[{"data":1,"prerenderedAt":846},["ShallowReactive",2],{"\u002Fzh\u002Fguides\u002Fspotting-fake-ea-results":3,"\u002Fzh\u002Fguides\u002Fspotting-fake-ea-results-surround":831,"\u002Fzh\u002Fguides\u002Fspotting-fake-ea-results-langs":842},{"id":4,"title":5,"authors":6,"badge":12,"body":14,"date":819,"description":820,"draft":821,"extension":822,"image":823,"meta":825,"navigation":826,"path":827,"seo":828,"stem":829,"updated":819,"__hash__":830},"posts\u002Fzh\u002F3.guides\u002F29.spotting-fake-ea-results.md","买 EA 之前，如何识破虚假业绩（{YEAR} 年版）",[7],{"name":8,"to":9,"avatar":10},"ShowMyTrades Team","https:\u002F\u002Fshowmytrades.com",{"src":11},"\u002Ficon-512.png",{"label":13},"尽职调查",{"type":15,"value":16,"toc":805},"minimark",[17,22,36,44,51,54,58,61,68,71,141,145,152,155,183,186,190,193,196,207,214,220,232,236,239,242,248,253,256,283,286,324,331,334,340,345,349,352,362,371,377,383,390,397,404,408,411,504,520,524,527,621,628,631,697,700,704,710,724,735,741,752,756,787,790],[18,19,21],"h2",{"id":20},"您买的不是一份回测","🤖 您买的不是一份回测",[23,24,25,30,31,35],"p",{},[26,27,29],"a",{"href":28},"\u002Fglossary\u002Fwhat-is-an-expert-advisor","EA（智能交易系统）","的销售页面一般就三样东西：一份策略测试器报告、一张增长图表、一个价格。三样里没有一样算得上",[26,32,34],{"href":33},"\u002Fglossary\u002Fwhat-is-a-track-record","业绩记录","。",[23,37,38,39,43],{},"要问的不是“这是不是骗局”，而是更具体的一点：",[40,41,42],"strong",{},"这个机器人放到实盘账户上","、放在我的经纪商那里，表现会不会和页面说的一样？大多数 EA 都过不了这一关，卖家却一句谎话都不用说：那些数字产生的条件，您根本复现不了。",[23,45,46,47,35],{},"本指南只讲 EA 营销特有的套路：不可能真实发生的回测，为一段不会重演的过去调出来的参数，藏在平滑曲线下面的手数递增，还有从十二个账户里挑出来的那个赚钱账户。信号和代管账户的问题，请看",[26,48,50],{"href":49},"\u002Fguides\u002Fverifying-trading-performance-claims","如何验证交易业绩宣传",[23,52,53],{},"下文所有基准数据都取自 2026 年 8 月的 ShowMyTrades 数据库：1 万+ 个已连接账户，1,540 万笔已同步交易，所有中位数都只统计有交易历史的公开账户。这些数据反映的是在本站公开的账户，不代表全体交易者；但它们真实、与经纪商同步，足以判断某个宣传是否已经超出正常分布。",[18,55,57],{"id":56},"回测与实盘您买到的其实是两者的差距","📉 回测与实盘：您买到的，其实是两者的差距",[23,59,60],{},"回测只是假设，实盘账户才是证据。两者之间的落差最能说明一个 EA 的真实水平，也正是销售页面最想模糊掉的地方。",[23,62,63,64,67],{},"大部分问题看三个日期就够了：这个版本的 EA 什么时候发布，实盘账户什么时候开始交易，回测区间到哪天结束。只有该版本发布",[40,65,66],{},"之后","的实盘交易才有参考价值。在那之前的一切，都是拿策略去拟合一段它早就见过的历史。",[23,69,70],{},"然后确认两边是同一个实验：品种、时间周期、set 文件、账户货币、经纪商类型都要一致。卖家常常拿一个品种的回测，配一个同时做四个品种的实盘账户。",[72,73,74,90],"table",{},[75,76,77],"thead",{},[78,79,80,84,87],"tr",{},[81,82,83],"th",{},"卖家给您看的",[81,85,86],{},"该向他要什么",[81,88,89],{},"怎样才算数",[91,92,93,105,116,130],"tbody",{},[78,94,95,99,102],{},[96,97,98],"td",{},"策略测试器报告",[96,100,101],{},"与经纪商连接的实盘账户，用同一个 set 文件",[96,103,104],{},"在所售版本发布之后才开始的实盘交易",[78,106,107,110,113],{},[96,108,109],{},"一张终端截图",[96,111,112],{},"带连续同步历史的公开网址",[96,114,115],{},"截图只是图片，不是记录",[78,117,118,121,124],{},[96,119,120],{},"一个“已验证业绩”链接",[96,122,123],{},"数据是不是从经纪商只读获取",[96,125,126],{},[26,127,129],{"href":128},"\u002Fguides\u002Fverified-vs-unverified-track-records","已验证与未验证的区别",[78,131,132,135,138],{},[96,133,134],{},"六周的实盘结果",[96,136,137],{},"十二个月，最差的那个月也要在里面",[96,139,140],{},"策略实实在在扛过的一轮回撤",[18,142,144],{"id":143},"️-曲线拟合参数早就知道答案","🎛️ 曲线拟合：参数早就知道答案",[23,146,147,148,35],{},"给优化器二十个输入参数、五年历史数据，它总能找到一组赚过钱的组合。这算不上发现，只是算术：这样的结果",[26,149,151],{"href":150},"\u002Fglossary\u002Fwhat-is-curve-fitting","把过去描述得完美无缺，对未来却毫无预测力",[23,153,154],{},"识别起来有几个很实在的信号：",[156,157,158,165,171,177],"ul",{},[159,160,161,164],"li",{},[40,162,163],{},"输入参数太多","。真正稳健的思路，四五个参数就立得住。二十个参数是在搜索，不是策略。",[159,166,167,170],{},[40,168,169],{},"数值精确得蹊跷","。止盈 37 点，止损 113 点，过滤值 62.5。靠推理得出的数字往往接近整数，这种数字是网格搜索搜出来的。",[159,172,173,176],{},[40,174,175],{},"测试区间正好止于优化结束的地方","。这样的报告，不过是一场已经赛完的比赛的记分牌。",[159,178,179,182],{},[40,180,181],{},"一碰就碎","。请卖家换一年再跑同样的回测，再把每个参数各调百分之十。稳健的系统成绩会下滑，拟合出来的系统会直接崩掉。",[23,184,185],{},"如果卖家说不清哪段数据用来找参数、哪段数据从头到尾没碰过，那就等于根本没做样本外测试。",[18,187,189],{"id":188},"tick-问题99-建模质量到底是什么意思","🧪 Tick 问题：“99% 建模质量”到底是什么意思",[23,191,192],{},"MT4 策略测试器里的建模质量，说的是测试器怎样在 K 线之间生成 tick：用更小的时间周期插值出来。它并不说明这些 tick 和您的经纪商实际报出的价格有多像。历史数据本身质量差，99% 的报告也只是把错误的价格算得 99% 精确。",[23,194,195],{},"更值得看的，是默认回测悄悄替您做了哪些假设：点差固定，没有佣金，没有库存费，没有滑点，没有重新报价，没有部分成交，新闻前后点差不扩大，成交即时完成。每一条都偏向您，也都偏离现实。",[23,197,198,199,202,203,206],{},"成本可不是零头。本站公开账户的交易者在 1,724,575 手交易上共支付了 ",[40,200,201],{},"$4,782,670 的佣金和 $862,547 的库存费","；在有库存费的账户里，",[40,204,205],{},"86.3% 是净支付库存费","，融资成本月复一月都对他们不利。一个剥头皮 EA 按零佣金回测，和按您经纪商的佣金实际运行，是两款不同的产品，而赚钱的只有一款。",[23,208,209],{},[210,211],"img",{"alt":212,"src":213},"价格横盘不动，账户余额却从周一到周五每晚往下掉一截，周三扣三倍；有库存费的账户中 86.3% 是在付库存费，而不是赚库存费","\u002Fimages\u002Farticles\u002Fswap-bleed.svg",[23,215,216],{},[217,218,219],"em",{},"默认回测把这条线设成零，可不管仓位最后做对还是做错，这笔钱都照扣不误。",[23,221,222,223,227,228,35],{},"MT5 用经纪商真实 tick 历史回测，确实是实打实的进步，但那终究只是一家经纪商的历史，也模拟不了您自己的成交。凡是固定点差、零成本的回测，都当营销素材看待，差额请自己算：用",[26,224,226],{"href":225},"\u002Ftools","交易计算器","，配合",[26,229,231],{"href":230},"\u002Fguides\u002Fforex-calculators-guide","计算器使用指南",[18,233,235],{"id":234},"马丁格尔的特征就在手数里","🎰 马丁格尔的特征就在手数里",[23,237,238],{},"靠净值曲线判断马丁格尔，是慢办法；看手数，是快办法。",[23,240,241],{},"在自定义分析里把账户筛选到单一品种，再顺着交易表格的交易量（手数）一列往下看：0.01、0.02、0.04、0.08，每一单都是在上一单还浮亏时开的。页面上叫它“智能恢复”也好，“网格平衡”“自适应敞口”也罢，本质都是逆势加仓摊平。",[23,243,244],{},[210,245],{"alt":246,"src":247},"两种仓位方式对比：一种每亏一次就把手数翻倍，直到某次亏损再也补不回来；另一种按净值的固定比例下单，挺过了同样的连续亏损","\u002Fimages\u002Farticles\u002Flot-progression.svg",[23,249,250],{},[217,251,252],{},"胜率一样，市场一样：两个账户的分别就在这一列里，一分钟都用不了就能看完。",[23,254,255],{},"另外还有四项佐证，几分钟就能全部看完：",[257,258,259,265,271,277],"ol",{},[159,260,261,264],{},[40,262,263],{},"净值与余额","。余额图表会把两条线画在一起。两线之间长期有缺口，说明亏损单一直扛着没平。",[159,266,267,270],{},[40,268,269],{},"回撤与余额回撤","。两项都在账户统计面板里。前者远大于后者，说明亏损藏在持仓里。",[159,272,273,276],{},[40,274,275],{},"Z 值","。平台把它当作随机性指标：接近零，说明盈亏随机出现；不论正负，离零越远，说明盈亏越是成串出现。网格在解套之前，会先拉出很长的连续亏损。",[159,278,279,282],{},[40,280,281],{},"持仓标签页","。同一品种上同时挂着几十个仓位，一片红，就是一个跑到一半的网格。",[23,284,285],{},"再把卖家的风险宣传和现实对照一下。全部公开账户的情况如下：",[72,287,288,298],{},[75,289,290],{},[78,291,292,295],{},[81,293,294],{},"曾达到的最深回撤",[81,296,297],{},"账户占比",[91,299,300,308,316],{},[78,301,302,305],{},[96,303,304],{},"低于 5%",[96,306,307],{},"38.5%",[78,309,310,313],{},[96,311,312],{},"超过 20%",[96,314,315],{},"38.2%",[78,317,318,321],{},[96,319,320],{},"超过 50%",[96,322,323],{},"17.6%",[23,325,326,327,330],{},"回撤中位数：",[40,328,329],{},"9.7%","。将近五分之一的公开账户曾经亏掉一半以上，而这些账户的自动化程度很高：自动交易占比中位数为 99%，53.9% 的账户自动化程度超过 90%。机器人并不比人安全，只是比人快。",[23,332,333],{},"第一行要细看：38.5% 的账户从来没有回撤超过 5%，所以光是回撤小，什么也证明不了。真正少见的，是回撤小、收益又高。一个号称年化 200%、最大回撤 5% 的 EA，等于说自己落在这个分布里几乎没人的那个角落。这不是不可能，但得拿出证据。",[23,335,336],{},[210,337],{"alt":338,"src":339},"ShowMyTrades 公开账户曾达到的最深回撤：38.5% 低于 5%，23.3% 在 5% 到 20% 之间，20.6% 在 20% 到 50% 之间，17.6% 超过 50%，中位数 9.7%","\u002Fimages\u002Farticles\u002Fdrawdown-distribution.svg",[23,341,342],{},[217,343,344],{},"看任何销售页面之前，先把这张分布图记在脑子里：卖家要您相信的，就是他们不属于这个群体。",[18,346,348],{"id":347},"模拟账户美分账户重置以及十二个账户里最好的那个","🎭 模拟账户、美分账户、重置，以及十二个账户里最好的那个",[23,350,351],{},"普通的业绩，换上四套“戏服”就能变成销售页面，而且一个数字都不用编。",[23,353,354,35,357,361],{},[40,355,356],{},"模拟账户冒充实盘",[26,358,360],{"href":359},"\u002Fglossary\u002Fwhat-is-a-demo-vs-live-account","模拟账户的成交很大方","，偏偏大方在实盘最吃亏的地方：开盘时、新闻期间、止损被跳空越过时。完美的模拟曲线只是模拟出了优势，证明不了优势。在 ShowMyTrades 上，账户标为真实账户还是模拟账户，依据的是经纪商报告的信息（账户的交易模式加上所连接的服务器），发布者自己选不了。到了卖家自己的网站上，就是他们说了算。",[23,363,364,370],{},[40,365,366],{},[26,367,369],{"href":368},"\u002Fguides\u002Fcent-accounts-explained","美分账户","。以美分计价的盈利，看上去比实际大一百倍。请看账户头部的货币和余额，别只看醒目的大标题。",[23,372,373,376],{},[40,374,375],{},"重置","。每隔几个月就从头再来的记录，其实是一连串尝试，失败的那几次都删掉了。识别信号：账户存续时间对不上宣称的历史，月度收益表有断档，回撤比策略自己最差的单月还小。",[23,378,379,382],{},[40,380,381],{},"十二个账户里最好的那个","。用略有差别的参数在十二个账户上跑同一个 EA，只公布跑赢的那个：每个数字都是真的，这样的展示却依然是谎言。挑选是 EA 圈最常见的不诚实手段，因为它根本不用造假。",[23,384,385,386,389],{},"请卖家把跑这个机器人的",[40,387,388],{},"所有","账户都拿出来，停掉的也算。在本站，交易者的公开主页会列出他公开的全部账户和投资组合，每个账户页面都链接到同一交易者的其他账户，投资组合则把多个账户合成一条曲线，回撤一点不少。",[23,391,392,393,396],{},"有一条边界，决定了任何诚实的平台能告诉您多少：我们",[40,394,395],{},"从不估计或指明某个账户上跑的是哪个 EA","，也从不按机器人做页面或排名。魔术号和它们的标签归账户所有者，所有者可以给自己的策略命名，愿意的话也可以公开这份分类统计。这样一来，举证责任就回到了它该在的地方：卖家要么拿出自己的账户给您看，要么您手里什么都没有。",[23,398,399,400,403],{},"一个问题就能戳穿上面四种手法：",[217,401,402],{},"把您最差那个月正在跑的账户给我看看","。心里没鬼的卖家，会直接回您一个链接。",[18,405,407],{"id":406},"在相信非凡之前先看看正常是什么样子","📐 在相信“非凡”之前，先看看“正常”是什么样子",[23,409,410],{},"没有基准，就没法判断一个宣传。下面是我们的基准，统计范围是有交易历史的公开账户。",[72,412,413,426],{},[75,414,415],{},[78,416,417,420,423],{},[81,418,419],{},"指标",[81,421,422],{},"中位数",[81,424,425],{},"怎么用",[91,427,428,439,449,460,471,482,493],{},[78,429,430,433,436],{},[96,431,432],{},"时间加权收益率",[96,434,435],{},"+3.2%（63.0% 为正）",[96,437,438],{},"两个季度赚钱只是一个样本，算不上优势",[78,440,441,444,446],{},[96,442,443],{},"最深回撤",[96,445,329],{},[96,447,448],{},"看任何收益宣传，都要连同回撤一起看",[78,450,451,454,457],{},[96,452,453],{},"盈利因子",[96,455,456],{},"1.28（100 笔以上交易时为 1.29）",[96,458,459],{},"要求 100 笔以上交易后几乎不变：高出这个中位数好几倍，就该打个问号",[78,461,462,465,468],{},[96,463,464],{},"胜率",[96,466,467],{},"68.8%",[96,469,470],{},"高胜率很常见，单凭它证明不了什么",[78,472,473,476,479],{},[96,474,475],{},"夏普比率",[96,477,478],{},"0.05",[96,480,481],{},"平滑的收益，比销售页面暗示的少见得多",[78,483,484,487,490],{},[96,485,486],{},"已平仓交易",[96,488,489],{},"每个账户 171 笔",[96,491,492],{},"不到 100 笔左右，统计数字只是摆设",[78,494,495,498,501],{},[96,496,497],{},"持仓时间",[96,499,500],{},"2.4 小时",[96,502,503],{},"看看它和卖家描述的策略是否对得上",[23,505,506,507,515,516,519],{},"这些账户是自愿公开的（大家只公开自己愿意给人看的），所以这些数字要说有偏差，也只会比现实好看。最有力的信号则很少见：在 1 万+ 个已连接账户中，",[40,508,509,510,514],{},"65 个带",[26,511,513],{"href":512},"\u002Fguides\u002Fthird-party-verification-explained","业绩记录已验证","徽章","，",[40,517,518],{},"264 个带交易权限已验证徽章","。两者都要账户所有者主动去做，所以才值得向卖家索要。",[18,521,523],{"id":522},"️-在-showmytrades-页面上做十分钟检查","🛠️ 在 ShowMyTrades 页面上做十分钟检查",[23,525,526],{},"如果卖家在本站公开了账户，上面说的一切点几下就能查到。各项内容的位置如下：",[72,528,529,539],{},[75,530,531],{},[78,532,533,536],{},[81,534,535],{},"去哪里看",[81,537,538],{},"能回答什么",[91,540,541,551,561,571,581,591,601,611],{},[78,542,543,548],{},[96,544,545],{},[40,546,547],{},"账户头部",[96,549,550],{},"真实账户还是模拟账户、业绩记录和交易权限徽章、平台、经纪商（悬停提示里有服务器名称）、货币、杠杆，以及自动交易或手动交易占比",[78,552,553,558],{},[96,554,555],{},[40,556,557],{},"账户统计",[96,559,560],{},"增长、绝对增长、日均 % 和月均 %；回撤和余额回撤；余额、净值、最高余额、盈利、入金、出金，信息图标里还有创建日期",[78,562,563,568],{},[96,564,565],{},[40,566,567],{},"图表",[96,569,570],{},"增长、余额、盈利、逐笔交易增长、回撤。余额视图把净值和余额画在一起，扛着不平的亏损单在这里一眼就能看出来",[78,572,573,578],{},[96,574,575],{},[40,576,577],{},"月度收益",[96,579,580],{},"亏损的月份、断档，还有最差的月份，而不是最好的",[78,582,583,588],{},[96,584,585],{},[40,586,587],{},"高级统计",[96,589,590],{},"盈利因子、夏普比率、Z 值、期望值、AHPR 与 GHPR 对比、总手数、佣金合计、已付库存费合计、平均持仓时间，以及按星期、按小时和按持续时间的分类",[78,592,593,598],{},[96,594,595],{},[40,596,597],{},"分类统计（按货币对）",[96,599,600],{},"每个品种的集中程度；所有者公开了自己的标签时，还有一个按魔术号分类的标签页",[78,602,603,608],{},[96,604,605],{},[40,606,607],{},"交易表格",[96,609,610],{},"订单号、品种、手数、止损、止盈、点数、盈利（毛额）、持续时间、库存费、佣金、魔术号和注释，持仓和已平仓交易分成两个标签页",[78,612,613,618],{},[96,614,615],{},[40,616,617],{},"自定义分析",[96,619,620],{},"可按日期范围、品种、魔术号、方向、星期、交易时段、手数、盈亏范围、持续时间和注释筛选：每一项统计都会按筛选出的子集重新计算",[23,622,623,624,35],{},"账户所有者可以隐藏手数、注释、止损止盈价位、笔记，甚至整张持仓或已平仓交易列表。这是正当权利，但本身也是信息：留意卖家藏起了什么。每个模块的详细说明见",[26,625,627],{"href":626},"\u002Fguides\u002Freading-a-trading-account-dashboard","如何读懂交易账户控制台",[23,629,630],{},"然后按下面的顺序逐项检查，碰到第一个硬伤就可以停了。",[257,632,633,639,645,651,657,662,673,679,685,691],{},[159,634,635,638],{},[40,636,637],{},"实盘还是模拟","？看账户头部的账户类型徽章。是模拟账户，到此为止。",[159,640,641,644],{},[40,642,643],{},"账户开了多久","？把创建日期和宣称的历史对一对。",[159,646,647,650],{},[40,648,649],{},"两种回撤","。把回撤和余额回撤放在一起看。差距大，说明有浮亏的持仓没平。",[159,652,653,656],{},[40,654,655],{},"月度表格","。历史很长却一个亏损月都没有，或者有说不清的断档，说明亏损藏在别处。",[159,658,659,661],{},[40,660,587],{},"。盈利因子是 1.28 中位数的好几倍，或者平均亏损远大于平均盈利，都得有个说法。",[159,663,664,667,668,672],{},[40,665,666],{},"成本","。把佣金合计、已付库存费合计和盈利放在一起比；自己要做实盘测试的话，先用",[26,669,671],{"href":670},"\u002Ftools\u002Fposition-size-calculator","仓位计算器","定好仓位。",[159,674,675,678],{},[40,676,677],{},"手数递增","。在自定义分析中筛选到一个品种，看交易量（手数）一列有没有成倍放大。",[159,680,681,684],{},[40,682,683],{},"持仓","。数一数。同一个货币对上一堆仓位、全部浮亏，就是跑到一半的网格。",[159,686,687,690],{},[40,688,689],{},"近期区间","。在自定义分析里设一个九十天的日期范围。优势还在不在？",[159,692,693,696],{},[40,694,695],{},"其他账户","。打开这位交易者公开的其他账户，再问他要没公开的那些。",[23,698,699],{},"追问了十分钟，卖家还拿不出一个与经纪商同步的实盘账户网址，答案其实已经有了。",[18,701,703],{"id":702},"常见问题","❓ 常见问题",[23,705,706,709],{},[40,707,708],{},"回测就完全没用吗？","\n不是。回测适合用来否定想法，不适合用来证实想法。干净的回测都过不了，策略可以直接判死刑；过了，也只是换来一次前向测试的资格。",[23,711,712,715,716,719,720,35],{},[40,713,714],{},"卖家的账户有验证徽章，这样就安全了吗？","\n更安全，但谈不上安全。验证说明数据是真的，而且账户确实在发布者手里。至于策略能不能熬过下一个市场阶段，这是不是十二个账户里挑出来的那一个，验证一概不管。请看",[26,717,718],{"href":128},"已验证与未验证","；想知道这个说法实际有多少分量，再看",[26,721,723],{"href":722},"\u002Fverified-track-record","经验证的业绩记录能证明什么",[23,725,726,729,730,734],{},[40,727,728],{},"EA 用了马丁格尔，而且明说了，怎么办？","\n摆在明面上的风险，总好过藏起来的。假设历史上最糟糕的那段行情会在您持有期间重演，先用",[26,731,733],{"href":732},"\u002Ftools\u002Fdrawdown-calculator","回撤计算器","把它模拟一遍。",[23,736,737,740],{},[40,738,739],{},"买之前能自己测一下 EA 吗？","\n先在模拟账户上跑，看运行机制；再用经纪商允许的最小仓位上实盘，看成交。模拟账户只能回答“它能不能成交”；只有真金白银才能回答“它能不能按我的价格成交”。",[23,742,743,746,747,751],{},[40,744,745],{},"我是卖 EA 的，怎样证明自己的业绩？","\n连接实盘账户，公开它，拿到两个徽章，步骤见",[26,748,750],{"href":749},"\u002Fguides\u002Fhow-to-verify-your-account","如何验证您的交易账户","。有了公开页面，这里的每一个质疑，买家还没开口就已经有了答案。",[18,753,755],{"id":754},"相关指南","🔗 相关指南",[156,757,758,764,769,775,780],{},[159,759,760,763],{},[26,761,762],{"href":49},"验证交易业绩宣传的终极指南","：针对信号和代管账户的取证手册",[159,765,766,768],{},[26,767,627],{"href":626},"：逐项讲解账户页面上的每个指标",[159,770,771,774],{},[26,772,773],{"href":128},"经验证与未经验证的业绩记录","：徽章能证明什么，不能证明什么",[159,776,777,779],{},[26,778,750],{"href":749},"：为自己的账户拿到两个徽章",[159,781,782,786],{},[26,783,785],{"href":784},"\u002Fguides\u002Ftracking-trading-performance","交易表现追踪","：机器人上了实盘之后该记录什么",[788,789],"hr",{},[23,791,792,35,795,799,800,804],{},[40,793,794],{},"用诚实的方式检验卖家",[26,796,798],{"href":797},"\u002Fexplore","探索","里的每个账户都是实盘、与经纪商同步、公开可查的，回撤也不例外。卖 EA？",[26,801,803],{"href":802},"\u002Fsignup","免费连接您的账户","，让业绩替您说话。",{"title":806,"searchDepth":807,"depth":807,"links":808},"",2,[809,810,811,812,813,814,815,816,817,818],{"id":20,"depth":807,"text":21},{"id":56,"depth":807,"text":57},{"id":143,"depth":807,"text":144},{"id":188,"depth":807,"text":189},{"id":234,"depth":807,"text":235},{"id":347,"depth":807,"text":348},{"id":406,"depth":807,"text":407},{"id":522,"depth":807,"text":523},{"id":702,"depth":807,"text":703},{"id":754,"depth":807,"text":755},"2026-08-19T00:00:00.000Z","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。",false,"md",{"src":824},"\u002Fimages\u002Fblog\u002Fspotting-fake-ea-results-hero.svg",{},true,"\u002Fzh\u002Fguides\u002Fspotting-fake-ea-results",{"title":5,"description":820},"zh\u002F3.guides\u002F29.spotting-fake-ea-results","sBcWYCc7mCgU9DBwS8OaUxD_R_IoLMwmYZd0udMQgN8",[832,837],{"title":833,"path":834,"stem":835,"description":836,"children":-1},"同一策略、两家经纪商、不同结果：如何做对比测试（{YEAR} 年版）","\u002Fzh\u002Fguides\u002Fsame-strategy-different-brokers","zh\u002F3.guides\u002F28.same-strategy-different-brokers","同一套策略，两家经纪商，跑出两条不同的曲线。本文讲如何让两个账户并行运行，并把差距拆解到仓位大小、成本、报价质量和执行延迟上。",{"title":838,"path":839,"stem":840,"description":841,"children":-1},"您的胜率在骗您：真实账户数据怎么说（{YEAR} 年版）","\u002Fzh\u002Fguides\u002Fwin-rate-is-not-an-edge","zh\u002F3.guides\u002F30.win-rate-is-not-an-edge","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。",[843,845],{"path":844},"\u002Fguides\u002Fspotting-fake-ea-results",{"path":827},1793898037885]