Where the price can go and where leveraged positions sit.
What a position can lose and how close it is to liquidation.
Real execution costs and whether a backtest result survives statistics.
At which prices are leveraged longs and shorts likely to be liquidated?
Held-out days 13–19 October, 30 perpetual contracts. Price crossing version 2.0 levels explains open-interest closes 3.4 times better than the previous version, with a positive effect on all 30 instruments. About 9% of a level's estimate closes when price crosses it (90% interval 4–13%). Reviewed every 14 days.
What range will the price cover, and how likely is it to touch a given level?
Horizons longer than one day run engine 2.0: a daily volatility level (HAR with the last 4h and 1h), 5m steps for the first day, then 1h and 4h steps. Out of sample on 30 perpetuals, January 2025 to October 2026, with the 27 days of history ArcDelta holds: CRPS 0.9% better than 1.1 at 7 days and 0.4% at 3 days, the 7-day upper 1% tail breached 1.4% of the time (1.1: 1.9%), and the run is about four times faster. Up to one day and for indicator distributions engine 1.1 runs. Reviewed every 14 days.
How much will this instrument move over the next hour, four hours and day, and are options pricing more or less than that?
On a held-out week across 28 perpetuals, HAR had a QLIKE loss 4.7% lower than GARCH at 1h, 3.3% at 4h and 3.7% at 24h; it won on 20, 17 and 13 of the 28 instruments.
How tied is each coin to BTC, in calm and in stress, and which coins move together once BTC is taken out?
On 134 perpetuals over 8 days the 134 × 134 matrix took under 0.2 s for every window. Between consecutive days the ranking of coins by correlation to BTC held (Spearman 0.77), while clusters agreed only partly (adjusted Rand 0.26).
Which hours of the day and days of the week are really different for this instrument?
Over 20 days on 28 perpetuals, 376 of 3,276 tests passed the false-discovery control at q = 0.10: higher turnover and volatility at the US open (13:00–15:00 UTC) on 26 and 25 instruments, quieter weekends on about half. Return effects passed 6 times in total and the up-hour share never.
How much of the current trading is caused by earlier trading, and is the flow heating up?
On simulated flow with known parameters the fit recovers the branching ratio within 0.15, the residuals pass the KS test and a plain Poisson fails it. A live check on real trades appears on this page for every run.
What would a simple two-sided quoting strategy earn or lose over the next minutes if the flow keeps its Hawkes clustering, calm and stressed?
Simulation only. The PnL split is exact (the three parts sum to the total) and is covered by tests; the price model is not validated against real order books.
What will holding this perpetual cost or pay in funding over the next day, on each venue?
0.6 × last settled rate + 0.4 × 3-day mean predicted the next-day average rate with an MAE of 0.51 bp, against 0.59 bp for the last rate and 0.78 bp for zero, on 488 venue symbols.
How much can this position lose over the next hour, day or week?
Same engine as Monte Carlo 1.1. Out of sample on 30 contracts from 1 September to 18 October 2026 the one-hour 5% loss tail was breached 4.4% of the time and the 1% tail 1.1%; at four hours 5.1% and 1.1%.
How likely is my leveraged position to be liquidated, and what leverage keeps that risk low?
Touch probabilities include intrabar highs and lows. On 30 perpetuals the predicted probability differed from the observed touch frequency by 1.8 points on average at 1h, 1.0 at 4h and 2.5 at 24h (engine 1.0 check; 1.1 bounds the variance reaction to single outlier bars).
How likely is my take to fill before my stop, and what is the trade worth in R?
On 28 perpetuals, about 27,000 setups per horizon, the predicted take-first probability differed from the observed frequency by 2.4 points on average at 1h, 1.3 at 4h and 3.3 at 24h.
What will a market order of my size really cost on each venue?
The trailing two-hour median predicted the cost one hour later with a median error of 6.6% ($10k) to 9.3% ($500k). Version 1.1 band: on the held-out day, 23 October, 79.8% of future costs fell inside the p10–p90 band against an 80% target (the old band 65.7%); BTC and ETH are still under-covered.
Is this backtest Sharpe real, or the best of many lucky tries?
In 1,000 null simulations the best of 50 variants passed plain PSR 92% of the time and DSR 0%; a real edge among 20 variants was found 37% of the time.
Model cards, versions and periodic review results are published in Research. If a model stops holding its calibration, it shows there first.
Today's term structure inside realised-vol percentiles.
Close-close, Parkinson, Garman-Klass and Yang-Zhang side by side.
Excursions beyond ±2σ flagged as they happen.
Where price sits inside its Bollinger bands.
Rescaled range slope: trending or mean-reverting.
VR(q) above 1 trends, below 1 reverts.
OU half-life: how fast a shock fades.
ACF by lag with the 95% significance band.
Skew and fat tails, read out live.
A smooth density built from individual samples.
Principal axes of a correlated return cloud.
Past windows that look most like today.
Calm, trend and stress regimes marked on the series.
Equity, trades and drawdowns under the high-water mark.
Each fold tests on data it never saw.
Every parameter cell scored; the best one kept.
In-sample vs out-of-sample, with a verdict.
How a growing order walks up the book.
Gross vs net equity: what fees cost over time.
How deep and how long below the last peak.
Tail loss at 95% and 99%, and what lies beyond.
A shock hits the book: losses, hedges, recovery.
The Kelly peak and the cost of over-betting.
Bootstrapped Sharpe: a range, not a point.
Block resampling builds alternative histories.
Each liquidation tips the next, faster and faster.
Square-root impact against a linear model.
Share of simulated paths that hit the ruin line.
Assets linked by how strongly they move together.
How predictable the tape is right now.
Each tree fits what the last ones missed.
Which inputs actually carry the model.
Overlapping labels removed around every test fold.
Every fold scored out of sample.
Take profit, stop or time — whichever comes first.
A second model decides which signals to trade and how big.
Two prices, one stationary spread.
A hidden state tracked through noisy prints.
Big moves come in clusters.
Hidden Markov states: calm, trend, stress.
Trends that persist and stretches that snap back.
How fast predictive power fades with horizon.
Rank correlation of forecast and outcome.
Trading more until costs eat the edge.
Best return for each level of risk.
Weights set so each asset adds equal risk.
Spread plus square-root impact by size.
Hits and misses as the threshold moves.
Does 70% mean 70%?
True positives against false alarms.
Sharpe adjusted for how many things were tried.
How often the in-sample winner loses out of sample.
Market sensitivity that changes over time.
Where returns really come from.
Assets that crash together.
A Pareto fit beyond the threshold.
Discontinuities separated from diffusion.
Activity by hour and weekday.
Data, metrics and limits on one page.
Drift between paper and reality, tracked.
Several models, one decision.
Each forecast traced split by split.