losses, exactly as they happened. Most services hide this page. We lead with it — because a track record you can check is worth more than any promise.
Tested with the same rigour whether the answer was yes or no.
Read the stages, not the sum. These rows span paper, testnet and live — different sizing, some not yet real money — so adding them up as one P&L is meaningless. This window: 7 wins, 13 losses. Improved from the previous fortnight, but still losing. Every stop — all 13 — fired at a pre-set level, none compounded. The Sep 3 cluster was sharp: 9 stops across four strategies in a single session as markets moved against SHORT positions. Money-Flow Long (paper) closed XAG +2.09% and XAUT +1.00% via take-profit. Smart-Money I closed INJ +0.82R via trailing stop on Sep 6. The live Liquidation-Zone (n=16) made no new entries — its most recent tape close remains XRP +9.9% on real capital. The sample is still building; stop discipline is demonstrable, edge confirmation takes more data.
The market context the discretionary overlay watches. A point-in-time snapshot — the specific entry orders, stops and targets behind it stay private.
| Asset | Watch zone | Position in zone | Status |
|---|---|---|---|
| BTC | $76.6k – $88.8k | 25% | Inside · Wait |
| ETH | $1,515 – $2,614 | 90% | Inside · Wait |
| SOL | $93.1 – $118.8 | 49% | Inside · Wait |
| BNB | R: $791–$929 · S: $601–$746 | — | Between zones |
| AVAX | R: $8–$10 · S: $6–$7 | — | Between zones |
| LINK | $11.4 – $14.4 | 63% | Inside · Wait |
| XRP | $1.27 – $1.55 | 51% | Inside · Wait |
| DOGE | $0.07 – $0.10 | 65% | Inside · Wait |
Watch zones = multi-week consolidation bands the overlay tracks for macro entries. Snapshot as of 2026-09-07; the specific ladder orders, stops and targets are not published.
Our strongest live performer so far. Real capital, small size, every trade published.
Read this number honestly. The +49% is the sum of per-trade returns on the margin deployed per trade — leveraged, deliberately small size while the strategy proves itself. It is not an account-level return, and 16 live trades is still an early sample. What it shows: the system is doing what it was designed to do — losses stopped early (−2.8%, −4.0%), winners allowed to reach target (+10%). Backtest for this strategy: 981 trades, PF 1.89, so live is currently outperforming its backtest — early samples flatter; we publish it either way.
The strong part first: 11 strategies cleared rigorous backtesting on years of data. The honest part: a strong backtest doesn't always survive live yet — and we show that gap rather than hide it. (Plain names; exact parameters stay private.)
| Strategy | Style | BT n | BT WR | BT PF | Live n | Live PF | Status |
|---|---|---|---|---|---|---|---|
| Higher-Timeframe Momentum | Trend | 39 | 46% | 1.31 | 32 | 0.83 | Paper · under kill review |
| Trend (Alligator) | Trend | 82 | 42% | 1.34 | 71 | 1.02 | Paper |
| Liquidation-Zone Reversion ★ | Mean-rev | 981 | 70% | 1.89 | 16 | 2.64 | Live · flagship |
| Liquidity Mean-Reversion | Mean-rev | 874 | 40% | 4.84 | 0 | — | Paper |
| Smart-Money Concepts I | SMC | 2,299 | 38% | 1.22 | 52 | 0.78 | Testnet |
| Smart-Money Concepts II | SMC | 490 | 36% | 1.18 | 28 | 0.25 | Testnet |
| Money-Flow Long | Trend | 14 | 50% | 1.19 | 25 | 7.53 | Paper |
| CME Gap Fill | Macro | 32 | 86% | 1.39 | 0 | — | Paper |
| ETH Funding Fade New | Carry | 34 | 59% | 1.96 | 0 | — | Paper |
| Alt EMA Cross New | Trend | 77 | 46% | 1.67 | 0 | — | Paper |
| BTC Macro Ladder | Macro | — | — | — | 0 | — | Live · testnet |
| Multi-leg Momentum (Cascade) | Mean-rev | 52 | 37% | 1.05 | 13 | 0.67 | Killed |
Flagship change (2026-09-11). Higher-Timeframe Momentum was our lead strategy; its live profit factor has fallen below its pre-committed threshold (PF < 1.5 at n=20 — now 0.83 at n=32), so per its own kill criteria it is demoted and under review. The book now leads with Liquidation-Zone Reversion — live on the desk's own capital since May with PF 2.64 (n=16, early sample). This is the system working as designed: rules decide, not attachment.
WR = win rate · PF = profit factor (gross wins ÷ gross losses) · BT = backtest · n = trade count. PF below 1.0 means losing. ★ = the strategy the copy book leads with.
The honest summary first. The full row-by-row tape is one click below — nothing hidden, just ordered so the picture is clear before the detail.
Read the stages, not the sum. These rows span paper, testnet and live — different sizing, some not real money — so adding them up as one P&L is meaningless. This window: 7 wins, 13 losses. Improved but still losing. Every stop — all 13 — fired at a pre-set level, none compounded. A single Sep 3 session drove the worst cluster: 9 stops across Smart-Money I & II, Higher-TF Momentum, and Trend (Alligator) as markets moved against SHORT entries. Money-Flow Long (paper) closed XAG +2.09% and XAUT +1.00% via take-profit. Smart-Money I closed INJ +0.82R via trailing stop on Sep 6. The live Liquidation-Zone (n=16) made no new entries — XRP +9.9% remains its most recent real-capital close. Stop discipline is the one lever the desk controls outright; this window shows it holding.
| Time (UTC) | Strategy | Stage | Coin | Side | Result | Exit |
|---|---|---|---|---|---|---|
| 2026-09-06 09:15 | Trend (Alligator) | Paper | BIO | LONG | −1.70% | Stop loss |
| 2026-09-06 00:00 | Smart-Money I | Testnet | INJ | SHORT | +0.82R | Trailing stop |
| 2026-09-05 17:00 | Smart-Money II | Testnet | RUNE | SHORT | −1.42R | Stop loss |
| 2026-09-05 08:00 | Smart-Money II | Testnet | SUI | SHORT | −1.25R | Stop loss |
| 2026-09-05 06:00 | Smart-Money I | Testnet | TIA | SHORT | −1.14R | Stop loss |
| 2026-09-05 06:00 | Smart-Money II | Testnet | TIA | SHORT | −1.22R | Stop loss |
| 2026-09-03 15:00 | Smart-Money II | Testnet | SOL | SHORT | −1.13R | Stop loss |
| 2026-09-03 14:50 | Money-Flow Long | Paper | XAG | LONG | +2.09% | Take profit |
| 2026-09-03 14:00 | Smart-Money I | Testnet | SOL | SHORT | −1.28R | Stop loss |
| 2026-09-03 13:56 | Higher-TF Momentum | Paper | INJ | SHORT | −1.70% | Stop loss |
| 2026-09-03 13:00 | Smart-Money I | Testnet | LINK | SHORT | −1.16R | Stop loss |
| 2026-09-03 13:00 | Smart-Money I | Testnet | RUNE | SHORT | −1.10R | Stop loss |
| 2026-09-03 13:00 | Smart-Money II | Testnet | LINK | SHORT | −1.13R | Stop loss |
| 2026-09-03 08:00 | Smart-Money II | Testnet | ATOM | SHORT | −1.15R | Stop loss |
| 2026-09-03 03:40 | Money-Flow Long | Paper | XAUT | LONG | +1.00% | Take profit |
| 2026-08-31 02:00 | Smart-Money II | Testnet | NEAR | LONG | +0.91R | Trailing stop |
| 2026-08-31 00:00 | Smart-Money I | Testnet | NEAR | LONG | −1.14R | Stop loss |
| 2026-08-28 16:00 | Smart-Money I | Testnet | NEAR | LONG | +0.81R | Trailing stop |
| 2026-08-27 17:00 | Smart-Money I | Testnet | SUI | SHORT | +0.87R | Trailing stop |
| 2026-08-26 19:00 | Smart-Money I | Testnet | ETH | SHORT | +0.74R | Trailing stop |
R = a trade's result measured against the risk taken on it (used for testnet rows) · % = the price move captured. Snapshot 2026-09-07.
Each strategy ships with pre-committed kill criteria. When they fire, it's discipline, not failure — and it's logged. This list only grows.
After a 66-variant research sweep: an ETH funding-fade (3-yr backtest PF 1.96, 92% bootstrap-positive) and an alt EMA-cross basket. Both go to paper first for forward validation — never straight to real money.
Deep-dive found it never earned a real-money slot: paper profit factor ~0.66, and even its best backtest regime barely broke even net of costs. Stopped and disabled. Reversible only if a genuine new edge thesis appears.
The first three live trades of a revised exit model all stopped out. Sample too small to act on — logged transparently, validation continues.
Data showed one strategy's market-regime gate was behind a string of stop-outs (it bought weakness while strength led). Removed on evidence, kept the filter that worked.
A direction filter had quietly stopped the Higher-TF Momentum strategy trading for six weeks. Diagnosed, A/B-tested, reverted — back to trading within 24 hours.
Transparency only counts if you can check it yourself.
The copy book runs on Binance, whose own lead-trader page reports ROI, PnL and win rate independently of us. Talk to Us and you get the live profile to inspect before you copy a cent.
As the Hyperliquid book comes online, every position and PnL is on-chain and public — verifiable by anyone, impossible to fake or quietly edit. Coming soon.
The kill log above only grows; entries are never deleted. The track record — wins and losses — is the asset, not any single call.
Write-ups from inside the desk — methodology, post-mortems, and what the data actually says.
The money flow index (MFI) measures buying and selling pressure using price and volume together. What it shows, what it misses, and why the 30-level matters.
Read →
Leverage amplifies every dollar — gains and losses alike. The liquidation maths, when leverage works, and when it quietly destroys accounts.
Read →
Trending vs ranging, what regime filters do, and the lesson from a filter we removed because the live data said to.
Read →
A trailing stop moves with the trade as it goes in your favour, locking in profit as momentum extends. How it works, what it costs you, and when to use it.
Read →
The same strategy with different position sizes produces completely different outcomes. The maths behind stake control — and the single rule that caps your downside.
Read →
Pre-committing to a stop removes the emotion. Why we've killed 5 strategies — and what to ask any service you follow.
Read →
We built a system to copy profitable traders — then killed it. A profitable trader isn't a copyable edge. The study, and what we do instead.
Read →
Gross wins ÷ gross losses. The single ratio that tells you whether a strategy has a real edge — and how to read it before you copy anyone.
Read →
Three halvings, a repeating chart pattern — and why three data points can't carry the precision most cycle forecasts require.
Read →
Spectacular backtest, immediate live losses. What overfitting is, how to spot it, and why the backtest-vs-live gap is the number that matters.
Read →
Plain-English guide: your funds never leave your own account. Why custody is the only catastrophic risk.
Read →
The most predictive number on any trader's profile — and the most ignored. The ratios that beat ROI alone.
Read →
The regulatory picture (VARA, ADGM), custody choices, and disciplined exposure without day-trading.
Read →
An honest comparison — effort, risk, cost, and who each approach actually suits.
Read →
Lending, staking, market-neutral strategies — and the risk behind each headline yield.
Read →
The 7-point checklist — custody, drawdown, fees, visible losses — from inside a desk, not a marketing department.
Read →
Most services show you wins. We publish the losses and the killed strategies — here's why that's the whole point.
Read →
It passed its backtest. We killed it anyway. Thin edges, trading costs, and pre-committed kill criteria.
Read →
What ~5,000 simulated trades taught us — why backtests flatter, and how we gate capital because of it.
Read →
No paywall. Leave your email and I'll send strategy write-ups, track-record updates and post-mortems as they're published.
Aligned with you, by design.
You keep custody of your funds in your own Binance account — Glasshouse never holds or can withdraw your money.
When the desk makes you money, it takes a performance fee — between 10% and 30% of profits, depending on your plan — on a high-water mark, settled by the exchange. It only earns on new gains, never the same ground twice. No profit, no fee; never a charge on losses, and the rest is yours.