[{"data":1,"prerenderedAt":1256},["ShallowReactive",2],{"\u002Fzh\u002Fglossary\u002Fwhat-is-forex-pair-correlation":3,"\u002Fzh\u002Fglossary\u002Fwhat-is-forex-pair-correlation-related":299,"\u002Fzh\u002Fglossary\u002Fwhat-is-forex-pair-correlation-langs":1253},{"id":4,"title":5,"body":6,"category":281,"date":282,"description":283,"draft":284,"extension":285,"meta":286,"navigation":287,"path":288,"related":289,"seo":294,"stem":295,"term":296,"updated":297,"__hash__":298},"glossary\u002Fzh\u002F5.glossary\u002Fwhat-is-forex-pair-correlation.md","什么是外汇货币对相关性？隐藏的集中风险",{"type":7,"value":8,"toc":273},"minimark",[9,19,23,26,37,45,57,60,63,69,72,85,88,99,119,134,138,186,196,227,230,266],[10,11,12,13,18],"p",{},"外汇货币对相关性，指两个货币对一起波动的程度，用介于 −1 和 +1 之间的系数表示。每个货币对都是两种货币的比价，所以只要两个货币对含有同一种货币，它们在一定程度上就是在押同一件事。持有",[14,15,17],"a",{"href":16},"\u002Fglossary\u002Fwhat-is-correlation-risk","三个相关仓位","的交易者，并没有把风险分散到三笔交易上：他持有的其实是一个更大的仓位，而这个仓位的大小是碰巧凑出来的，不是自己决定的。",[20,21,22],"h2",{"id":22},"工作原理",[10,24,25],{},"这个系数等于两组收益序列的协方差，除以二者标准差的乘积。",[27,28,33],"pre",{"className":29,"code":31,"language":32},[30],"language-text","ρ = cov(A, B) \u002F (σA × σB)\n\nρ = +1.0        走势完全相同\nρ =  0.0        互不相关\nρ = −1.0        走势完全相反\n|ρ| > 0.7       计算仓位时把两者视为一个仓位\n","text",[34,35,31],"code",{"__ignoreMap":36},"",[10,38,39,40,44],{},"背后的机制是算术，不是市场情绪。做多 EUR\u002FUSD、做多 GBP\u002FUSD、做多 AUD\u002FUSD，是做空美元的三种方式，美元一走强，三笔同时亏。做多 EUR\u002FUSD、做多 EUR\u002FGBP、做多 EUR\u002FJPY，则是做多欧元的三种方式。还有些货币对从构成上就是负相关的：EUR\u002FUSD 和 USD\u002FCHF 里的美元在相反的两侧，两个都做多，几乎等于什么都没持有，却要付两份",[14,41,43],{"href":42},"\u002Fglossary\u002Fwhat-is-a-spread","点差","。",[10,46,47,48,52,53,56],{},"它的危险来自两个特性。第一，相关性",[49,50,51],"strong",{},"不稳定","：最近一个月测出的系数，说的不是这两个货币对的固有属性，只是最近一个月的观察结果。第二，它",[49,54,55],{},"在压力下会升高","：正当最需要分散化的时候，资金成批涌入或撤出美元，各货币对的相关性都向 1 靠拢。",[20,58,59],{"id":59},"为什么重要",[10,61,62],{},"相关仓位的风险是相乘的，不是相加的。合并后的风险敞口，遵循的是和任何双资产投资组合相同的公式。",[27,64,67],{"className":65,"code":66,"language":32},[30],"σ_total = √( σA² + σB² + 2 × ρ × σA × σB )\n",[34,68,66],{"__ignoreMap":36},[10,70,71],{},"ρ = 0 时，两个 1% 的仓位合起来约为 1.4% 的风险敞口；ρ = 0.9 时，合起来约为 1.95%，几乎等于两者相加。三笔这样的交易，单笔都在合理的风险限额之内，合起来却突破了账户的限额，而每一条单独的规则都没有违反。",[10,73,74,75,79,80,84],{},"这就是为什么",[14,76,78],{"href":77},"\u002Fglossary\u002Fwhat-is-position-sizing","仓位管理","很有纪律的账户，照样会出现大幅回撤。规则一条没破，只是数的是仓位个数，而不是风险敞口；整个持仓组合的",[14,81,83],{"href":82},"\u002Fglossary\u002Fwhat-is-volatility","波动率","，是交易者自以为的两到三倍。",[20,86,87],{"id":87},"数据显示了什么",[10,89,90,91,94,95,98],{},"以 ShowMyTrades 上有交易历史的公开账户为样本（2026 年 8 月），总交易量为 ",[49,92,93],{},"1,724,575 手","，账户已平仓交易数的中位数为 ",[49,96,97],{},"171 笔","。这个数量足以让集中度在品种分类统计里显现出来，也足以让一组相关仓位主导整份记录。",[10,100,101,102,106,107,110,111,114,115,118],{},"没有明说的集中度，会在回撤分布里露出马脚。",[14,103,105],{"href":104},"\u002Fglossary\u002Fwhat-is-maximum-drawdown","历史最深回撤的中位数","为 ",[49,108,109],{},"9.7%","，但 ",[49,112,113],{},"38.2%"," 的账户曾从峰值下跌超过 20%，",[49,116,117],{},"17.6%"," 曾亏掉峰值的一半以上。单笔手数都不大的账户亏到这种程度，通常意味着许多仓位在同进同退。",[10,120,121,122,125,126,129,130,133],{},"用反向持仓抵消风险敞口而不平仓，是要持续付费的，这在总额里看得出来。这些账户共支付了 ",[49,123,124],{},"$4,782,670 的佣金","和 ",[49,127,128],{},"$862,547 的库存费","；在隔夜持仓的账户中，",[49,131,132],{},"86.3%"," 的净库存费为负。两个货币对互相对冲着持有，费用照样一天天累计，而当初开仓想承担的风险，大半却被抵消掉了。",[20,135,137],{"id":136},"在-showmytrades-上的位置","在 ShowMyTrades 上的位置",[10,139,140,143,144,147,148,151,152,155,156,159,160,151,163,166,167,170,171,174,175,178,179,155,182,185],{},[49,141,142],{},"分类统计","模块就是检查集中度的工具。它的",[49,145,146],{},"按品种","标签页列出账户交易过的每个品种，分",[49,149,150],{},"做多","、",[49,153,154],{},"做空","和",[49,157,158],{},"合计","三栏，每栏都有",[49,161,162],{},"交易",[49,164,165],{},"点数","和盈利，另有",[49,168,169],{},"盈利单（%）","。先看",[49,172,173],{},"品种","列：好几行都含 USD、方向又全都一样，那就是一个美元仓位拆到了几个代码上。",[49,176,177],{},"图表","视图会画出",[49,180,181],{},"按货币对统计交易",[49,183,184],{},"按货币对统计盈利","，不读表格也能看出交易是否集中在最大的几个品种上。",[10,187,188,195],{},[14,189,191,194],{"href":190},"\u002Fguides\u002Fmagic-numbers-and-strategy-labels",[49,192,193],{},"按魔术号","标签页","对策略做同样的检查：这些相关的行是同一个 EA（智能交易系统）做出来的，还是几个 EA 撞在了同一种货币上，一看便知。",[10,197,198,199,202,203,206,207,210,211,214,215,218,219,222,223,226],{},"在",[49,200,201],{},"交易表格","里看",[49,204,205],{},"开仓时间","列，留意扎堆的地方：同一分钟内在不同品种上开出好几张订单，这是一次相关的入场，不是分散的入场。",[49,208,209],{},"自定义分析","侧边面板可以按品种、魔术号以及",[49,212,213],{},"买入（做多）"," \u002F ",[49,216,217],{},"卖出（做空）"," 筛选，排除某一组货币后重新计算统计数据，看看这份记录有多少是靠它撑着的。再看",[49,220,221],{},"图表模块","的",[49,224,225],{},"回撤","视图，效果就出来了：相关的持仓组合是一下子跌下去的，而不是分好几步。",[20,228,229],{"id":229},"常见误解",[231,232,233,241,250,260],"ul",{},[234,235,236,237,240],"li",{},"“",[49,238,239],{},"一个货币对就是一个交易品种。","”它是两种货币。EUR\u002FUSD 同时是一个欧元仓位和一个美元仓位，所以凡是报价里含有这两种货币之一的品种，它都会跟着一起动。",[234,242,236,243,246,247,249],{},[49,244,245],{},"相关系数能告诉我仓位该开多大。","”它只说明两个仓位怎样一起波动，不说明波动多远。定仓位要把相关性和",[14,248,83],{"href":82},"结合起来看：两个都很平静的品种即使高度相关，问题也不大。",[234,251,236,252,255,256,259],{},[49,253,254],{},"每个货币对只有一个相关系数。","”每个货币对在",[49,257,258],{},"每个时间窗口","里各有一个数：两个品种按小时看可能毫不相干，按周看却同进同退。",[234,261,236,262,265],{},[49,263,264],{},"这只适用于主要外汇货币对。","”以美元报价的黄金、指数和加密货币对，和 EUR\u002FUSD 归在同一组的情况，比交易者想象的更常见。",[10,267,268,269,44],{},"如果想从交易量而不是盈利的角度看集中度，请参阅",[14,270,272],{"href":271},"\u002Fguides\u002Ftrading-volume-analysis","交易量分析",{"title":36,"searchDepth":274,"depth":274,"links":275},2,[276,277,278,279,280],{"id":22,"depth":274,"text":22},{"id":59,"depth":274,"text":59},{"id":87,"depth":274,"text":87},{"id":136,"depth":274,"text":137},{"id":229,"depth":274,"text":229},"策略","2026-08-19T00:00:00.000Z","含有同一种货币的货币对会一起涨跌。相关性衡量的是这种联动有多紧，三笔看似各自独立的交易，就是因为它变成了同一个仓位。",false,"md",{},true,"\u002Fzh\u002Fglossary\u002Fwhat-is-forex-pair-correlation",[290,291,292,293],"what-is-correlation-risk","what-is-volatility","what-is-position-sizing","what-is-leverage",{"title":5,"description":283},"zh\u002F5.glossary\u002Fwhat-is-forex-pair-correlation","外汇货币对相关性 (Forex Pair Correlation)","2026-09-24T00:00:00.000Z","EO9V84b99m2LdKbsdx3ArhjOdhSDLW4oTuHtQ1aHh5M",[300,616,834,1035],{"id":301,"title":302,"body":303,"category":604,"date":282,"description":605,"draft":284,"extension":285,"meta":606,"navigation":287,"path":607,"related":608,"seo":612,"stem":613,"term":614,"updated":297,"__hash__":615},"glossary\u002Fzh\u002F5.glossary\u002Fwhat-is-correlation-risk.md","什么是相关性风险？分散投资为何失效",{"type":7,"value":304,"toc":597},[305,308,310,318,324,327,394,397,400,442,445,447,453,461,468,470,488,496,508,510,559,561,590],[10,306,307],{},"相关性风险（correlation risk）是指：一些仓位或账户看起来互不相干，背后却受同一个因素驱动，于是一起亏损。这就是为什么看似分散的持仓，到头来只出现一次毫不分散的回撤：风险是按仓位一个个算的，来的时候却是一齐到来。",[20,309,22],{"id":22},[10,311,312,313,317],{},"两个货币对之间的相关系数，属于",[14,314,316],{"href":315},"\u002Fglossary\u002Fwhat-is-forex-pair-correlation","外汇货币对相关性","的话题。放到整个持仓上，要看的是最后实际剩下几个独立的赌注。",[27,319,322],{"className":320,"code":321,"language":32},[30],"独立赌注数 = n ÷ ( 1 + (n − 1) × ρ )\n\nn = 仓位数量\nρ = 仓位之间的平均相关系数\n",[34,323,321],{"__ignoreMap":36},[10,325,326],{},"三个仓位，每个承担账户 1% 的风险：",[328,329,330,346],"table",{},[331,332,333],"thead",{},[334,335,336,340,343],"tr",{},[337,338,339],"th",{},"平均相关系数",[337,341,342],{},"合计风险",[337,344,345],{},"独立赌注数",[347,348,349,361,372,383],"tbody",{},[334,350,351,355,358],{},[352,353,354],"td",{},"ρ = 0",[352,356,357],{},"1.73%",[352,359,360],{},"3.0",[334,362,363,366,369],{},[352,364,365],{},"ρ = 0.7",[352,367,368],{},"2.68%",[352,370,371],{},"1.25",[334,373,374,377,380],{},[352,375,376],{},"ρ = 1",[352,378,379],{},"3.00%",[352,381,382],{},"1.0",[334,384,385,388,391],{},[352,386,387],{},"ρ = −0.5（三个仓位时的下限）",[352,389,390],{},"≈ 0%",[352,392,393],{},"—",[10,395,396],{},"交易者以为自己冒了三次 1% 的风险。ρ = 1 时，其实是一次冒了 3%。平台上没有任何提示会告诉您这个差别。",[10,398,399],{},"相关的风险敞口有五个来源，大多数持仓同时占了好几个：",[231,401,402,408,419,425,436],{},[234,403,404,407],{},[49,405,406],{},"共同的货币腿","。同时做多 EUR\u002FUSD、GBP\u002FUSD 和 AUD\u002FUSD，其实是换了三身行头的同一个做空美元仓位。",[234,409,410,413,414,418],{},[49,411,412],{},"一个策略跑多个品种","。一个 ",[14,415,417],{"href":416},"\u002Fglossary\u002Fwhat-is-an-expert-advisor","EA（智能交易系统）","交易十二个货币对，同一个信号条件就生成十二个仓位。",[234,420,421,424],{},[49,422,423],{},"一个策略跑多个账户","。同一个 EA 复制到四个账户上，就是把一个赌注放大了四倍；要看的是承担风险的资金总额，而不是每个账户各有多少。",[234,426,427,430,431,435],{},[49,428,429],{},"依赖同一种市场状态","。持仓里所有",[14,432,434],{"href":433},"\u002Fglossary\u002Fwhat-is-trend-following","趋势跟踪","系统，不管交易什么，遇到震荡行情都会一起吃亏。",[234,437,438,441],{},[49,439,440],{},"同一个交易场所","。同一家经纪商、同一个报价源、同一套执行。报价源一中断，或者来一波重新报价，所有仓位同时受影响。",[10,443,444],{},"而且 ρ 不是常数：市场平静时，各个仓位各走各的，持仓看上去确实分散；一旦出现流动性事件，它们就会向 1 靠拢。",[20,446,59],{"id":59},[10,448,449,450,452],{},"相关性风险会让",[14,451,78],{"href":77},"赖以成立的算术失效。“每笔交易 1%”的规则是对最坏情况的承诺，但前提是交易彼此独立。十个相关仓位各押 1%，规则一丝不差地执行，一天照样可能亏掉 10%。",[10,454,455,456,460],{},"对",[14,457,459],{"href":458},"\u002Fglossary\u002Fwhat-is-margin-level","保证金比例","的冲击也是同样道理：相关的仓位会一起陷入浮亏，整个账户的净值一齐下滑，已用保证金却纹丝不动，强平线逼近的速度比逐个分析仓位时预想的快得多。",[10,462,463,464,467],{},"它还会让业绩记录误导人。四个账户各自只有不大的",[14,465,225],{"href":466},"\u002Fglossary\u002Fwhat-is-drawdown","，可能全都发生在同样的三天里：分开看是四个数据点，实际上只有一个。",[20,469,87],{"id":87},[10,471,472,473,476,477,480,481,476,484,487],{},"在 ShowMyTrades 上（2026 年 8 月），",[49,474,475],{},"699 名用户运行不止一个账户","，",[49,478,479],{},"429 名用户在不止一家经纪商开有账户","。在有交易历史的公开账户中，自动交易占比的中位数为 ",[49,482,483],{},"99%",[49,485,486],{},"53.9% 的账户自动化程度超过 90%","。统计对象是在本站公开的账户，并不代表全体交易者。",[10,489,490,491,495],{},"这些数字说明了规模。超过一半的用户持有多个账户，约三分之一把账户分散在不同的经纪商，通常是为了不依赖单一交易对手。但这分散的是交易对手，不是策略：如果两边跑的是同一个自动化系统，",[14,492,494],{"href":493},"\u002Fguides\u002Fsame-strategy-different-brokers","两家经纪商只是把同一条净值曲线交出两遍","；而自动交易占比几乎是百分之百，这种情况就成了常态。",[10,497,498,499,106,502,110,504,507],{},"结果也印证了这一点。",[14,500,501],{"href":104},"最深回撤的中位数",[49,503,109],{},[49,505,506],{},"17.6% 的账户超过了 50%","。这么深的跌幅，很少是某一个仓位出了错，而是一组仓位到头来被证明是同一个仓位。",[20,509,137],{"id":136},[231,511,512,525,538,545,554],{},[234,513,514,516,517,519,520,524],{},[49,515,142],{},"模块里的",[49,518,193],{},"标签页，按策略而不是按交易品种拆分记录。两个",[14,521,523],{"href":522},"\u002Fglossary\u002Fwhat-is-a-magic-number","魔术号","如果好时段、坏时段都一样，就是同一个系统贴了两个标签。",[234,526,527,532,533,155,535,537],{},[14,528,530],{"href":529},"\u002Fguides\u002Fcustom-analysis-and-filter-presets",[49,531,209],{},"可以按",[49,534,173],{},[49,536,523],{},"筛选。排除一个魔术号后重新计算各项数字，就能看出这份记录有多少是靠它撑着的。已保存的预设列表在公开页面上不显示，筛选功能则照常可用。",[234,539,540,541,544],{},"图表模块的",[49,542,543],{},"增长","视图，打开两个账户页面对照着看。两条曲线如果在相同日期出现相同的高峰和低谷，不管叫什么名字，都是同一个策略。",[234,546,547,553],{},[14,548,550],{"href":549},"\u002Fglossary\u002Fwhat-is-a-trading-portfolio",[49,551,552],{},"投资组合","把多个账户合并成一个汇总视图：净值曲线、统计和指标全都合在一起。相关性风险针对的是这个总数，而不是每个账户各自的数字。",[234,555,556,558],{},[49,557,201],{},"里的魔术号和注释两列，能看出哪个系统在什么时候开了什么仓。",[20,560,229],{"id":229},[231,562,563,569,575,581],{},[234,564,236,565,568],{},[49,566,567],{},"我交易八个货币对，已经分散了。","”如果其中七个都以美元报价，您手里其实是一个放大了八倍的美元仓位。",[234,570,236,571,574],{},[49,572,573],{},"用不同的经纪商，风险就分散了。","”分散的是交易对手和执行，这确实有价值。但同一个策略在两边一模一样地跑，策略本身的风险并没有分散。",[234,576,236,577,580],{},[49,578,579],{},"相关性够稳定，可以拿来做计划。","”它是在正常行情下测出来的，而在造成大幅回撤的行情里，它会向 1 靠拢。",[234,582,236,583,586,587,589],{},[49,584,585],{},"做多一个货币对、同时做空一个相关的货币对，风险就抵消了。","”这只是把方向性赌注换成了交叉赌注，而持有它要付两份",[14,588,43],{"href":42},"和两份库存费。",[10,591,592,593,44],{},"想看到合并后的整体风险敞口，而不是四个分开的页面，请阅读",[14,594,596],{"href":595},"\u002Fguides\u002Ftracking-multiple-accounts-portfolio","在投资组合中追踪多个账户",{"title":36,"searchDepth":274,"depth":274,"links":598},[599,600,601,602,603],{"id":22,"depth":274,"text":22},{"id":59,"depth":274,"text":59},{"id":87,"depth":274,"text":87},{"id":136,"depth":274,"text":137},{"id":229,"depth":274,"text":229},"风险","相关性风险，是指几个仓位或账户看似各自独立，实际押的是同一注。本文讲解相关的风险敞口如何放大回撤，并附数千名交易者的数据。",{},"\u002Fzh\u002Fglossary\u002Fwhat-is-correlation-risk",[609,610,292,611],"what-is-forex-pair-correlation","what-is-drawdown","what-is-margin-level",{"title":302,"description":605},"zh\u002F5.glossary\u002Fwhat-is-correlation-risk","相关性风险 (Correlation Risk)","O47Vu9WYin37xzzy2BkP-7d68-Yw2EaPgPLKtMpFkVw",{"id":617,"title":618,"body":619,"category":604,"date":282,"description":824,"draft":284,"extension":285,"meta":825,"navigation":287,"path":826,"related":827,"seo":830,"stem":831,"term":832,"updated":282,"__hash__":833},"glossary\u002Fzh\u002F5.glossary\u002Fwhat-is-leverage.md","外汇中的杠杆是什么？公式与真实数字",{"type":7,"value":620,"toc":817},[621,624,626,629,635,638,686,689,692,704,706,709,712,728,730,733,740,743,745,748,782,784,810],[10,622,623],{},"杠杆（leverage）是仓位规模与持仓所需资金之间的比率。经纪商提供 1:100 杠杆，您拿出 $1,000 作保证金，就能控制 $100,000 的货币。杠杆是一种借款工具，不是策略：它只改变一笔交易占用多少资金，其他一概不变。",[20,625,22],{"id":22},[10,627,628],{},"保证金是持仓期间经纪商冻结的押金。算法很简单。",[27,630,633],{"className":631,"code":632,"language":32},[30],"仓位价值   = 合约规模 × 手数 × 价格\n所需保证金 = 仓位价值 ÷ 杠杆\n\n有效杠杆   = 全部持仓价值 ÷ 账户净值\n",[34,634,632],{"__ignoreMap":36},[10,636,637],{},"一标准手 EUR\u002FUSD 是 100,000 单位。价格在 1.0850 时，仓位价值 $108,500。",[328,639,640,653],{},[331,641,642],{},[334,643,644,647,650],{},[337,645,646],{},"账户杠杆",[337,648,649],{},"冻结的保证金",[337,651,652],{},"50 点不利波动的成本",[347,654,655,666,676],{},[334,656,657,660,663],{},[352,658,659],{},"1:30",[352,661,662],{},"$3,616.67",[352,664,665],{},"$500",[334,667,668,671,674],{},[352,669,670],{},"1:100",[352,672,673],{},"$1,085.00",[352,675,665],{},[334,677,678,681,684],{},[352,679,680],{},"1:500",[352,682,683],{},"$217.00",[352,685,665],{},[10,687,688],{},"三行的仓位一模一样。杠杆动的是保证金，从来不是风险。",[10,690,691],{},"真正反映风险敞口的是有效杠杆：所有持仓的名义价值除以净值。在 1:500 的账户上用 $10,000 做 0.05 手，有效杠杆只有 0.54:1，敞口比全额现金买入同样货币的人还小。在 1:30 的账户上用同样的 $10,000 做 2.5 手，有效杠杆是 27:1，已经顶到监管设定的上限。杠杆档位说明不了这两个人的任何问题，说明问题的是仓位大小。",[10,693,694,695,155,699,703],{},"用",[14,696,698],{"href":697},"\u002Ftools\u002Fleverage-calculator","杠杆计算器",[14,700,702],{"href":701},"\u002Ftools\u002Fmargin-calculator","保证金计算器","，拿您自己交易的品种算一算。",[20,705,59],{"id":59},[10,707,708],{},"高杠杆本身不会亏钱。它拿走的是原本能拦住您的那道限制。",[10,710,711],{},"在 1:30 的账户上，$10,000 净值最多只能开约 2.7 标准手 EUR\u002FUSD，再多保证金就不够了：等于经纪商替您设了一个仓位上限。换成 1:500，同样的 $10,000 能撑起 46 手。什么都没变好，只是上限没了。",[10,713,714,715,719,720,722,723,727],{},"把账户掏空的链条从来都一样：杠杆允许开大仓位，大仓位带来扛不住的亏损，由此造成的回撤所需要的收益，算术上根本给不出来：",[14,716,718],{"href":717},"\u002Fguides\u002Fmaximum-drawdown-explained","亏损 50% 需要盈利 100% 才能回本","。杠杆是链条的起点，但真正造成伤害的那一环是仓位规模。所以值得掌握的控制手段是",[14,721,78],{"href":77},"，而",[14,724,726],{"href":725},"\u002Fglossary\u002Fwhat-is-a-margin-call","追加保证金通知","只是症状，不是病因。",[20,729,87],{"id":87},[10,731,732],{},"截至 2026 年 8 月，连接到 ShowMyTrades 的 1 万+ 个账户分布在 703 个不同的经纪商服务器上，涵盖 MT4、MT5、cTrader 和 TradeLocker。市面上几乎所有杠杆档位都能在其中找到，从受监管的 1:30 上限，到离岸的 1:500 乃至更高。",[10,734,735,736,739],{},"结果并不随档位排序。在有交易历史的公开账户中，",[14,737,738],{"href":104},"最深回撤","的中位数为 9.7%，但 38.5% 的账户从未回撤超过 5%，38.2% 回吐了峰值的五分之一以上，17.6% 回吐了一半以上。",[10,741,742],{},"两组账户数量几乎一样，来自同一批杠杆设置，结局却差了一个数量级。把它们分开的，不是经纪商允许了什么。这些数字描述的是在 ShowMyTrades 上公开的账户，并不代表全体交易者；在这批账户里，真正起变化的变量是仓位规模。",[20,744,137],{"id":136},[10,746,747],{},"经纪商给的杠杆以徽章形式显示在账户头部，就在经纪商名称和账户货币旁边。这个数值取自账户数据，不是账户主人手动填写的。",[10,749,750,751,754,755,758,759,155,761,767,768,774,775,155,778,781],{},"交易者拿杠杆做了什么，要看别的地方。高级统计里的",[49,752,753],{},"总手数","是",[14,756,757],{"href":271},"实际交易的累计交易量","。账户统计面板里的",[49,760,225],{},[14,762,764],{"href":763},"\u002Fglossary\u002Fwhat-is-drawdown-on-balance",[49,765,766],{},"余额回撤","显示这些交易量在最糟糕时刻付出的代价：前者按包含持仓的净值计算，后者只按已平仓余额计算。图表模块里的",[14,769,771],{"href":770},"\u002Fglossary\u002Fwhat-is-an-equity-curve",[49,772,773],{},"净值曲线","把",[49,776,777],{},"余额",[49,779,780],{},"净值","画在一起，两条线之间的空隙，就是过大的持仓在平仓前藏身的地方。",[20,783,229],{"id":229},[231,785,786,792,798,804],{},[234,787,236,788,791],{},[49,789,790],{},"1:500 比 1:30 风险更大。","”账户设置不等于风险。两个账户持有完全相同的仓位，不管经纪商允许多高的杠杆，风险都完全一样。更高的杠杆只是让您能选错的仓位范围更大了。",[234,793,236,794,797],{},[49,795,796],{},"杠杆越高，赚得越多。","”杠杆高，只意味着被占作保证金的资金更少。盈亏由仓位大小和价格波动决定，两者都和杠杆档位无关。",[234,799,236,800,803],{},[49,801,802],{},"可用保证金是闲着的购买力。","”可用保证金是您离强制平仓还有多远。把它用掉，可控的亏损就是这样变成爆仓的。",[234,805,236,806,809],{},[49,807,808],{},"经纪商把我的杠杆限制在 1:30，所以我很安全。","”监管上限限制的是总仓位的最大规模，而不是任何一笔交易的风险。哪怕远在 1:30 的上限之内，一个仓位开得不当，照样能把账户亏光。",[10,811,812,813,44],{},"上面每个数字背后都有对应的计算器。想了解它们如何配合使用，请参阅",[14,814,816],{"href":815},"\u002Fguides\u002Fforex-calculators-guide","外汇计算器指南",{"title":36,"searchDepth":274,"depth":274,"links":818},[819,820,821,822,823],{"id":22,"depth":274,"text":22},{"id":59,"depth":274,"text":59},{"id":87,"depth":274,"text":87},{"id":136,"depth":274,"text":137},{"id":229,"depth":274,"text":229},"杠杆是仓位规模与背后所需资金之间的比率。本文给出保证金公式、计算示例，以及数千个账户的回撤数据。",{},"\u002Fzh\u002Fglossary\u002Fwhat-is-leverage",[828,292,829],"what-is-a-margin-call","what-is-maximum-drawdown",{"title":618,"description":824},"zh\u002F5.glossary\u002Fwhat-is-leverage","杠杆 (Leverage)","1K2tTcEclXFA4muLuzmkGaobTWxAuO3stEO8wQ-6IfM",{"id":835,"title":836,"body":837,"category":604,"date":282,"description":1026,"draft":284,"extension":285,"meta":1027,"navigation":287,"path":1028,"related":1029,"seo":1031,"stem":1032,"term":1033,"updated":282,"__hash__":1034},"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":838,"toc":1019},[839,845,847,850,856,859,865,873,876,896,903,906,908,911,918,920,932,935,948,950,957,982,984,1014],[10,840,841,842,844],{},"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",[14,843,225],{"href":466},"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。",[20,846,22],{"id":22},[10,848,849],{},"分三步，按顺序来。",[27,851,854],{"className":852,"code":853,"language":32},[30],"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",[34,855,853],{"__ignoreMap":36},[10,857,858],{},"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：",[27,860,863],{"className":861,"code":862,"language":32},[30],"风险金额   = 10,000 × 0.01      = $100\n仓位大小   = 100 \u002F (50 × 10)    = 0.20 手\n",[34,864,862],{"__ignoreMap":36},[10,866,867,868,872],{},"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",[14,869,871],{"href":870},"\u002Fglossary\u002Fwhat-is-a-stop-loss","止损","位置不能分开决定。",[10,874,875],{},"实践中最常用的有三种方法：",[231,877,878,884,890],{},[234,879,880,883],{},[49,881,882],{},"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。",[234,885,886,889],{},[49,887,888],{},"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。",[234,891,892,895],{},[49,893,894],{},"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。",[10,897,898],{},[899,900],"img",{"alt":901,"src":902},"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",[10,904,905],{},"一个账户用的是哪一种，光看手数就能看出来。",[20,907,59],{"id":59},[10,909,910],{},"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%。",[10,912,913,917],{},[14,914,916],{"href":915},"\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。",[20,919,87],{"id":87},[10,921,922,923,925,926,476,929,931],{},"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），",[49,924,113],{}," 曾",[14,927,928],{"href":104},"回撤超过 20%",[49,930,117],{}," 亏掉了峰值的一半以上。",[10,933,934],{},"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。",[10,936,937,938,476,940,943,944,947],{},"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 ",[49,939,483],{},[49,941,942],{},"53.9%"," 的账户自动化程度超过 90%，低于 10% 的则有 ",[49,945,946],{},"42.2%","。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。",[20,949,137],{"id":136},[10,951,952,956],{},[14,953,955],{"href":954},"\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。",[10,958,959,960,962,963,966,967,969,970,973,974,516,977,981],{},"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。",[49,961,201],{},"有一列",[49,964,965],{},"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。",[49,968,209],{},"提供",[49,971,972],{},"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。",[49,975,976],{},"高级统计",[14,978,979],{"href":271},[49,980,753],{},"给出汇总值；所有已发布账户累计交易了 1,724,575 手。",[20,983,229],{"id":229},[231,985,986,992,1002,1008],{},[234,987,236,988,991],{},[49,989,990],{},"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。",[234,993,236,994,997,998,1001],{},[49,995,996],{},"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。",[14,999,1000],{"href":16},"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。",[234,1003,236,1004,1007],{},[49,1005,1006],{},"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。",[234,1009,236,1010,1013],{},[49,1011,1012],{},"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。",[10,1015,1016,1017,44],{},"全套风险与仓位管理工具，以及各自适用的场景，请参阅",[14,1018,816],{"href":815},{"title":36,"searchDepth":274,"depth":274,"links":1020},[1021,1022,1023,1024,1025],{"id":22,"depth":274,"text":22},{"id":59,"depth":274,"text":59},{"id":87,"depth":274,"text":87},{"id":136,"depth":274,"text":137},{"id":229,"depth":274,"text":229},"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",[610,829,1030,293],"what-is-a-stop-loss",{"title":836,"description":1026},"zh\u002F5.glossary\u002Fwhat-is-position-sizing","仓位管理 (Position Sizing)","5hi2OmhGtFDhESyalituW-Nh51a9wIit5SKZxEMEQyk",{"id":1036,"title":1037,"body":1038,"category":1244,"date":282,"description":1245,"draft":284,"extension":285,"meta":1246,"navigation":287,"path":1247,"related":1248,"seo":1249,"stem":1250,"term":1251,"updated":297,"__hash__":1252},"glossary\u002Fzh\u002F5.glossary\u002Fwhat-is-volatility.md","什么是波动率？交易中如何衡量",{"type":7,"value":1039,"toc":1237},[1040,1043,1045,1048,1054,1057,1064,1066,1072,1078,1081,1083,1094,1111,1126,1128,1152,1179,1201,1203,1231],[10,1041,1042],{},"波动率是收益围绕自身平均值的离散程度：看的是单个结果离均值有多远，而不是朝哪个方向。它用标准差来衡量，而且是对称的：偶尔单日赚 8% 的策略，和偶尔单日亏 8% 的策略，波动率完全一样。所以波动率不等于风险。它是一个乘数，决定了某个仓位大小会变成多大的风险。",[20,1044,22],{"id":22},[10,1046,1047],{},"取一组收益（按日、按月或按笔），求出均值，再看各个结果离均值的典型距离。",[27,1049,1052],{"className":1050,"code":1051,"language":32},[30],"σ = √( Σ (rᵢ − r̄)² \u002F (n − 1) )\n\nrᵢ = 第 i 期的收益\nr̄  = 全部 n 期的平均收益\nn  = 期数\n",[34,1053,1051],{"__ignoreMap":36},[10,1055,1056],{},"标准差的单位和输入一致：输入百分比，得到百分比；输入每笔交易的盈亏，得到账户货币。要在不同时间周期之间比较，就得按时间的平方根换算，所以用日数据算年化数值时要乘以 √252，也就是一年的交易日数。",[10,1058,1059,1060,1063],{},"实践中有两个特性很重要。第一，波动率",[49,1061,1062],{},"不是恒定的","，它会扎堆：平静的几周后面往往还是平静的几周，剧烈的日子则接二连三。第二，它是所有风险调整指标的分母：夏普比率这一类指标，正是拿收益除以这个数字，所以两个增长相同的策略，评价可能天差地别。",[20,1065,59],{"id":59},[10,1067,1068,1069,1071],{},"离开波动率，仓位大小就没有意义。每天波动 40 点的货币对上的 1 手，和每天波动 140 点的货币对上的 1 手，标签相同，却是两笔完全不同的押注。按手数而不是按预期波动来定仓位，最容易让账户的风险状况在交易者什么都没决定的情况下悄悄走样，这也是",[14,1070,78],{"href":77},"规则按承担风险的金额来写、而不是按手数来写的原因。",[10,1073,1074,1075,1077],{},"波动率最终会变成",[14,1076,225],{"href":466},"。下跌的深度，大致等于波动率乘以风险敞口，再乘以判断错误的持续时间。把所交易品种的波动率减半，仓位不变，您要扛过的回撤也会减半，收益也跟着减半。",[10,1079,1080],{},"最后，波动率会在您不知不觉中改变。在平静的季度里仓位刚好合适的策略，波动率一翻倍就成了重仓，它不用做任何新决定，就已经变得危险。止损拉宽，点差拉宽，同样的手数，承担的风险已经是当初允许的两倍。",[20,1082,87],{"id":87},[10,1084,1085,1086,1089,1090,1093],{},"在 ShowMyTrades 上有交易历史的公开账户中（2026 年 8 月），",[49,1087,1088],{},"夏普比率","中位数为 ",[49,1091,1092],{},"0.05","。这样的收益和周围的噪音几乎分不出来：中间水平的账户，每期的平均结果和这些期间的离散程度比起来微不足道。",[10,1095,1096,1097,1100,1101,476,1103,1106,1107,1110],{},"分布的其他数字也是一样的结论：时间加权收益率中位数 ",[49,1098,1099],{},"+3.2%","，最深回撤中位数 ",[49,1102,109],{},[49,1104,1105],{},"171 笔已平仓交易","，持仓时间中位数 ",[49,1108,1109],{},"2.4 小时","。优势很小，承受的风险却是所得结果的好几倍。",[10,1112,1113,1114,1117,1118,1121,1122,1125],{},"离散程度还解释了这些账户为什么赢得多、最后却剩得少。胜率中位数是 ",[49,1115,1116],{},"68.8%","，盈利因子中位数是 ",[49,1119,1120],{},"1.28","：大多数交易以盈利平仓，而少数亏损单亏得够大，足以吃掉大部分收益。",[49,1123,1124],{},"63.0%"," 的账户长期为正，算是多数，但领先的幅度，比背后结果的离散程度薄得多。",[20,1127,137],{"id":136},[10,1129,1130,222,1132,1135,1136,155,1142,1144,1145,1151],{},[49,1131,976],{},[49,1133,1134],{},"表现指标","部分，",[14,1137,1139],{"href":1138},"\u002Fglossary\u002Fwhat-is-standard-deviation-in-trading",[49,1140,1141],{},"标准差",[49,1143,1088],{},"上下相邻。这里的标准差以账户货币显示，衡量的是单笔交易结果的离散程度：标准差很大，",[14,1146,1148],{"href":1147},"\u002Fglossary\u002Fwhat-is-expectancy",[49,1149,1150],{},"期望值","却很小，说明这个策略的结果很大程度上取决于样本里恰好有哪些交易。",[10,1153,1154,1157,1158,155,1160,1163,1164,1167,1168,1171,1172,1178],{},[49,1155,1156],{},"账户统计","面板在",[49,1159,543],{},[49,1161,1162],{},"绝对增长","下面两行显示",[49,1165,1166],{},"日均 %"," 和",[49,1169,1170],{},"月均 %","。这两个是平均值，所以要对照",[14,1173,1175],{"href":1174},"\u002Fguides\u002Fmonthly-returns-table-explained",[49,1176,1177],{},"月度收益","表来看，逐月的波动在那里一目了然：月均百分比相同的两个账户，每一行的数字可能天差地别。",[10,1180,1181,1183,1184,1187,1188,1190,1191,1193,1194,155,1197,1200],{},[49,1182,221],{},"是一个可以切换视图的面板：",[49,1185,1186],{},"逐笔交易增长","能清楚看出波动扎堆，",[49,1189,225],{},"则显示这些离散付出了多大代价。",[49,1192,976],{},"还提供",[49,1195,1196],{},"星期",[49,1198,1199],{},"小时","视图，可以看出结果集中在什么时候，而那通常也是波动集中的时候。",[20,1202,229],{"id":229},[231,1204,1205,1211,1217,1223],{},[234,1206,236,1207,1210],{},[49,1208,1209],{},"波动率就是风险。","”风险是遭受您承受不起的亏损的可能性。波动率是双向的离散程度，完全不涉及您的资金规模。",[234,1212,236,1213,1216],{},[49,1214,1215],{},"波动率低就安全。","”卖出尾部风险的策略（网格、马丁格尔、不对冲的套息交易），在那次注定让它们输掉的事件到来之前，波动率读数一直很低。",[234,1218,236,1219,1222],{},[49,1220,1221],{},"夏普比率高，交易者就更优秀。","”它只说明在所测样本里，每单位离散程度换来的收益更高。只有 171 笔交易，这个估计噪音很大。",[234,1224,236,1225,1228,1229,44],{},[49,1226,1227],{},"波动率是交易品种本身的属性。","”它既取决于品种，也取决于时间段。同一个货币对一年里会几次切换市场状态，含同一种货币的货币对还会一起变化。请参阅",[14,1230,316],{"href":315},[10,1232,1233,1234,44],{},"波动率只有变成真实亏损时才真正要紧，这部分该怎么看，请参阅",[14,1235,1236],{"href":717},"最大回撤详解",{"title":36,"searchDepth":274,"depth":274,"links":1238},[1239,1240,1241,1242,1243],{"id":22,"depth":274,"text":22},{"id":59,"depth":274,"text":59},{"id":87,"depth":274,"text":87},{"id":136,"depth":274,"text":137},{"id":229,"depth":274,"text":229},"指标","波动率是收益围绕平均值的离散程度，用标准差来衡量。它本身不是风险，但同样的手数意味着什么，由它决定。",{},"\u002Fzh\u002Fglossary\u002Fwhat-is-volatility",[610,292,829,609],{"title":1037,"description":1245},"zh\u002F5.glossary\u002Fwhat-is-volatility","波动率 (Volatility)","XCZ9I_SBNYpYPqBkPwcZX0S3YAivr3b6366b9-tzqnU",[1254,1255],{"path":315},{"path":288},1793898068501]