[{"data":1,"prerenderedAt":1143},["ShallowReactive",2],{"\u002Fzh\u002Fglossary\u002Fwhat-is-ahpr-and-ghpr":3,"\u002Fzh\u002Fglossary\u002Fwhat-is-ahpr-and-ghpr-related":234,"\u002Fzh\u002Fglossary\u002Fwhat-is-ahpr-and-ghpr-langs":1140},{"id":4,"title":5,"body":6,"category":216,"date":217,"description":218,"draft":219,"extension":220,"meta":221,"navigation":222,"path":223,"related":224,"seo":229,"stem":230,"term":231,"updated":232,"__hash__":233},"glossary\u002Fzh\u002F5.glossary\u002Fwhat-is-ahpr-and-ghpr.md","什么是 AHPR 和 GHPR？公式与两者的差距",{"type":7,"value":8,"toc":208},"minimark",[9,13,17,20,31,40,48,51,57,68,71,74,87,95,98,114,134,138,157,160,163,200],[10,11,12],"p",{},"AHPR（平均持仓期收益率）和 GHPR（几何持仓期收益率）都是单笔交易的平均收益，以该笔交易之前的资金为基数，用百分比表示：AHPR 取算术平均，GHPR 取几何平均。两者衡量的是同一批交易，回答的却是不同的问题；两者之间的差距，反映了结果的波动有多大。",[14,15,16],"h2",{"id":16},"工作原理",[10,18,19],{},"每笔已平仓交易都有一个持仓期收益率，即该笔净结果除以交易前的资金：",[21,22,27],"pre",{"className":23,"code":25,"language":26},[24],"language-text","HPRᵢ = (盈利 + 库存费 + 佣金)ᵢ \u002F 第 i 笔交易之前的资金\n\nAHPR = ( Σ HPRᵢ \u002F N ) × 100\nGHPR = ( [ Π (1 + HPRᵢ) ] ^ (1\u002FN) − 1 ) × 100\n","text",[28,29,25],"code",{"__ignoreMap":30},"",[10,32,33,34,39],{},"成本都在分子里：佣金和",[35,36,38],"a",{"href":37},"\u002Fglossary\u002Fwhat-is-a-swap-in-forex","库存费","先计入每笔交易，再算 HPR，所以两个数字都是净值。",[10,41,42,43,47],{},"基数有个细节要注意。我们的实现以账户的",[44,45,46],"strong",{},"历史累计入金总额","作为滚动资金的起点，再逐笔加上每笔交易的净结果。分期注资的账户，早期交易是拿还没到账的钱来衡量的，两个数字都会被拉向零；只入金一次的账户，基数就和公式字面上的意思完全一致。",[10,49,50],{},"GHPR 永远不会高于 AHPR。这是算术，不是交易本身的规律：正数的几何平均数最多等于算术平均数，只有所有数字都相同时两者才相等。差额大约是收益方差的一半：",[21,52,55],{"className":53,"code":54,"language":26},[24],"GHPR ≈ AHPR − σ² \u002F 2\n",[28,56,54],{"__ignoreMap":30},[10,58,59,60,63,64,67],{},"两笔交易就能说明问题。先 +50% 再 −50%，AHPR 正好是 ",[44,61,62],{},"0%","，账户却只剩起点的 0.75。GHPR 是 √(1.50 × 0.50) − 1 = ",[44,65,66],{},"−13.4%","，复利两次正好落在 0.75。算术平均描述的是谁都没拿到的结果，几何平均才还原了真实的余额。",[14,69,70],{"id":70},"为什么重要",[10,72,73],{},"算术平均数是花不出去的。只有几何平均值按交易笔数乘方之后，才能还原账户实际的表现，所以该引用的是 GHPR，AHPR 是拿来对照的标尺。两者差距大，就是所谓的波动率税：方差在复利落到余额之前先拿走的那一部分。",[10,75,76,77,81,82,86],{},"到了尾部，这种不对称会变得非常残酷。只要有一笔交易把资金全部亏光，乘积里就多了一个零，乘积也就是零：之前一千笔交易做得再好，GHPR 也显示 −100%。这就是",[35,78,80],{"href":79},"\u002Fglossary\u002Fwhat-is-position-sizing","仓位管理","比胜率更重要的道理，只是换成乘法来讲：一种方法即使 AHPR 不错，只要",[35,83,85],{"href":84},"\u002Fglossary\u002Fwhat-is-standard-deviation-in-trading","标准差","大到偶尔会押上全部资金，几何期望照样是负的。",[10,88,89,90,94],{},"AHPR 也不是摆设：它就是每笔交易的",[35,91,93],{"href":92},"\u002Fglossary\u002Fwhat-is-expectancy","期望值","，只是用资金的百分比来表示，而不是用点数或货币，所以能在不同规模的账户之间比较。两者放在一起看，能告诉您平均每笔交易赚了多少，以及扣掉方差的代价之后还剩多少。",[14,96,97],{"id":97},"数据显示了什么",[10,99,100,101,104,105,109,110,113],{},"以下数字只描述在 ShowMyTrades 上发布的账户，不代表全体交易者。中位数的已发布账户有 ",[44,102,103],{},"171 笔已平仓交易","，",[35,106,108],{"href":107},"\u002Fglossary\u002Fwhat-is-time-weighted-return","时间加权收益率","中位数为 ",[44,111,112],{},"+3.2%","。把这个收益平摊到这么多笔交易上，每笔的几何收益约为 +0.018%，面板四舍五入后显示 +0.02%。这只是示意，不是实测的中位数，但它如实反映了这些数字的量级，也说明了为什么它们在几乎所有真实账户上都显得不起眼。",[10,115,116,117,109,121,124,125,128,129,133],{},"同一批账户的",[35,118,120],{"href":119},"\u002Fglossary\u002Fwhat-is-a-profit-factor","盈利因子",[44,122,123],{},"1.28","（已扣除库存费和佣金），胜率中位数为 ",[44,126,127],{},"68.8%","。所以一个页面上",[35,130,132],{"href":131},"\u002Fguides\u002Fwin-rate-is-not-an-edge","胜率看起来很高","、GHPR 却几乎为零，两者并不矛盾。要看的是正负号、和 AHPR 的差距，以及背后有多少笔交易。",[14,135,137],{"id":136},"在-showmytrades-上的位置","在 ShowMyTrades 上的位置",[10,139,140,141,144,145,148,149,152,153,156],{},"两者都在每个已发布账户页面的",[44,142,143],{},"高级统计","模块里，标签就是 ",[44,146,147],{},"AHPR"," 和 ",[44,150,151],{},"GHPR","，位于",[44,154,155],{},"表现指标","一栏底部，排在盈利因子、标准差、夏普比率、Z 值（概率）和期望值之后。两者都显示为带正负号、保留两位小数的百分比，大于等于零为绿色，小于零为红色。",[10,158,159],{},"这个模块有四个视图：表格、星期、小时和持续时间。AHPR 和 GHPR 在表格视图里；另外三个视图能看出是哪些交易造成了拉开两者差距的方差。两个数值都由我们的统计服务在账户同步时算好，不是在您的浏览器里计算。",[14,161,162],{"id":162},"常见误解",[164,165,166,174,188,194],"ul",{},[167,168,169,170,173],"li",{},"“",[44,171,172],{},"AHPR 是平均值，所以它才老实。","”这是一个谁都赚不到的平均值。只要结果有波动，它就会高估复利增长，高出的幅度约为方差的一半。",[167,175,169,176,179,180,183,184,187],{},[44,177,178],{},"它们是月收益或年收益。","”它们是按笔计算的。同样的 +0.02%，放在 20 笔交易和 2,000 笔交易上意义完全不同，所以要和同一模块",[44,181,182],{},"交易","一栏的",[44,185,186],{},"总交易数","对照着看。",[167,189,169,190,193],{},[44,191,192],{},"GHPR 接近零，说明系统不行。","”在正常的单笔风险下，两个数字本来就很小，面板又只保留两位小数。一个账户在几百笔交易上每笔显示 +0.02%，累计涨幅已是两位数。",[167,195,169,196,199],{},[44,197,198],{},"AHPR 为正、GHPR 为负，是 bug。","”这正是大幅波动的特征：收益平均下来高于零，复利结果却依然向下。这也是这对指标能告诉您的最有用的信息。",[10,201,202,203,207],{},"单笔交易指标如何放进一份要给别人看的业绩记录里，请参阅",[35,204,206],{"href":205},"\u002Fguides\u002Ftracking-trading-performance","交易表现追踪","。",{"title":30,"searchDepth":209,"depth":209,"links":210},2,[211,212,213,214,215],{"id":16,"depth":209,"text":16},{"id":70,"depth":209,"text":70},{"id":97,"depth":209,"text":97},{"id":136,"depth":209,"text":137},{"id":162,"depth":209,"text":162},"指标","2026-08-19T00:00:00.000Z","AHPR 和 GHPR 是每笔交易的算术平均收益和几何平均收益。本文介绍公式、几何平均为什么总是更低，以及数千个账户的实测结果。",false,"md",{},true,"\u002Fzh\u002Fglossary\u002Fwhat-is-ahpr-and-ghpr",[225,226,227,228],"what-is-expectancy","what-is-standard-deviation-in-trading","what-is-position-sizing","what-is-a-z-score-in-trading",{"title":5,"description":218},"zh\u002F5.glossary\u002Fwhat-is-ahpr-and-ghpr","AHPR 与 GHPR","2026-09-17T00:00:00.000Z","-kX8ydWcwYp-iVY5E0IA7RC9EMTk0Vt5UJs8vlpjUJ4",[235,439,678,890],{"id":236,"title":237,"body":238,"category":216,"date":217,"description":428,"draft":219,"extension":220,"meta":429,"navigation":222,"path":430,"related":431,"seo":435,"stem":436,"term":437,"updated":232,"__hash__":438},"glossary\u002Fzh\u002F5.glossary\u002Fwhat-is-a-z-score-in-trading.md","交易中的 Z 值是什么？连胜连亏还是偶然",{"type":7,"value":239,"toc":421},[240,243,245,251,257,264,267,287,294,296,304,312,320,323,325,343,359,361,379,386,388,414],[10,241,242],{},"交易中的 Z 值，看的是盈利和亏损是扎堆出现还是交替出现：先数出实际序列里有多少个游程，再和同样数量的盈亏随机排列时应有的游程数比较。它不说明账户赚没赚钱，只说明上一笔的结果对下一笔有没有提示作用。",[14,244,16],{"id":16},[10,246,247,250],{},[44,248,249],{},"游程","指一段连续的同号结果。序列 W W L W L L 有四个游程。所用的统计量是 Wald–Wolfowitz 游程检验：",[21,252,255],{"className":253,"code":254,"language":26},[24],"R  = 2WL \u002F N + 1                    （期望游程数）\nσ² = 2WL(2WL − N) \u002F ( N² (N − 1) )  （方差）\nZ  = (游程数 − R) \u002F σ\n\nW = 盈利交易数，L = 亏损交易数，N = W + L\n",[28,256,254],{"__ignoreMap":30},[10,258,259,260,263],{},"交易按平仓时间排序，盈亏按",[44,261,262],{},"净","结果判定，佣金和库存费都已计入。一笔仓位毛利为正、扣掉成本后净亏，在这里就算亏损，并且会打断一段连胜的游程。",[10,265,266],{},"看符号就知道怎么读：",[164,268,269,275,281],{},[167,270,271,274],{},[44,272,273],{},"Z 为负","：游程比随机情况少。结果扎堆，赢了接着赢，亏了接着亏。",[167,276,277,280],{},[44,278,279],{},"Z 为正","：游程比随机情况多。盈亏交替得比抛硬币还频繁。",[167,282,283,286],{},[44,284,285],{},"接近零","：和相互独立的交易没有区别。",[10,288,289,290,293],{},"账户页面上，Z 值后面的括号里还有一个百分比：把同一个值压缩到 0–100 的刻度上，",[44,291,292],{},"50% 表示和抛硬币一样随机","。低于 50% 说明比随机更容易连续，高于 50% 说明更容易交替；Z 为 −2.41 时显示 8.24%，Z 为 +1.00 时显示 73.11%。这只是换个更好读的说法，不是 p 值，所以 8.24% 不是显著性水平。",[14,295,70],{"id":70},[10,297,298,299,303],{},"策略统计几乎都默认交易之间相互独立。Z 值一旦显示并非如此，页面上其他数字的含义就会悄悄变样。亏损扎堆出现时，",[35,300,302],{"href":301},"\u002Fglossary\u002Fwhat-is-drawdown","回撤","会比同样胜率下预想的更深，因为回撤不光看亏了几笔，还看这些亏损是怎么排列的。",[10,305,306,307,311],{},"Z 值明显为负，原因往往并不陌生。网格、逆势加仓摊平和篮子系统会围绕同一个判断开出很多仓位，再一起平掉，于是一个决定就产生一长串同号结果。凡是在亏损仓位上逐步加仓的做法都是如此，这类账户通常在净值回撤和",[35,308,310],{"href":309},"\u002Fglossary\u002Fwhat-is-drawdown-on-balance","余额回撤","之间也有很大差距。",[10,313,314,315,319],{},"Z 值明显为正，通常说明账户里跑的根本不是一个序列。对冲仓位会在同一时刻各记下一笔盈利和一笔亏损；两个互不相关的策略在同一个登录号上交错运行，会制造出一种不属于任何一方的人为交替。交易如果带有",[35,316,318],{"href":317},"\u002Fglossary\u002Fwhat-is-a-magic-number","魔术号","，只有按交易流分开看，才能得到一个可信的数值。",[10,321,322],{},"它的实际用途不宽，但很实在：Z 值为负，是给连续亏损设规则的统计依据；接近零，则说明这类规则没有必要。",[14,324,97],{"id":97},[10,326,100,327,104,329,332,333,335,336,339,340,207],{},[44,328,103],{},[35,330,331],{"href":131},"胜率","为 ",[44,334,127],{},"，大约是 118 笔盈利、53 笔亏损。按这个比例，随机情况下预计约有 ",[44,337,338],{},"74 个游程","，一个标准差约 ",[44,341,342],{},"5.6 个游程",[10,344,345,346,349,350,353,354,358],{},"也就是说，在中位数账户上，游程数要缩到 63 个左右或超过 85 个，Z 值才会到 ±2。落在这两个界限之间，都和抛硬币没有区别。这实实在在限制了这个数字的分量，而历史越长，界限收得越紧：同样的规律出现在 1,700 笔交易上算是发现，出现在 171 笔上只算线索。还有一点：这些账户中 ",[44,347,348],{},"53.9%"," 的自动化程度超过 90%，",[44,351,352],{},"42.2%"," 低于 10%。扎堆现象在前一组里要常见得多，因为",[35,355,357],{"href":356},"\u002Fglossary\u002Fwhat-is-an-expert-advisor","机器人","平掉一篮子订单时，会一次往序列里写进好几个相同的结果。",[14,360,137],{"id":136},[10,362,363,364,366,367,370,371,373,374,378],{},"这项指标在每个已发布账户页面的",[44,365,143],{},"模块里，名为 ",[44,368,369],{},"Z 值（概率）","，显示两位小数的数值，旁边是换算后的百分比，位于",[44,372,155],{},"一栏，和盈利因子、标准差、夏普比率、期望值、",[35,375,377],{"href":376},"\u002Fglossary\u002Fwhat-is-ahpr-and-ghpr","AHPR 和 GHPR"," 排在一起。数值由我们的统计服务在账户同步时算好，不是在您的浏览器里计算。",[10,380,381,382,385],{},"同一页面的 ",[44,383,384],{},"Smart Insights"," 面板也会据此给出一条结论：账户至少有 30 笔已平仓交易、且数值在 ±2 以内时，会提示交易模式稳定。",[14,387,162],{"id":162},[164,389,390,396,402,408],{},[167,391,169,392,395],{},[44,393,394],{},"Z 值为负是坏事。","”它不衡量好坏，只描述交易之间是否相互关联，和赚不赚钱无关。很多优秀的趋势系统，Z 值都是负的。",[167,397,169,398,401],{},[44,399,400],{},"它能预测下一笔交易。","”它总结的是已经发生的序列。游程检验里没有任何预测的成分。",[167,403,169,404,407],{},[44,405,406],{},"这个百分比就是系统有效的概率。","”它只是同一个值换了个刻度，仅此而已。百分之五十是随机的中点，不是及格线。",[167,409,169,410,413],{},[44,411,412],{},"顺序无所谓。","”顺序恰恰是唯一要紧的。同一秒内平仓的几笔交易，彼此之间没有真正意义上的先后，所以批量平掉篮子订单的系统，游程数可能因为和策略无关的原因而变化。",[10,415,416,417,207],{},"这类统计量需要多少笔交易才有意义，请参阅",[35,418,420],{"href":419},"\u002Fguides\u002Fhow-much-history-a-track-record-needs","一份业绩记录需要多长的历史",{"title":30,"searchDepth":209,"depth":209,"links":422},[423,424,425,426,427],{"id":16,"depth":209,"text":16},{"id":70,"depth":209,"text":70},{"id":97,"depth":209,"text":97},{"id":136,"depth":209,"text":137},{"id":162,"depth":209,"text":162},"交易中的 Z 值把盈亏的连续程度和随机情况做比较。本文介绍我们的 Z 值（概率）怎么算，以及为什么 171 笔交易的样本还很单薄。",{},"\u002Fzh\u002Fglossary\u002Fwhat-is-a-z-score-in-trading",[432,433,434,225],"what-is-ahpr-and-ghpr","what-is-drawdown-on-balance","what-is-a-magic-number",{"title":237,"description":428},"zh\u002F5.glossary\u002Fwhat-is-a-z-score-in-trading","Z 值 (Z-Score)","EaJTDRCihu0CO4ef0bSKbKcJ2LSNVoL-2S08WChZfBg",{"id":440,"title":441,"body":442,"category":216,"date":217,"description":667,"draft":219,"extension":220,"meta":668,"navigation":222,"path":669,"related":670,"seo":674,"stem":675,"term":676,"updated":232,"__hash__":677},"glossary\u002Fzh\u002F5.glossary\u002Fwhat-is-expectancy.md","什么是期望值？公式、示例与真实数据",{"type":7,"value":443,"toc":660},[444,447,449,455,462,465,471,474,476,482,490,498,500,512,528,547,549,561,603,623,625,654],[10,445,446],{},"期望值就是一笔交易的平均结果，可以用货币或点数表示：账户每开一笔仓位，平均能赚多少。它由胜率、平均盈利、亏损率、平均亏损四个数字算出，回答的是“这套系统每做一笔到底赚不赚钱”，而且结果可以直接乘以交易笔数。正期望值重复得足够多，就是一门生意；负期望值则是在倒数计时。",[14,448,16],{"id":16},[21,450,453],{"className":451,"code":452,"language":26},[24],"期望值 = (胜率 × 平均盈利) − (亏损率 × 平均亏损)\n\n亏损率 = 1 − 胜率\n",[28,454,452],{"__ignoreMap":30},[10,456,457,458,461],{},"假设一个系统胜率 40%，每笔盈利单赚 $300，每笔亏损单亏 $100：(0.40 × $300) − (0.60 × $100) = $120 − $60 = ",[44,459,460],{},"每笔交易 +$60","。一年做四百笔，预期结果就是 $24,000，这还没考虑仓位大小的变化，也没算上任何意外。",[10,463,464],{},"把同一个公式改用 R 倍数来写（R 即一笔平均亏损），账户货币的影响就去掉了，不同系统也能放在一起比较：",[21,466,469],{"className":467,"code":468,"language":26},[24],"期望值 (R) = (胜率 × 盈亏比) − (1 − 胜率)\n\n盈亏比 = 平均盈利 \u002F 平均亏损\n",[28,470,468],{"__ignoreMap":30},[10,472,473],{},"算出来的数字有没有意义，取决于两点。第一，输入里必须已经扣除成本：点差、佣金和库存费每笔都要付，盈利单因此变小，亏损单因此变大。第二，平均值背后要有足够多的交易，不能让某一笔特别大的结果单独左右它。",[14,475,70],{"id":70},[10,477,478,479,481],{},"统计数字要变成交易计划，靠的就是期望值。",[35,480,120],{"href":119},"告诉您赚的钱和亏的钱之比；期望值告诉您再多做一笔交易值多少钱，把它乘以交易频率，就能得出一个月的预期结果。",[10,483,484,485,489],{},"胜率掩盖的取舍，也会在期望值里暴露出来。胜率只是优势的一半，另一半是",[35,486,488],{"href":487},"\u002Fglossary\u002Fwhat-is-risk-reward-ratio","风险回报比","，期望值正是两者的交汇点。动其中任何一个，结果都会跟着变；两个都不动，故事讲得再好，结果也不会变。",[10,491,492,493,497],{},"期望值还划出了成本的底线。如果一个策略每笔的期望是 +$8，而一个来回的",[35,494,496],{"href":495},"\u002Fglossary\u002Fwhat-is-a-spread","点差","加佣金就要 $7，剩下的那点东西已经算不上策略了。",[14,499,97],{"id":97},[10,501,502,503,104,506,104,509,207],{},"以下数字描述的是 ShowMyTrades 上已发布的账户，不代表全体交易者。在有交易历史的已发布账户中（2026 年 8 月）取中间水平：",[44,504,505],{},"胜率中位数 68.8%",[44,507,508],{},"盈利因子中位数 1.28",[44,510,511],{},"已平仓交易数中位数 171 笔",[10,513,514,515,518,519,522,523,207],{},"有了这几个数字，这个“中间账户”的期望值就能算出来。把盈利因子公式反推回去，平均盈利约为平均亏损的 ",[44,516,517],{},"0.58 倍","；剩下的用 R 形式一算便知：(0.688 × 0.58) − 0.312 = ",[44,520,521],{},"+0.087R","。也就是说，每做一笔交易，只赚到一笔平均亏损的百分之九。171 笔交易下来，全部利润加起来大约只相当于十五笔平均亏损，所以尽管近七成交易以盈利平仓，",[44,524,525,527],{},[35,526,108],{"href":107},"的中位数也只有 +3.2%",[10,529,530,531,534,535,538,539,542,543,546],{},"优势薄到这个程度，成本就不是可以忽略的零头了。已发布账户累计支付了 ",[44,532,533],{},"$4,782,670 佣金","和 ",[44,536,537],{},"$862,547 库存费","；在产生过库存费的账户中，",[44,540,541],{},"86.3%"," 的净库存费为负。持仓时间中位数只有 ",[44,544,545],{},"2.4 小时","，在这种节奏下，0.087R 的优势要想保住，一个来回的成本相对平均亏损就必须足够小。",[14,548,137],{"id":136},[10,550,551,552,554,555,557,558,560],{},"每个已发布账户页面的",[44,553,143],{},"模块里都有",[44,556,93],{},"这一行，位于",[44,559,155],{},"分组。它在同一行里给出两种单位（先点数，后账户货币）：点数便于在不同品种之间比较，货币金额便于在不同仓位大小之间比较，单看哪一个都不完整。",[10,562,563,564,566,567,569,570,569,573,576,577,579,580,586,587,576,590,593,594,569,597,576,599,602],{},"它的四个输入在同一模块的",[44,565,182],{},"分组里：",[44,568,331],{},"、",[44,571,572],{},"平均盈利",[44,574,575],{},"平均亏损","和",[44,578,186],{},"，旁边还有",[35,581,583],{"href":582},"\u002Fglossary\u002Fwhat-is-average-trade-length",[44,584,585],{},"平均持仓时间","，能看出这点薄薄的单笔优势是几小时赚到的，还是几周才赚到的。已经扣除的成本显示在",[44,588,589],{},"佣金合计",[44,591,592],{},"已付库存费合计","两行，每笔交易的明细则在交易表格的",[44,595,596],{},"盈利（毛额）",[44,598,38],{},[44,600,601],{},"佣金","列。",[10,604,605,606,612,613,569,616,576,619,622],{},"想检验一个期望值到底是一种优势，还是几种优势平均出来的，可以打开",[35,607,609],{"href":608},"\u002Fguides\u002Fcustom-analysis-and-filter-presets",[44,610,611],{},"自定义分析","侧边面板：按品种、魔术号、方向、日期范围、星期或小时筛选出子集，整个区块就会重新计算。高级统计还提供",[44,614,615],{},"星期",[44,617,618],{},"小时",[44,620,621],{},"持续时间","视图，不用筛选就能把同一结果拆开来看。",[14,624,162],{"id":162},[164,626,627,636,642,648],{},[167,628,169,629,632,633,635],{},[44,630,631],{},"正期望值就意味着我会赚钱。","”它只说明平均每笔交易是盈利的。仓位定错了，正期望值的系统照样会亏光：能不能活下来，是",[35,634,80],{"href":79},"的问题，不是期望值的问题。",[167,637,169,638,641],{},[44,639,640],{},"期望值就是我的平均每笔交易。","”前提是平均值按净额计算。按毛利润算就漏掉了佣金和库存费，而薄优势恰恰是在这里消失的。",[167,643,169,644,647],{},[44,645,646],{},"高胜率就意味着正期望值。","”单看胜率说明不了任何问题。盈亏比为 0.58 时，60% 的胜率是个亏钱的系统。",[167,649,169,650,653],{},[44,651,652],{},"这个数字很稳定。","”只有几百笔交易时并不稳定。一笔特别大的盈利单就能把整个记录的期望值抬上去，把它去掉再算一遍，是个公平的检验。",[10,655,656,657,207],{},"哪些指标值得长期追踪、该按什么顺序看，请参阅",[35,658,659],{"href":205},"交易表现追踪完全指南",{"title":30,"searchDepth":209,"depth":209,"links":661},[662,663,664,665,666],{"id":16,"depth":209,"text":16},{"id":70,"depth":209,"text":70},{"id":97,"depth":209,"text":97},{"id":136,"depth":209,"text":137},{"id":162,"depth":209,"text":162},"期望值是单笔交易的平均盈亏：(胜率 × 平均盈利) − (亏损率 × 平均亏损)。本文给出公式、完整算例和公开账户的真实数据。",{},"\u002Fzh\u002Fglossary\u002Fwhat-is-expectancy",[671,672,227,673],"what-is-a-profit-factor","what-is-risk-reward-ratio","what-is-drawdown",{"title":441,"description":667},"zh\u002F5.glossary\u002Fwhat-is-expectancy","期望值 (Expectancy)","4gLLgYBgHrrvWKTKFixk2tUfM9kXODaMx4LcyeirnhQ",{"id":679,"title":680,"body":681,"category":878,"date":217,"description":879,"draft":219,"extension":220,"meta":880,"navigation":222,"path":881,"related":882,"seo":886,"stem":887,"term":888,"updated":217,"__hash__":889},"glossary\u002Fzh\u002F5.glossary\u002Fwhat-is-position-sizing.md","什么是仓位管理？公式与手数计算",{"type":7,"value":682,"toc":871},[683,689,691,694,700,703,709,717,720,740,747,750,752,755,762,764,779,782,794,796,803,831,833,864],[10,684,685,686,688],{},"仓位管理（position sizing）决定一笔交易开多大，依据是这笔交易失败时您愿意亏掉账户的多少。它用两个数把风险百分比换算成手数：到止损的距离，以及这个手数下一个点值多少钱。账户的",[35,687,302],{"href":301},"特征就由这一个变量决定，而且它是在交易前选定的，不是交易后才发现的。",[14,690,16],{"id":16},[10,692,693],{},"分三步，按顺序来。",[21,695,698],{"className":696,"code":697,"language":26},[24],"1. 风险金额   = 账户余额 × 风险 %\n2. 点值　　   = 合约规模 × 以报价货币计的一个点 × 换算为账户货币的汇率\n3. 仓位大小   = 风险金额 \u002F (止损距离点数 × 每手点值)\n",[28,699,697],{"__ignoreMap":30},[10,701,702],{},"举个例子：$10,000 的账户，每笔交易风险 1%，EUR\u002FUSD 止损 50 点，一标准手每点 $10：",[21,704,707],{"className":705,"code":706,"language":26},[24],"风险金额   = 10,000 × 0.01      = $100\n仓位大小   = 100 \u002F (50 × 10)    = 0.20 手\n",[28,708,706],{"__ignoreMap":30},[10,710,711,712,716],{},"止损放宽到 100 点，同样 $100 的风险就只能开 0.10 手。风险不变，手数跟着调整。整个思路就在这种反向关系里，也正因为如此，仓位管理和",[35,713,715],{"href":714},"\u002Fglossary\u002Fwhat-is-a-stop-loss","止损","位置不能分开决定。",[10,718,719],{},"实践中最常用的有三种方法：",[164,721,722,728,734],{},[167,723,724,727],{},[44,725,726],{},"固定手数","。不管止损多远、余额多少，永远 0.10 手。简单，但止损和余额一变，单笔风险就跟着漂移。",[167,729,730,733],{},[44,731,732],{},"百分比风险（固定比例）","。也就是上面的公式。风险始终与账户成比例，账户缩水时，按金额计的亏损也随之变小，回本在算术上才有可能。",[167,735,736,739],{},[44,737,738],{},"马丁格尔递增","。亏损后加大手数，想把亏损扳回来。盈利频繁而且每笔都小，净值曲线漂亮得无可挑剔，但结果的分布有一条很厚的左尾，而且会一下子全部兑现。",[10,741,742],{},[743,744],"img",{"alt":745,"src":746},"每次亏损后手数加倍，对比按净值固定比例下单：胜率相同，市场相同，结局相反","\u002Fimages\u002Farticles\u002Flot-progression.svg",[10,748,749],{},"一个账户用的是哪一种，光看手数就能看出来。",[14,751,70],{"id":70},[10,753,754],{},"入场决定您能不能赢，仓位管理决定您还能不能留下来做下一笔。两个交易者用同样的信号、同样的止损，胜率一模一样，账户却天差地别，因为一个每笔只冒 0.5% 的风险，另一个冒 8%。",[10,756,757,761],{},[35,758,760],{"href":759},"\u002Fglossary\u002Fwhat-is-leverage","杠杆","真正造成伤害，也是通过仓位管理。高杠杆本身不是风险，只是允许您去冒险。风险来自用这份许可开出的仓位：按它的止损距离，一笔亏损大到账户连亏两次都扛不住。",[14,763,97],{"id":97},[10,765,766,767,770,771,104,775,778],{},"仓位管理会在回撤的尾部留下痕迹。在 ShowMyTrades 上有交易历史的已发布账户中（2026 年 8 月；统计对象是在本站公开的账户，并不代表所有交易者），",[44,768,769],{},"38.2%"," 曾",[35,772,774],{"href":773},"\u002Fglossary\u002Fwhat-is-maximum-drawdown","回撤超过 20%",[44,776,777],{},"17.6%"," 亏掉了峰值的一半以上。",[10,780,781],{},"规模定得合理的固定比例账户，很少会落到最后这一档。按每笔 1% 的风险亏掉 50%，大约要连亏 69 次。落在这条尾部里的账户，一般另有原因：亏损后不断加码，或者定手数时根本没考虑止损。",[10,783,784,785,104,788,790,791,793],{},"再看看仓位大小是谁定的：这些账户自动交易占比的中位数为 ",[44,786,787],{},"99%",[44,789,348],{}," 的账户自动化程度超过 90%，低于 10% 的则有 ",[44,792,352],{},"。在一半以上的已发布账户上，仓位大小只是 EA 设置里的一行，设好一次，余额怎么变都不再调整。",[14,795,137],{"id":136},[10,797,798,802],{},[35,799,801],{"href":800},"\u002Ftools\u002Fposition-size-calculator","仓位计算器","用实时价格、以您的账户货币套用上面的公式，风险可以按百分比或固定金额设定，止损可以用价格或点数表示。",[10,804,805,806,809,810,813,814,816,817,820,821,823,824,830],{},"在已发布账户的页面上，仓位管理可以实际核查，不必凭空猜测。",[44,807,808],{},"交易表格","有一列",[44,811,812],{},"交易量","，列出每笔交易的手数，每亏一次就加一级的手数一眼就能看出来。",[44,815,611],{},"提供",[44,818,819],{},"手数","的最小\u002F最大筛选：把下限设在账户正常手数之上，看看剩下的交易是不是都紧跟在亏损之后出现，这就是马丁格尔的特征。",[44,822,143],{},"模块里的",[35,825,827],{"href":826},"\u002Fguides\u002Ftrading-volume-analysis",[44,828,829],{},"总手数","给出汇总值；所有已发布账户累计交易了 1,724,575 手。",[14,832,162],{"id":162},[164,834,835,841,852,858],{},[167,836,169,837,840],{},[44,838,839],{},"风险 2%，就是动用账户的 2%。","”不是。2% 是止损被触发时的亏损。仓位占用的保证金是另一个完全不同的数字，由杠杆决定。",[167,842,169,843,846,847,851],{},[44,844,845],{},"我每笔只冒 1% 的风险，所以最坏也就亏 1%。","”只有同时只开一个仓位时才成立。",[35,848,850],{"href":849},"\u002Fglossary\u002Fwhat-is-correlation-risk","五个相关的仓位","各占 1%，其实是一笔换了装扮的 5% 交易，而相关的货币对偏偏在最伤人的时候同涨同跌。",[167,853,169,854,857],{},[44,855,856],{},"不设止损，就没有风险要算。","”没有止损，仓位大小公式就没有分母：风险不是很小，而是根本无法定义，它的真实数值就是追加保证金通知。",[167,859,169,860,863],{},[44,861,862],{},"复利就是不断加仓。","”固定比例的仓位管理是双向调整的。赢了加手数、亏了却不减，这不是复利，而是一个慢慢收紧的棘轮，把账户推向一场您定仓位时从未考虑过的回撤。",[10,865,866,867,207],{},"全套风险与仓位管理工具，以及各自适用的场景，请参阅",[35,868,870],{"href":869},"\u002Fguides\u002Fforex-calculators-guide","外汇计算器指南",{"title":30,"searchDepth":209,"depth":209,"links":872},[873,874,875,876,877],{"id":16,"depth":209,"text":16},{"id":70,"depth":209,"text":70},{"id":97,"depth":209,"text":97},{"id":136,"depth":209,"text":137},{"id":162,"depth":209,"text":162},"风险","仓位管理就是根据止损距离和点值，把风险百分比换算成手数。本文介绍公式、三种常见方法，以及仓位定错的代价。",{},"\u002Fzh\u002Fglossary\u002Fwhat-is-position-sizing",[673,883,884,885],"what-is-maximum-drawdown","what-is-a-stop-loss","what-is-leverage",{"title":680,"description":879},"zh\u002F5.glossary\u002Fwhat-is-position-sizing","仓位管理 (Position Sizing)","5hi2OmhGtFDhESyalituW-Nh51a9wIit5SKZxEMEQyk",{"id":891,"title":892,"body":893,"category":216,"date":217,"description":1129,"draft":219,"extension":220,"meta":1130,"navigation":222,"path":1131,"related":1132,"seo":1136,"stem":1137,"term":1138,"updated":232,"__hash__":1139},"glossary\u002Fzh\u002F5.glossary\u002Fwhat-is-standard-deviation-in-trading.md","交易中的标准差是什么？公式与数据",{"type":7,"value":894,"toc":1122},[895,898,900,906,926,961,964,966,969,980,983,1010,1012,1021,1024,1033,1035,1058,1083,1085,1116],[10,896,897],{},"标准差衡量的是单笔交易的结果通常离账户的平均交易结果有多远。ShowMyTrades 以账户货币显示它，所以它回答的问题很直接：在这个账户上，一笔交易和平常的交易通常相差多少？数值小，说明结果比较集中；数值大，说明这段历史是由离群值堆出来的。",[14,899,16],{"id":16},[21,901,904],{"className":902,"code":903,"language":26},[24],"标准差 = √( Σ (rᵢ − 均值)² \u002F n )\n",[28,905,903],{"__ignoreMap":30},[10,907,908,909,912,913,917,918,921,922,925],{},"其中 ",[28,910,911],{},"rᵢ"," 是第 ",[914,915,916],"em",{},"i"," 笔已平仓交易的净结果，",[28,919,920],{},"均值"," 是这些结果的平均值，",[28,923,924],{},"n"," 是已平仓交易笔数。这个数字代表什么，取决于四个细节：",[164,927,928,936,942,955],{},[167,929,930,207,933,935],{},[44,931,932],{},"每笔交易都按扣除成本后的净额计入",[28,934,911],{}," 等于盈利 + 佣金 + 库存费，所以佣金或库存费高的交易，压缩了多少盈利，就同样放大了多少离散程度。",[167,937,938,941],{},[44,939,940],{},"只统计已平仓的买入和卖出交易","。入金和出金在生成序列之前就已剔除，所以往账户里注资不会变成一个巨大的离群值。",[167,943,944,947,948,950,951,954],{},[44,945,946],{},"它用的是总体标准差","，除以 ",[28,949,924],{}," 而不是 ",[28,952,953],{},"n − 1","。已平仓交易不足两笔时，显示为 0.00。",[167,956,957,960],{},[44,958,959],{},"它以货币计，不是百分比","。同样是 $50 的标准差，在 $2,000 的账户和 $200,000 的账户上，说明的是完全不同的交易方式。一定要对照账户余额来看。",[10,962,963],{},"由于每个距离都先平方再平均，一笔异常交易对结果的影响远超十几笔普通交易，所以它特别容易揪出那笔格格不入的交易。",[14,965,70],{"id":70},[10,967,968],{},"“稳定”这个词是真是假，标准差一测便知：谁都可以说自己的净值曲线很平稳，标准差则用一个货币数值告诉您，这些交易到底有多平稳。",[10,970,971,972,976,977,979],{},"它也是我们的",[35,973,975],{"href":974},"\u002Fglossary\u002Fwhat-is-the-sharpe-ratio","夏普比率","的分母：夏普比率就是用这个数字作单位来衡量的平均交易结果。它还是",[35,978,93],{"href":92},"的天然搭档：期望值给出每笔交易的平均结果，却看不出这个平均结果有多靠得住。",[10,981,982],{},"最有用的读法是比较，拿它和几行之外的两个平均值对照：",[164,984,985,991,1004],{},[167,986,987,990],{},[44,988,989],{},"接近平均盈利","：结果很集中，账户大部分时候表现一致。",[167,992,993,996,997,1000,1001,207],{},[44,994,995],{},"是平均盈利的好几倍","：整段历史靠少数几笔交易撑着。相信总数之前，先看",[44,998,999],{},"最佳交易（盈亏）"," 和",[44,1002,1003],{},"最差交易（盈亏）",[167,1005,1006,1009],{},[44,1007,1008],{},"远大于平均亏损","：至少有一笔亏损远远超出平常水平。",[14,1011,97],{"id":97},[10,1013,1014,1015,104,1018,207],{},"以下数字描述的是 ShowMyTrades 上已发布的账户，不代表全体交易者。在有交易历史的已发布账户中（2026 年 8 月），",[44,1016,1017],{},"夏普比率中位数为 0.05",[44,1019,1020],{},"已平仓交易数中位数为 171 笔",[10,1022,1023],{},"按我们自己的定义倒推，这组数字能把离散程度说得很准：夏普比率是平均交易结果除以标准差，中位数 0.05 意味着，中间水平的账户平均每笔交易只值交易标准差的二十分之一左右。也就是说，一笔交易和下一笔之间的正常波动，大约是从中赚到的优势的二十倍。",[10,1025,1026,1032],{},[44,1027,1028,1031],{},[35,1029,1030],{"href":582},"持仓时间","中位数为 2.4 小时","，所以看一整天的结果，交易者学不到任何东西：一天的结果完全落在正常波动范围之内。同样道理，新账户的标准差低，也证明不了风险受控：样本里可能只是还没出现这个策略会产生的那个离群值，而距离要平方，它一旦出现，动静就会很大。",[14,1034,137],{"id":136},[10,1036,1037,1038,1040,1041,1043,1044,1046,1047,1049,1050,1052,1053,1057],{},"账户页面的",[44,1039,143],{},"模块里，",[44,1042,155],{},"一栏中有一行，标签正是",[44,1045,85],{},"，以账户货币显示，并附有提示说明。它紧挨在以它为分母的",[44,1048,975],{},"上方、",[44,1051,120],{},"下方，由我们的服务器根据存储的已平仓交易计算。在已发布的账户页面上，这个模块默认显示，不过",[35,1054,1056],{"href":1055},"\u002Fguides\u002Fpublic-profile-and-custom-slugs","账户主人可以隐藏它","；可嵌入的小组件里没有这个模块。",[10,1059,1060,1061,1063,1064,569,1066,569,1068,569,1070,1000,1072,1074,1075,1077,1078,1082],{},"把它和同一模块的",[44,1062,182],{},"一栏对照着看：",[44,1065,572],{},[44,1067,575],{},[44,1069,999],{},[44,1071,1003],{},[44,1073,186],{},"为这个数字提供了参照。模块里的",[44,1076,621],{},"视图能把离散程度直接画出来：每笔已平仓交易是一个点，横轴是持仓小时数，纵轴是盈亏，离散程度就从一个数字变成了一个形状。",[35,1079,1080],{"href":608},[44,1081,611],{},"会在筛选出的子集上重新计算整个模块，可按日期范围、品种、魔术号、交易类型、手数等筛选，由此看出离散程度来自策略本身，还是来自某一个交易品种。",[14,1084,162],{"id":162},[164,1086,1087,1093,1104,1110],{},[167,1088,169,1089,1092],{},[44,1090,1091],{},"它是一个百分比。","”我们的标准差是每笔交易的金额。旁边没有余额作参照，在不同规模的账户之间比较毫无意义。",[167,1094,169,1095,1098,1099,1103],{},[44,1096,1097],{},"它衡量的是市场波动率。","”它衡量的是您交易结果的",[35,1100,1102],{"href":1101},"\u002Fglossary\u002Fwhat-is-volatility","波动率","。市场再平静，只要仓位越加越大，数值照样会很大。",[167,1105,169,1106,1109],{},[44,1107,1108],{},"标准差低，风险就低。","”只有已平仓交易计入其中。扛着亏损单不平的系统，已平仓结果个个都小而整齐，真正的风险敞口却藏在浮动仓位里。",[167,1111,169,1112,1115],{},[44,1113,1114],{},"30 笔交易下数值很低，可以放心。","”不到几百笔交易时，只能把它当作暂时的结果。",[10,1117,1118,1119,1121],{},"一组统计数据要有多少笔交易才有意义，请参阅",[35,1120,420],{"href":419},"指南。",{"title":30,"searchDepth":209,"depth":209,"links":1123},[1124,1125,1126,1127,1128],{"id":16,"depth":209,"text":16},{"id":70,"depth":209,"text":70},{"id":97,"depth":209,"text":97},{"id":136,"depth":209,"text":137},{"id":162,"depth":209,"text":162},"标准差衡量单笔交易的结果离您的平均交易结果有多远。本文介绍我们使用的公式（已扣除成本），以及如何在数千个账户上解读它。",{},"\u002Fzh\u002Fglossary\u002Fwhat-is-standard-deviation-in-trading",[1133,1134,225,1135],"what-is-the-sharpe-ratio","what-is-volatility","what-is-the-sortino-ratio",{"title":892,"description":1129},"zh\u002F5.glossary\u002Fwhat-is-standard-deviation-in-trading","标准差 (Standard Deviation)","d23H7esNsSG175eWPyoZzoJTWWR3SEfy5D-ZKP-x9wU",[1141,1142],{"path":376},{"path":223},1793898058362]