FMZ均幅框架：ATOM Hyperliquid 4小时双向移植

Hypothesis

忠实移植善2019-07-19《均幅指标策略框架》的My语言代码，N=14、M=6为基线。选择ATOMUSD.HYPERLIQUID 4小时为市场与周期适配，原文未给出周期。原LOTS公式约对应10倍权益名义金额，为符合信号退出风险路线(b)，明确改为固定50%权益仓位、leverage=1；不声称复现原仓位风险。保留收盘退出，不加盘中止损、止盈、趋势过滤或持仓时限。完成K线收盘提交市场单，按连续市场的平台收盘成交约定执行。已查看三图：封面没有可识别标的/周期/交易标记；公式图确认ATR为TR的简单平均，MATR未被代码使用；回测图为国内商品期货Futures_CTP_FUTURES_CNY、MA888日志，未标明K线周期，其收益与回撤未经验证。独立策略链接128136在来源清单中HTTP404，未读取其配置。

按明确种子规则的可测试性接纳，不运行预筛回测。Hyperliquid占11.3%低于20%目标，双向设计缓解长仓78.1%的偏斜。ATOM交叉柱固定ATR锚定不同于所列AVAX每日Turtle、LINK/BCH星期开盘区间及ETH成交量EMA策略。corpus_concentration为空，无法声称ATOM未测试或其周期占比低于15%；避开拥挤BTC。保持单标的来源规则，不为跨市场或期权配额凭空添加腿。

negative_expectancy: This ATOM Hyperliquid 4H transplant at fixed 50% equity sizing loses money across 313 closed positions over 950 days. Average net trade return is -0.564% against the +0.15% review floor, PF is 0.746 and total return is -62.526%. QA’s economic concern is confirmed on full history, including severe modeled execution-cost drag. Optimization is not justified by this measured negative expectancy and the absence of a specific supported improvement path. This rejects the authorized market/timeframe/sizing transplant, not the source author’s strategy on its original commodity market.

Implementation

Ports the FMZ EMA(close,n)-EMA(open,n) zero-crossing framework to ATOM Hyperliquid 4-hour bars. Crossing-bar high/low plus or minus 0.1*m*SMA(TR,n) establish fixed entry and close-exit levels. Holds one long or short position at 50% of settled equity, without pyramiding, targets, intrabar stops or time limits. Market orders fill at the completed signal bar's close under the continuous-market platform convention.

Implements the source at https://www.fmz.com/digest-topic/4053 with n=14 and m=6. EMA series seed from their first completed prices; ATR is a simple rolling mean, and VALUEWHEN levels persist as scalar state until the next corresponding crossing. Long exits on a downward zero crossing or close at/below PRICE_SP; short exits on an upward crossing or close at/above PRICE_BP. Only n and m are tunable under the research plan. The original approximately 10x LOTS sizing is explicitly replaced by fixed 50% equity exposure and leverage=1, satisfying signal-exit risk route (b). No claim of profitability or replication of original position risk is made.

Verification Results

The smoke test reports -0.524% net return per closed position, PF 0.760, and 69.14% drawdown. Modeled impact totals $81,680.93 and reported capacity is $6,959.41; full-history review should assess execution costs at the intended capital allocation. Hyperliquid taker commission is approximately 0.09% round trip. These are economic concerns, not demonstrated implementation defects.

Backtest Review

Sharpe

-1.07

Total return

-62.53%

Max drawdown

70.55%

Trades

313

Win rate

27.2%

Profit factor

0.75

313 closed positions, split 160 long and 153 short, provide a usable review sample. Inspected trades show the authorized capital-relative 50% sizing.

Accounting reconciles and no terminal unrealized gain remains.

QA’s concern is borne out: average net trade return -0.564% versus the +0.15% Hyperliquid review floor, PF 0.746, Sharpe -1.071 and total return -62.526%.

Modeled impact totals $79,220, or 474.5% of gross PnL; reported capacity is $4,441 versus the $100,000 tested account.

Drawdown is 70.548% versus the contracted 20% limit. Buy-and-hold returned -85.164%, with approximately 91.189% drawdown on the stored benchmark curve.

No verified implementation defect or demonstrated viable tuning region supports a specific iteration.

Outcome Summary

This market, timeframe and sizing adaptation needs evidence of viable net expectancy and execution capacity before parameter optimization is justified.

Backtest review voted to abandon because expectancy was negative, average trade return missed the +0.15% review floor and drawdown exceeded the 20% limit, with no verified implementation defect or supported improvement path. Optimization and subsequent analyst and risk reviews were not reached.

FmzAtomMeanAmplitude adapted the FMZ mean-amplitude framework to ATOM Hyperliquid 4-hour long/short trading using EMA crossovers, ATR-based thresholds, bar-close exits and fixed 50% equity sizing at 1x leverage.

Across 950 days and 313 closed positions, the backtest returned -62.526%, with Sharpe -1.071, profit factor 0.746 and maximum drawdown 70.548%. Average net trade return was -0.564%, while modeled impact totaled $79,220.

Analysis

均幅指标策略框架