RESEARCH · DATA TO 6 OCTOBER 2026
What the data says about Guard’s rules
Four studies on real Hyperliquid data. Each lists its method, sample, trials and limits, and distinguishes pre-registered measurements from exploratory subgroups. Aggregates only: no account is named or shown.
Three findings to start with
Defaults held up out of sample
None beat the defaults out of sample: the best challenger scored +0.023 (95% interval −0.121 to +0.227). Recommendation: keep the current defaults.
Sample and trials
262 Hyperliquid accounts (train 136, validation 63, held-out test 63); each account’s last 180 days to 6 October 2026.
5,946 rule-set variants (5,940 grid, 6 ablation); the deviations in the method count as trials too.
Liquidations and rule violations
91% of liquidated positions against 38% of the others; relative risk 15.7 (95% interval 15.0–16.4). An association, not a claim that Guard prevents liquidations.
Sample and trials
All of Hyperliquid over 30 days, 4 Sep – 3 October 2026: 95,095 liquidated positions and 7,942,674 position-days that were not; 62% of liquidations covered.
18 comparisons reported, listed in the method.
Bot results remain exploratory
22 likely bots, 151 likely humans, 150 unclear. For likely bots nothing robust can be said yet; for likely humans the protection results hold.
Sample and trials
323 accounts from the three finished studies.
94 measurements, each also an extra trial of its parent study.
The principal trade-off
It costs the best traders part of their gains: in the replay the median long-term winner’s return fell from 43% to 8%, mostly through the drawdown halt (with the halt opened alone, 13%). 9 winners, training and validation accounts.
Only about 10% of entries go through at full size (95% interval 3–18%). Of all judged entries, 43% were held by a halt (39% by the drawdown halt, once an account was already 25% down), 21% refused and 13% resized.
See the cost alongside the protection →These are historical measurements. The full findings below retain their intervals, methods and limits; they do not promise future results.
| Study | Data | Trials | Design |
|---|---|---|---|
| Guard defaults study | 6 October 2026 | 5,946 | Pre-registered Train, validation and a held-out test |
| How traders compare with Guard’s defaults | 6 October 2026 | 24 | Pre-registered Two cohorts, replayed |
| Liquidation autopsy | 4 Sep – 3 October 2026 | 18 | Pre-registered A count over the whole market |
| Bots and humans in the studies (exploratory) | 6 October 2026 | 94 | Exploratory Post-hoc subgroups of the finished studies |
Trials: every variant or measurement the study took, as its method lists them. Each study’s protocol summary is at the top of its method page.
Guard defaults study · 6 October 2026
Defaults against alternatives
We tested 5,940 alternative rule sets on the real trades of 262 Hyperliquid accounts. None beat our defaults on accounts they weren’t tuned on.
A replay of each account’s last 180 days to 6 October 2026 through Guard’s engine. “Beat” means our pre-registered score, with its constraints, on accounts the rule set was not chosen on; the best challenger scored +0.023 (95% interval −0.12 to +0.23). So we kept them.
How to read this
The dot is the best challenger’s score against the defaults, the bar its 95% interval. The defaults sit at 0. An interval that crosses 0 gives no evidence that the challenger is better.
Pre-registered score, best challenger minus the defaults. Right of 0 favours the challenger.
Guard defaults study · held-out test, 63 accounts
The held-out test
Held-out test: 63 accounts looked at once, after the rules were fixed. Medians; 95% bootstrap intervals. A mixed sample: ordinary active traders, long-term winners and recently liquidated accounts.
How to read this
Left: each measure with its 95% bootstrap interval; the drawdown cut is in percentage points, the rest are shares. Right: the median account’s worst day as traded and behind the defaults; switch a series off with its key. Hover, focus or tap a row for the exact values.
- −42 ptsmax drawdown, median accountcut by 42 points (95% interval 22–63)63 accounts held out
- −60% → −18%worst day, median accountas traded, then behind the defaults; no interval computed for this difference63 accounts held out
- 60%return per unit of drawdown improvedshare of accounts (95% interval 49–71%)63 accounts held out
- 68%liquidation events with the guarded account flatin 136 of 201 liquidation events, the guarded account held no position in that coin (95% interval 26–90%). Not “prevented”: it is mostly the drawdown halt, and part of the sample was picked for having been liquidated.63 accounts held out
Medians and shares, 63 accounts held out; 95% bootstrap intervals.
Compliance study · data as of 6 October 2026
Winners and liquidated accounts
Long-term winners mostly sized their trades within Guard’s limits; liquidated accounts mostly did not.
- 83%
- long-term winners · median share of entries within Guard’s limits (95% interval 19–100%) · 15 accounts
- 0%
- liquidated accounts · the same share (95% interval 0–4%) · 57 accounts
How to read this
Each bar counts the accounts whose share of entries within Guard’s limits falls in that 10-point band. The cohorts differ in size (15 and 57): switch to “Share of cohort” to compare their shapes. Under each cohort, the dot is its median and the bar the median’s 95% bootstrap interval.
Limits of this result
“Within” counts Guard’s own 2% stop: most winners used no resting stop. Winners are survivors of today’s leaderboard, with much larger accounts. Last 180 days to 6 October 2026; descriptive, not a promise.
Long-term winners n = 15
median 83% · 95% interval 19%–100%
Liquidated accounts n = 57
median 0% · 95% interval 0%–4%
Share of each account’s entries that fit Guard’s limits with Guard’s 2% stop, in 10-point bands; the median and its 95% interval below each cohort.
Liquidation autopsy · 4 Sep – 3 October 2026
Before liquidation
91% of liquidated positions broke Guard’s leverage or liquidation-buffer default that morning, against 38% of the others.
- 91%
- liquidated positions over the limits · leverage above 5× or liquidation within 10% of entry (95% interval 90–91%) · 95,095 positions
- 38%
- positions not liquidated that day · the same rules, the base rate (95% interval 37–38%) · 7,942,674 position-days
How to read this
The share of positions over Guard’s leverage or liquidation-buffer default at that morning’s snapshot, for positions liquidated later that day and for all others. The samples are large, so the 95% intervals are narrower than the dots.
Limits of this result
Over the 30 days 4 Sep – 3 October 2026, all of Hyperliquid. Read from that morning’s snapshot of open positions, not at entry; covers the 62% of liquidations whose position was open at the snapshot. An association, not a claim that Guard prevents liquidations.
Share of positions over the limits at the morning snapshot, with 95% intervals.
Guard defaults study · the cost
The cost of protection
It costs the best traders part of their gains: in the replay the median long-term winner’s return fell from 43% to 8%, mostly through the drawdown halt (with the halt opened alone, 13%). 9 winners, training and validation accounts.
Only about 10% of entries go through at full size (95% interval 3–18%). Of all judged entries, 43% were held by a halt (39% by the drawdown halt, once an account was already 25% down), 21% refused and 13% resized.
The “active” preset lets more trades through (paper and testnet only for now).
How to read this
Left: the median long-term winner’s return as traded, behind the defaults, and behind the defaults with only the drawdown halt opened. Right: what Guard did with the entries it judged; “by the drawdown halt” is part of “held by a halt”.
Bot split · exploratory
Bots and humans (exploratory)
Of the 323 accounts the studies measured, 22 look like bots, 151 like people using a web interface, 150 are unclear. For likely bots nothing robust can be said yet; for likely humans the protection results hold.
Limits of this result
Exploratory: post-hoc subgroups of the three finished studies, from a public-data heuristic with no ground truth. Every number counts as an extra trial of its parent study. A dedicated bot study is running.
- Likely bots227%
- Likely humans15147%
- Unclear15046%
Guard defaults study · a second preset
The “active” preset (paper and testnet only for now)
For traders who find the defaults too restrictive: more of your trades go through unchanged (about twice as many in our replay), at the cost of somewhat deeper drawdowns. Liquidation protection stays about the same.
- On the held-out test: 24% of entries at full size (95% interval 11–41%) against 10% for the defaults; liquidation events with the guarded account flat 138 against 136 of 201; the worst decile’s drawdown 49% against 45%.
- Measured with a 1.5% attached stop, a result the study could not confirm (the replay misses more stop-outs at 1.5%); offered here with the defaults’ 2% stop.
- Not the default: it breaks the study’s pre-registered worst-case constraint. Its limits are looser than Guard’s mainnet ceiling, so it is for paper and testnet until we decide otherwise.
Entries through at full size, held-out test, with 95% intervals.
| Rule | Defaults | Active |
|---|---|---|
| Loss at the stop | 2% | 3% |
| Position size | 200% | 500% |
| Open risk | 6% | 15% |
| Daily loss stop | 6% | 4% |
| Drawdown halt | 25% | 35% |
| Leverage, attached stop, liquidation buffer | as the defaults | as the defaults |
What these numbers can and cannot carry
- What-if replays: each account’s own trades over the 180 days to 6 October 2026, judged by Guard’s engine; later decisions are kept as traded.
- The replay sees Guard’s attached stops only at the account’s own fills. One-second prices show that 16% of the default stops the replay never fired were touched in between: real Guard would have closed those positions at a loss and missed any recovery. This flatters Guard somewhat.
- Main-dex perps only: HIP-3, spot and outcome markets are left out, as Guard does not cover them yet.
- The samples are mostly manual traders: by a public-data heuristic (exploratory), 22 of the 323 accounts measured (7%) look like bots. A dedicated bot study is running.
- The liquidation autopsy covers 4 Sep – 3 October 2026, read at a daily snapshot.
- Past behaviour, not a forecast, and no promise of returns.
Each study’s own limits, deviations and trial count are on its method page: Guard defaults study · How traders compare with Guard’s defaults · Liquidation autopsy · Bots and humans in the studies (exploratory).