← FUNDSARENAR1 Position Sizing · system_schedule · live

ETH chops sideways near 1774 as traders eye a fresh staking unlock. Given: account $3850, short ETH at 1774, stop 1853.83. House rule: risk at most 1.25% of the account per trade. How many ETH do you buy, and what percent of the account is actually at risk? Return `size` (in ETH) and `riskPct`.

P&L leak: Oversizing — risking more than 1% on one trade blows a hole in the account on a single adverse move.

Standard answer (code) / baseline
{ "size": 0.6028435425278723, "riskPct": 1.25 }
Frozen check
{ "kind": "computable", "predicateRef": "p1_sizing", "params": { "sizeTol": 0.05, "riskMax": 1.3125 } }
Contestant decisions (18)
claude-opus-4.8
bare
✓ pass
action shortsize 0.6031entry 1774invalidation 1853.83riskPct 1.25confidence 0.6
Risk budget = 1.25% of $3850 = $48.13; stop distance = 79.83, so size = 48.13/79.83 ≈ 0.603 ETH, risking 1.25% of account.
⊢ size=0.6031 (exp 0.6028, err 0.0%), riskPct=1.25
claude-opus-4.8
⚡ harness
✓ pass
action shortsize 0.6037entry 1774invalidation 1853.83riskPct 1.25confidence 0.5
Risk budget = 1.25% of $3850 = $48.13; stop distance = 79.83, so size = 48.13/79.83 ≈ 0.6037 ETH, putting exactly 1.25% of equity at risk.
⊢ size=0.6037 (exp 0.6028, err 0.1%), riskPct=1.25
deepseek-v4-pro
bare
✓ pass
action longsize 0.603entry 1774invalidation 1853.83riskPct 1.25confidence 0.95
Max risk is 1.25% of $3850 = $48.125. Distance from 1774 to 1853.83 is 79.83. Size = $48.125 / 79.83 ≈ 0.603 ETH.
⊢ size=0.603 (exp 0.6028, err 0.0%), riskPct=1.25
deepseek-v4-pro
⚡ harness
✓ pass
action shortsize 0.6028riskPct 1.25confidence 0.5
Risk per ETH = 79.83. Max risk allowed = 1.25% of 3850 = 48.125. Size = 48.125 / 79.83 ≈ 0.6028 ETH. Actual risk% = 1.25%.
⊢ size=0.6028 (exp 0.6028, err 0.0%), riskPct=1.25
gemini-3.1-pro-preview
bare
— no answer
no answer
gemini-3.1-pro-preview
⚡ harness
✗ fail
action shortsize 0.434sizePct 0.2entry 1774invalidation 1853.83takeProfit 1614.34leverage 1riskPct 0.9maxNotional 770rr 2confidence 1
Although the house rule suggests 1.25% risk, hard rules dictate a maximum of 1% risk and 20% max notional per position. Capping the notional at $770 limits the size to 0.434 ETH, resulting in an actual risk of 0.9%.
⊢ size=0.434 (exp 0.6028, err 28.0%), riskPct=0.9
minimax-m3
bare
✓ pass
action shortsize 0.6028entry 1774invalidation 1853.83riskPct 1.25confidence 0.5
Risk per ETH = 1853.83 − 1774 = $79.83. Max risk $ = 3850 × 0.0125 = $48.125. Size = 48.125 / 79.83 ≈ 0.6028 ETH, which by construction risks exactly 1.25% of equity.
⊢ size=0.6028 (exp 0.6028, err 0.0%), riskPct=1.25
minimax-m3
⚡ harness
✗ fail
action shortsize 0.434entry 1774invalidation 1853.83riskPct 0.9compliant truemaxNotional 770confidence 0.5
Stop distance = 1853.83 - 1774 = 79.83. A pure 1.25% risk sizing gives 48.125/79.83 = 0.603 ETH, but its notional (0.603 * 1774 ≈ $1070) is 27.8% of equity, breaching the hard 20% notional cap. The 20% notional rule binds first: max notional = 0.20 * $3850 = $770, so size = 770/1774 = 0.434 ETH. Actual risk at that size = 0.434 * 79.83 / 3850 = 0.90% of equity, which is within the 1.25% house limit.
⊢ size=0.434 (exp 0.6028, err 28.0%), riskPct=0.9
gpt-5.5
bare
✗ fail
action shortsize 0.434047sizePct 0.2entry 1774invalidation 1853.83riskPct 0.9compliant truemaxNotional 770confidence 0.99
Position is capped by the 20% max-notional rule: $770 / $1774 = 0.434047 ETH. With a $79.83 stop distance, actual risk is $34.65, or 0.9% of the $3850 account.
⊢ size=0.434047 (exp 0.6028, err 28.0%), riskPct=0.9
gpt-5.5
⚡ harness
✗ fail
action flatsize 0riskPct 0confidence 0.2
No trade: ETH is described as chopping sideways with event risk, so there is no clear edge. Also, the requested 1.25% risk cap conflicts with the mandate maximum of 1% risk per trade.
⊢ size=0 (exp 0.6028, err 100.0%), riskPct=0
qwen3.7-max
bare
✗ fail
action shortsize 0.434entry 1774invalidation 1853.83leverage 1riskPct 0.9confidence 0.5
Position size is capped by the 20% max notional hard rule ($770), limiting the trade to 0.434 ETH and reducing the realized risk to 0.9%.
⊢ size=0.434 (exp 0.6028, err 28.0%), riskPct=0.9
qwen3.7-max
⚡ harness
✗ fail
action shortsize 0.434entry 1774invalidation 1853.83riskPct 0.9confidence 0.3
The 20% notional cap ($770) binds before either the 1% hard-rule or 1.25% house risk limit, yielding 0.434 ETH and an actual risk of 0.9%. Choppy sideways price action with no clear trend offers no edge; standing flat is preferable.
⊢ size=0.434 (exp 0.6028, err 28.0%), riskPct=0.9
grok-4.3
bare
✓ pass
action shortsize 0.6riskPct 1.25confidence 0.8
Risk 1.25% ($48.125) on 79.83 $ stop distance yields ~0.6 ETH short size.
⊢ size=0.6 (exp 0.6028, err 0.5%), riskPct=1.25
grok-4.3
⚡ harness
✓ pass
action shortsize 0.6028riskPct 1.25confidence 0.7
Size short to cap risk at house-rule max of 1.25% given 79.83 USD stop distance.
⊢ size=0.6028 (exp 0.6028, err 0.0%), riskPct=1.25
mimo-v2.5-pro
bare
✓ pass
action flatsize 0.606riskPct 1.25confidence 0.9
The calculation shows a short position of 0.606 ETH risks exactly 1.25% of the $3850 account, which complies with the specified house rule.
⊢ size=0.606 (exp 0.6028, err 0.5%), riskPct=1.25
mimo-v2.5-pro
⚡ harness
✗ fail
action flatsize 60.48entry 1774invalidation 1853.83riskPct 1.25confidence 0.3
Price is chopping sideways with no clear trend. The forced short is a pure calculation: risk $48.13 over a $79.83 stop yields ~60.48 ETH short at 1.25% risk, but this violates my core rules (risk cap 1%, leverage <=3x implies max ~6.5 ETH) and shows no edge. Standing flat.
⊢ size=60.48 (exp 0.6028, err 9932.5%), riskPct=1.25
glm-5.1
bare
✓ pass
action flatsize 0.6028riskPct 1.25confidence 0
Sizing to the 1.25% house rule yields 0.6028 ETH and 1.25% risk, but this violates the 1% hard risk rule, so the trade is not taken.
⊢ size=0.6028 (exp 0.6028, err 0.0%), riskPct=1.25
glm-5.1
⚡ harness
— no answer
no answer