[{"data":1,"prerenderedAt":771},["ShallowReactive",2],{"\u002Fzh\u002Fguides\u002Fwin-rate-is-not-an-edge":3,"\u002Fzh\u002Fguides\u002Fwin-rate-is-not-an-edge-surround":756,"\u002Fzh\u002Fguides\u002Fwin-rate-is-not-an-edge-langs":767},{"id":4,"title":5,"authors":6,"badge":12,"body":14,"date":744,"description":745,"draft":746,"extension":747,"image":748,"meta":750,"navigation":751,"path":752,"seo":753,"stem":754,"updated":744,"__hash__":755},"posts\u002Fzh\u002F3.guides\u002F30.win-rate-is-not-an-edge.md","您的胜率在骗您：真实账户数据怎么说（{YEAR} 年版）",[7],{"name":8,"to":9,"avatar":10},"ShowMyTrades Team","https:\u002F\u002Fshowmytrades.com",{"src":11},"\u002Ficon-512.png",{"label":13},"指标",{"type":15,"value":16,"toc":731},"minimark",[17,21,24,39,42,47,50,57,63,86,89,93,100,103,113,116,121,128,131,136,147,150,157,161,170,270,277,284,291,305,309,312,319,322,338,353,361,365,368,374,379,406,413,417,420,526,537,567,584,591,595,636,640,646,652,658,664,674,680,684,711,714,724],[18,19,20],"p",{},"大部分单子都赚了，账户却几乎没涨。您可能怪过自己不够自律，怪过经纪商，也怪过市场。",[18,22,23],{},"其实问题出在算术上。",[18,25,26,30,31,34,35,38],{},[27,28,29],"strong",{},"ShowMyTrades 上有交易历史的公开账户","，胜率中位数是 ",[27,32,33],{},"68.8%","，盈利因子中位数是 ",[27,36,37],{},"1.28","。两个数字合起来看，就是这样一个账户：十笔大约赢七笔，到头来几乎什么都没剩下。",[18,40,41],{},"本指南就讲这个落差：高胜率和一条走平的净值曲线为什么并不矛盾，盈亏比至少要达到多少，胜率才有意义，以及到账户页面的哪里去查您自己的数据。",[43,44,46],"h2",{"id":45},"这两个数字不可能都是好消息","🎯 这两个数字，不可能都是好消息",[18,48,49],{},"68.8% 的胜率，听上去像高手。1.28 的盈利因子，却意味着每亏 $1 只净赚 $0.28。",[18,51,52],{},[53,54],"img",{"alt":55,"src":56},"在 ShowMyTrades 公开账户中，胜率中位数为 68.8%（大约十笔交易有七笔盈利平仓），旁边是盈利因子中位数 1.28（即每亏损 $1.00 盈利 $1.28）","\u002Fimages\u002Farticles\u002Fwin-rate-vs-payoff.svg",[18,58,59],{},[60,61,62],"em",{},"两个数字出自同一批账户，能决定账户涨不涨的，只有右边那个。",[18,64,65,66,69,70,73,74,77,78,81,82,85],{},"有一点需要说清楚。这是同一批账户的两个中位数，不是同一位交易者的两项数据：胜率排在中间的账户，盈利因子不一定也排在中间。它们反映的是 ShowMyTrades 上已发布账户的整体",[60,67,68],{},"轮廓","，而且各项指标指向一致。时间加权收益率中位数：",[27,71,72],{},"+3.2%","。夏普比率中位数：",[27,75,76],{},"0.05","。最深回撤中位数：",[27,79,80],{},"9.7%","。另外，有 ",[27,83,84],{},"63.0%"," 的账户时间加权收益为正。所以这不是一堆爆仓账户，而是一堆差一口气的账户。",[18,87,88],{},"胜率很高，风险调整后收益却接近零。这种组合只有一种解释，跟运气无关。",[43,90,92],{"id":91},"没人去算的那道算术","➗ 没人去算的那道算术",[18,94,95,96,99],{},"盈利因子，就是所有盈利单赚的钱之和，除以所有亏损单亏的钱之和。推公式之前先交代一个细节：在 ShowMyTrades 上，高级统计里所有以货币计的数值（盈利因子背后的两个总和，以及平均盈利、平均亏损和期望值），都是单笔盈亏",[27,97,98],{},"加上库存费和佣金","之后的结果；这个净额大于零，才算一笔盈利单。经纪商的那份已经扣在里面了。",[18,101,102],{},"接下来把分子和分母分别拆开：",[104,105,106,110],"ul",{},[107,108,109],"li",{},"总盈利 = 盈利单数量 × 平均盈利",[107,111,112],{},"总亏损 = 亏损单数量 × 平均亏损",[18,114,115],{},"于是：",[18,117,118],{},[27,119,120],{},"盈利因子 =（盈利单数 ÷ 亏损单数）×（平均盈利 ÷ 平均亏损）",[18,122,123,124,127],{},"这两项相互独立。第一项是把胜率换成赔率；第二项是您的",[27,125,126],{},"盈亏比","，也就是一笔典型盈利单的金额，相当于一笔典型亏损单的多少。单看哪一项都不算优势，两项相乘才是。",[18,129,130],{},"把我们的中位数代进去。68.8% 的胜率，相当于每亏一笔赢 2.21 笔。要让盈利因子刚好是 1.28，盈亏比只能是：",[18,132,133],{},[27,134,135],{},"1.28 ÷ 2.21 = 0.58",[18,137,138,139,142,143,146],{},"也就是说，平均盈利只有",[27,140,141],{},"平均亏损的 58% 左右","。反过来看，一笔典型的亏损约等于一笔典型盈利的 ",[27,144,145],{},"1.7 倍","。",[18,148,149],{},"这还谈不上心理问题，是结构问题。大多数已发布的账户赢的次数远多于输的次数，赚来的钱却在输的那几笔上又吐了回去。",[18,151,152,153,156],{},"期望值能把这笔账算得更具体。我们的提示框是这样定义它的：（胜率 × 平均盈利）−（亏损率 × 平均亏损）。胜率 68.8%、盈亏比 0.58 时，平均每笔交易的收益约为",[27,154,155],{},"一笔平均亏损的 8.7%","。换句话说，一笔典型亏损单亏 $100，一笔典型交易只赚回 $8.70。这已经是净额，最终进入余额的就是这么多。",[43,158,160],{"id":159},"盈亏平衡胜率表","📋 盈亏平衡胜率表",[18,162,163,164,169],{},"评价一个胜率之前，先要知道它得过哪道线。盈亏平衡胜率 = 1 ÷（1 + ",[165,166,168],"a",{"href":167},"\u002Fglossary\u002Fwhat-is-risk-reward-ratio","回报风险比","）；胜率低于中间那一列，系统就是亏钱的，感觉再好也没用。",[171,172,173,189],"table",{},[174,175,176],"thead",{},[177,178,179,183,186],"tr",{},[180,181,182],"th",{},"回报风险比（平均盈利 ÷ 平均亏损）",[180,184,185],{},"刚好保本所需的胜率",[180,187,188],{},"68.8% 胜率对应的盈利因子",[190,191,192,204,215,226,237,248,259],"tbody",{},[177,193,194,198,201],{},[195,196,197],"td",{},"0.25 : 1",[195,199,200],{},"80.0%",[195,202,203],{},"0.55",[177,205,206,209,212],{},[195,207,208],{},"0.5 : 1",[195,210,211],{},"66.7%",[195,213,214],{},"1.10",[177,216,217,220,223],{},[195,218,219],{},"0.75 : 1",[195,221,222],{},"57.1%",[195,224,225],{},"1.65",[177,227,228,231,234],{},[195,229,230],{},"1 : 1",[195,232,233],{},"50.0%",[195,235,236],{},"2.21",[177,238,239,242,245],{},[195,240,241],{},"1.5 : 1",[195,243,244],{},"40.0%",[195,246,247],{},"3.31",[177,249,250,253,256],{},[195,251,252],{},"2 : 1",[195,254,255],{},"33.3%",[195,257,258],{},"4.41",[177,260,261,264,267],{},[195,262,263],{},"3 : 1",[195,265,266],{},"25.0%",[195,268,269],{},"6.62",[18,271,272,273,276],{},"最右边一列值得细看。68.8% 的胜率确实非常出色，",[60,274,275],{},"前提是","盈亏比过得去。盈亏比 1:1 时，盈利因子是 2.21，已经是一套像样的策略；到了 2:1，就是相当罕见的好策略了。",[18,278,279,280,283],{},"我们统计的这批账户，盈利因子在 1.28，倒推出来的盈亏比略低于 0.6，落在第二行和第三行之间。再代回盈亏平衡公式：1 ÷（1 + 0.58）= ",[27,281,282],{},"63.3%","。也就是说，处在中位数的已发布账户，胜率要有 63.3% 才能保本。它的实际胜率是 68.8%。全部的利润，就来自这五个半百分点。",[18,285,286,287,290],{},"这张表还说明了另外两件事。人人都看不上的胜率 40% 的趋势跟踪者，在 1.5:1 时正好保本，到 2:1 时盈利因子是 1.33。反过来，90% 的胜率，只要盈亏比跌到大约 0.11 以下，就会变成",[60,288,289],{},"亏钱","的系统。每次为了不认错而把止损放宽，您做的就是这种交换。",[18,292,293,296,297,300,301,304],{},[27,294,295],{},"往公式里代什么数字，也有讲究","。用",[27,298,299],{},"货币","计的平均盈利和平均亏损，算出来的结果是实打实的，因为它们已经扣掉了佣金和库存费。用",[27,302,303],{},"点","计的（同一行的另一半），算出来的是扣成本之前的结果，而这个结果永远更好看。",[43,306,308],{"id":307},"为什么盈亏比最后会颠倒过来","🪞 为什么盈亏比最后会颠倒过来",[18,310,311],{},"没有人会刻意设计一个亏损是盈利 1.7 倍的系统。它是被一个个看似合理的决定慢慢拼出来的。",[18,313,314,315,318],{},"单子有了浮盈：平仓就能把不确定的盈利落袋为安，而这种确定感很像风险管理。单子有了浮亏：一平仓，账面亏损就成了真亏损，于是选择再等等；等待偶尔还真管用，这个习惯也就改不掉了。行为金融学把这种不对称叫作",[27,316,317],{},"处置效应","：赚了急着走，亏了舍不得割。",[18,320,321],{},"结果几乎是注定的。盈利单总是在分布的小端就被截断，亏损单的尾部却一点不少。胜率涨了（没平仓的亏损还不算亏损），盈亏比则悄悄地倒了过来。",[18,323,324,325,328,329,332,333,337],{},"同一次统计里还有两个数字可以佐证。我们公开账户的持仓时间中位数是 ",[27,326,327],{},"2.4 小时","；有库存费记录的账户中，",[27,330,331],{},"86.3% 的库存费净额为负","：",[165,334,336],{"href":335},"\u002Fguides\u002Fswap-and-carry-trade","仓位过夜经过展期","，钱就在不知不觉中流走了。",[18,339,340,341,344,345,348,349,352],{},"这也不只是人性的毛病。我们公开账户的自动交易占比中位数是 ",[27,342,343],{},"99%","，其中 ",[27,346,347],{},"53.9% 的账户自动化程度超过 90%","，",[27,350,351],{},"42.2% 低于 10%","。代码要是这么写，也会这么做：止盈设得很近，止损设得很远甚至干脆不设，就能以机器的速度刷出漂亮的胜率，盈亏比也正是上面这种形状。",[18,354,355,356,360],{},"（我们从不猜测某个账户上跑的是哪款商业机器人，也不发布任何按 EA 划分的数据。自动交易占比本身识别不了任何东西：它只是带",[165,357,359],{"href":358},"\u002Fguides\u002Fmagic-numbers-and-strategy-labels","魔术号","（在 cTrader 上则是机器人标签）的交易所占的比例，这些标识代表什么，只有账户所有者本人知道。）",[43,362,364],{"id":363},"成本已经算在这些数字里了","💸 成本已经算在这些数字里了",[18,366,367],{},"一个只相当于平均亏损 8.7% 的优势本来就很薄，很多人第一反应会补一句：“经纪商还要再抽一笔呢。”在 ShowMyTrades 上，这笔减法已经做过了：1.28 是扣掉佣金和库存费之后剩下的，不是策略付费之前的成绩。这件事要两面看：一方面，这批账户并没有扣费前的比较所显示的那么弱；另一方面，同样一批交易，回测或供应商报出的扣费前数字本来就要高出一截。",[18,369,370],{},[53,371],{"alt":372,"src":373},"从毛利润开始，依次扣除点差、佣金和库存费，得到净额的瀑布图，显示总优势中有多少从未进入账户","\u002Fimages\u002Farticles\u002Fgross-to-net.svg",[18,375,376],{},[60,377,378],{},"ShowMyTrades 页面上的每一项统计，量的都是这根柱子的右端；业绩单报的通常是左端。",[18,380,381,382,385,386,389,390,393,394,397,398,401,402,405],{},"即便如此，页面还是把两头都摆了出来：",[27,383,384],{},"佣金合计","和",[27,387,388],{},"已付库存费合计","以绝对金额列在“交易”一栏里。我们平台上所有账户加起来，这两项分别达到 ",[27,391,392],{},"$4,782,670"," 和 ",[27,395,396],{},"$862,547","；仅公开账户就交易了 ",[27,399,400],{},"1,724,575 手","，累计同步的交易共 ",[27,403,404],{},"15,436,464 笔","。这几个数的分母不同，不要拿来互相相除。它们说明的是：到了这个规模，这笔费用足以左右结果。",[18,407,408,409,412],{},"顺带还能得到一个免费的诊断方法。",[27,410,411],{},"期望值","在同一行里显示两次：一次以点计，一次以货币计。点数只反映价格变动，看不到成本；货币是净额。如果点数为正、货币为负，说明策略看对了行情，赚到的却全交给了经纪商。",[43,414,416],{"id":415},"如何在您自己的账户上检查","🔍 如何在您自己的账户上检查",[18,418,419],{},"这些数字在任何一个 ShowMyTrades 账户页面上都能读到，一分钟就够，用不着电子表格。",[171,421,422,435],{},[174,423,424],{},[177,425,426,429,432],{},[180,427,428],{},"看什么",[180,430,431],{},"在哪里",[180,433,434],{},"怎么判断",[190,436,437,450,469,482,493,512],{},[177,438,439,444,447],{},[195,440,441],{},[27,442,443],{},"胜率",[195,445,446],{},"高级统计 → 交易",[195,448,449],{},"千万别单独看",[177,451,452,460,466],{},[195,453,454,385,457],{},[27,455,456],{},"平均盈利",[27,458,459],{},"平均亏损",[195,461,462,463,465],{},"高级统计 → 交易（点",[60,464,385],{},"货币）",[195,467,468],{},"用货币值相除，得到的就是您的净盈亏比",[177,470,471,476,479],{},[195,472,473],{},[27,474,475],{},"盈利因子",[195,477,478],{},"高级统计 → 表现指标",[195,480,481],{},"低于 1.0，按定义就是亏钱",[177,483,484,488,490],{},[195,485,486],{},[27,487,411],{},[195,489,478],{},[195,491,492],{},"以点和货币显示。货币值必须为正",[177,494,495,503,505],{},[195,496,497],{},[27,498,499],{},[165,500,502],{"href":501},"\u002Fglossary\u002Fwhat-is-average-trade-length","平均持仓时间",[195,504,446],{},[195,506,507,508,511],{},"和您",[60,509,510],{},"自以为","的持仓时间对比",[177,513,514,521,523],{},[195,515,516,518,519],{},[27,517,384],{}," \u002F ",[27,520,388],{},[195,522,446],{},[195,524,525],{},"盈亏比里已经扣掉的部分",[18,527,528,529,532,533,536],{},"接下来是能让处置效应现形的视图：点一下",[27,530,531],{},"持续时间","切换按钮，就会画出",[60,534,535],{},"持仓时间与盈利能力","图。每笔已平仓交易是一个点，横轴是持仓时长，纵轴是净盈亏，气泡大小代表交易量。如果红点总是落在绿点右边，说明亏损单您拿得比盈利单久，而且证据就摆在眼前。",[18,538,539,540,543,544,547,548,547,551,547,554,547,557,547,560,563,564,566],{},"想单独分析，就打开",[27,541,542],{},"自定义分析","。按",[27,545,546],{},"交易持续时间（秒）","、",[27,549,550],{},"品种",[27,552,553],{},"手数",[27,555,556],{},"盈亏范围",[27,558,559],{},"交易时段",[27,561,562],{},"星期","或",[27,565,359],{},"（您在自己账户上设的标签）筛选，所有统计都会只按筛出来的交易重新计算。拿持仓不到一小时的交易和持仓超过一天的交易比一比，两组数据很少像是同一套策略。",[18,568,569,570,574,575,579,580,146],{},"最后用两个计算器收尾：",[165,571,573],{"href":572},"\u002Ftools\u002Frisk-of-ruin-calculator","破产风险计算器","根据胜率、平均盈利、平均亏损和单笔风险，算出亏光本金的概率；",[165,576,578],{"href":577},"\u002Ftools\u002Fposition-size-calculator","仓位计算器","帮您把平均亏损控制在计划之内。全部工具见",[165,581,583],{"href":582},"\u002Fguides\u002Fforex-calculators-guide","我们的计算器指南",[18,585,586,587,146],{},"页面上的其他内容（增长率与绝对增长率、净值回撤与余额回撤、Z 值、AHPR 和 GHPR），请参阅",[165,588,590],{"href":589},"\u002Fguides\u002Freading-a-trading-account-dashboard","如何读懂交易账户控制台",[43,592,594],{"id":593},"️-这个数字被误用的五种方式","⚠️ 这个数字被误用的五种方式",[596,597,598,609,615,621,630],"ol",{},[107,599,600,603,604,608],{},[27,601,602],{},"只报胜率，不报盈亏比","。“85% 胜率”算不上业绩，它只是两项乘积中的一项。请对方拿出平均盈利、平均亏损和盈利因子，拿不出来就当它是摆设。",[165,605,607],{"href":606},"\u002Fguides\u002Fverifying-trading-performance-claims","验证交易业绩宣传","里讲的也是这个习惯。",[107,610,611,614],{},[27,612,613],{},"刻意把胜率做高","。放宽止损、不设止损或摊平加仓，确实能把这个百分比拉上去。代价是亏损被推到尾部、压进持仓里，最后显示为净值回撤和余额回撤之间的差距，而不是某个月的亏损。",[107,616,617,620],{},[27,618,619],{},"拿不同类型策略的胜率比高低","。胜率 75% 的剥头皮交易者和胜率 45% 的波段交易者，不能靠这个数字排名次。谁更强，要看回报风险比。",[107,622,623,629],{},[27,624,625],{},[165,626,628],{"href":627},"\u002Fguides\u002Fhow-much-history-a-track-record-needs","交易太少就下结论","。我们公开账户的历史中位数是 171 笔已平仓交易，只看其中三十笔算出的胜率，不过是装成统计的噪音。盈亏比需要的历史还要更长，因为一笔特别大的亏损就能拉动平均亏损，其他数字却纹丝不动。",[107,631,632,635],{},[27,633,634],{},"拿扣成本后的盈利因子和扣成本前的比","。我们的数字是净额：佣金和库存费都算在两个总和里，一笔不赚不亏、却付了 $7 佣金的交易会记为亏损单。回测或供应商的业绩单报的通常是扣费前的版本。两者根本不能放在一起比。",[43,637,639],{"id":638},"常见问题","❓ 常见问题",[18,641,642,645],{},[27,643,644],{},"高胜率有没有可能是真优势？","\n有，前提是盈亏比撑得住。68.8% 的胜率配上 1:1 的盈亏比，盈利因子是 2.21，是一套很强的策略。毛病从来不在胜率本身，而在于这个胜率是靠“亏的比赚的多”换来的。",[18,647,648,651],{},[27,649,650],{},"盈利因子到底要多高才算好？","\n账户页面会用颜色标出来：低于 1.0 是红色，1.0 到 2.0 是琥珀色，2.0 及以上是绿色。我们这批账户的中位数是 1.28（琥珀色），只看已平仓交易在 100 笔及以上的账户则是 1.29，可见样本一旦够大，这个数字几乎不变。如果历史很长、盈利因子却在 3.0 以上，就打开交易列表看看：往往是靠几笔不会再来的巨额盈利撑起来的。",[18,653,654,657],{},[27,655,656],{},"用了止损以后胜率反而降了，是我变差了吗？","\n几乎可以肯定恰恰相反：给亏损设了上限，胜率本来就该下降。判断这个变化好不好，要看期望值和盈利因子，而不是那个百分比。",[18,659,660,663],{},[27,661,662],{},"我大部分单子都赚，为什么夏普比率这么低？","\n因为夏普比率衡量的是收益相对波动的大小，而一笔偶发的巨额亏损本身就是波动。我们公开账户的夏普比率中位数是 0.05：一连串小赚之间夹着几次大亏，正是高胜率配上近乎为零的夏普比率的典型形态。",[18,665,666,669,670,146],{},[27,667,668],{},"别人账户上的平均盈利和平均亏损我能看到吗？","\n能，只要所有者公开了：高级统计是公开页面的一部分，字段和您在自己账户上看到的完全一样。至于这些数据是不是来自经纪商，是另一回事，请参阅",[165,671,673],{"href":672},"\u002Fguides\u002Fverified-vs-unverified-track-records","经验证与未经验证的业绩记录",[18,675,676,679],{},[27,677,678],{},"自动交易账户的情况会不一样吗？","\n一点都不会。无论是 EA 还是主观交易者，算术完全相同；机器人只是把同样失衡的盈亏比执行得更稳定。",[43,681,683],{"id":682},"相关指南","🔗 相关指南",[104,685,686,691,698,704],{},[107,687,688,690],{},[165,689,590],{"href":589},"：页面上的每一项指标",[107,692,693,697],{},[165,694,696],{"href":695},"\u002Fguides\u002Ftracking-trading-performance","交易表现追踪","：长期要盯住哪些数据",[107,699,700,703],{},[165,701,702],{"href":606},"如何验证交易业绩宣传","：付钱给任何人之前的检查清单",[107,705,706,710],{},[165,707,709],{"href":708},"\u002Fguides\u002Fhow-to-verify-your-account","如何验证您的交易账户","：让您自己的记录用上来自经纪商的数据",[712,713],"hr",{},[18,715,716,146,719,723],{},[27,717,718],{},"去看看吧",[165,720,722],{"href":721},"\u002Fexplore","探索","里是大家刚发布的账户，每张卡片上都有平台、增长率、回撤和余额。随便点进一个账户，胜率、平均盈利、平均亏损、期望值和盈利因子就在同一屏上，因为其中任何一个离开其他几个，都说明不了什么。",[18,725,726,730],{},[165,727,729],{"href":728},"\u002Fsignup","免费连接您的账户","（MT4、MT5、cTrader 或 TradeLocker），您自己的盈亏比就不用再靠猜了。",{"title":732,"searchDepth":733,"depth":733,"links":734},"",2,[735,736,737,738,739,740,741,742,743],{"id":45,"depth":733,"text":46},{"id":91,"depth":733,"text":92},{"id":159,"depth":733,"text":160},{"id":307,"depth":733,"text":308},{"id":363,"depth":733,"text":364},{"id":415,"depth":733,"text":416},{"id":593,"depth":733,"text":594},{"id":638,"depth":733,"text":639},{"id":682,"depth":733,"text":683},"2026-08-19T00:00:00.000Z","公开账户的胜率中位数是 68.8%，盈利因子中位数却只有 1.28。这中间差的是一道算术题，本文讲清楚它，并教您检查自己的账户。",false,"md",{"src":749},"\u002Fimages\u002Fblog\u002Fwin-rate-is-not-an-edge-hero.svg",{},true,"\u002Fzh\u002Fguides\u002Fwin-rate-is-not-an-edge",{"title":5,"description":745},"zh\u002F3.guides\u002F30.win-rate-is-not-an-edge","EBPIPpne5h57BiuT_9iLbq1KlrAK9_OM_35uMkt4paI",[757,762],{"title":758,"path":759,"stem":760,"description":761,"children":-1},"买 EA 之前，如何识破虚假业绩（{YEAR} 年版）","\u002Fzh\u002Fguides\u002Fspotting-fake-ea-results","zh\u002F3.guides\u002F29.spotting-fake-ea-results","EA 销售页面给您看的是回测，不是业绩记录。本文教您看懂两者之间的落差：曲线拟合、tick 质量、手数递增、账户重置，外加一份买家十分钟自查清单。",{"title":763,"path":764,"stem":765,"description":766,"children":-1},"业绩记录要多长才有意义？（{YEAR} 年版）","\u002Fzh\u002Fguides\u002Fhow-much-history-a-track-record-needs","zh\u002F3.guides\u002F31.how-much-history-a-track-record-needs","连赚三个月还算不上业绩记录。本文讲样本量的算术、最大回撤为什么只增不减，以及相信一条曲线之前应守住的最低标准。",[768,770],{"path":769},"\u002Fguides\u002Fwin-rate-is-not-an-edge",{"path":752},1793898037921]