A key task in value investing is determining the “normal” earning power of a company in any given year.价值投资的一个重要任务是要确定在任何一年的公司“正常”的盈利能力。
People use a variety of methods to do this, the most famous being Benjamin Graham's average earnings from the company's previous 7-10 years of operation.人们使用各种方法来做到这一点，最著名的是本杰明格雷厄姆从该公司以前的7-10年的经营的平均收入。
Here's a technique I use to get a sense of a company's historical normal earning power while weighing the highest and lowest earning years a little less.这里我使用的技术得到了公司的正常盈利能力的历史意义，而体重的最高和最低的年收入少。
The method is called the Program Evaluation and Review Technique (PERT) and is a project management tool invented by Booz Allen Hamilton and the US Navy to build the Polaris missile system in the late 1950s.这种方法被称为计划评审技术（PERT），是一个项目管理工具，博思艾伦和美国海军发明建立在50年代后期的北极星导弹系统。 It was designed to determine likely schedule outcomes for specific project tasks based on estimates of the most likely, optimistic, and pessimistic times for completion.它旨在确定具体项目上最有可能的，乐观的估计为基础的任务可能安排的结果，并完成悲观倍。 [1] [1]
Here's how it works.下面是它的工作原理。
Let's say you have earnings from operations per share for a company for a 10-year period.假设你有一公司的10年期间的每股运营盈余。
Instead of using a straight average, note the best year and the worst year.而不是使用直线平均，注意最好的一年，最糟糕的一年。
Label the best year the optimistic earning power of the company and the worst year the pessimistic earning power of the company.标签最好的一年该公司的盈利能力持乐观态度，而最差的一年，该公司悲观的盈利能力。
You could determine the most likely earning power in a variety of ways.你可以决定在以各种方式最有可能的盈利能力。
What I do is take the average of the remaining 8 numbers.我要做的就是走，其余的8数字的平均值。
Then multiply the most likely earning power by 4 and add the optimistic and pessimistic numbers to that figure.然后乘以4，最有可能的盈利能力，并添加乐观和悲观的数字，这个数字。
Divide the total by 6.除以6个。
The result is called the expected value and is a normalized answer.结果称为预期值，是一个标准化的答案。 [2] [2]
For example, let's say the reported earnings per share for a company for the years 1998-2007 are:例如，假设每股报告的一家公司的年收入为1998-2007如下：
$1.22, $2.12, $2.62, $2.75, $2.59, $3.42, $3.26, $3.82, $4.25, and $.72. 1.22美元，2.12美元，2.62美元，2.75美元，2.59美元，3.42美元，3.26美元，3.82美元，4.25美元和0.72美元。
$4.25 is your optimistic year and $.72 is your pessimistic one. 4.25美元，今年是你的乐观和0.72美元是您的悲观。
The straight average of the rest is $2.73 and is your most likely number.其余的直平均2.73美元，是你最有可能数目。
4 X 2.73= 10.92. 4 × 2.73 = 10.92。
Adding $4.25 and $.72 gives you $15.89.添加4.25美元和0.72美元给你十五点八九美元。
Now divide by 6 to get $2.65.现在，除以6得到2.65美元。
Using the PERT method, $2.65 would be your mean or “normalized” earnings per share for this company based on the 10-year historical record.使用PERT方法，将2.65美元的平均或“标准化”的每股收益在10年的历史纪录的这家公司。 [3] [3]
The standard deviation (SD) is equal to the pessimistic number minus the optimistic number divided by 6.标准偏差（处）等于减去除以6悲观的乐观数字号码。
In this case the SD is (.72-4.25)/6= -.59在这种情况下，SD是（.72-4.25）/ 6 = -. 59
If this company's earnings are normally distributed, that means that 68.4% of the time they will fall within plus or minus 1 SD of the $2.65 mean (between $2.06 & $3.24), 95.4% of the time they will fall within plus or minus 2 SD of the mean (between $1.47 and $3.83), and 99.7% of the time they will fall within plus or minus 3 SD of the mean ($.88 and $4.42). 如果这家公司的盈利通常分布，这意味着68.4％的时间则属于正负1 2.65美元平均统计（至2.06美元及3.24美元），95.4％的时间则属于正负2标清平均数（1.47美元至3.83美元的），和99.7的时间则属于正负3％的平均统计（0.88美元美元和4.42）。 [4] [4]
But don't get too attached to this number: a business is a very dynamic organism and earnings may not necessarily follow a normal distribution.但不要太重视这个数目：一个企业是一个非常动态的有机体和收益不一定服从正态分布。
Even if they do that .3% does happen: the company in our example earned $.72 in 2007, below the 99.7% probability low of $.88.即使他们这样做，0.3％确实发生：在我们的例子，该公司2007年盈利0.72美元，低于99.7％的概率0.88美元低。
Plus, statistically speaking, 10 data points is not a whole lot.另外，从统计数字上看，10个数据点不是一大堆。
Regardless, I've found the method a useful place to start, especially when examining companies with stable track records and business models.无论如何，我发现该方法的有益的开端，特别是在审查与稳定的业绩和商业模式的公司。
If you're interested in learning more about the PERT method and the normal and beta distributions as they relate to project management check out the following link:如果你有兴趣了解PERT方法和正常和β分布更因为这涉及到项目管理进行检查以下链接：
http://www.interventions.org/pertcpm.html http://www.interventions.org/pertcpm.html
[1] [1] http://en.wikipedia.org/wiki/Program_Evaluation_and_Review_Technique http://en.wikipedia.org/wiki/Program_Evaluation_and_Review_Technique
[2] [2] The PERT calculations are based on a beta distribution but the result is normal in the statistical sense.在PERT的计算是基于对Beta分布，但结果是在统计意义上正常。 My thanks to Mr. Ajit Mani of Intervention Ltd. of Bangalore, India for helping me understand why: “The weights are based on an approximation of the Beta Distribution, NOT the Normal.我感谢先生的干预玛尼阿吉特班加罗尔，帮助我理解为什么说：“加权是一个基于Beta分布近似印度公司，而不是正常。 The originators of PERT selected the Beta distribution for the following qualities: a) It is unimodal; b) Has finite and non-negative end points; c) Is not necessarily symmetrical. PERT技术的原创者选择了以下素质Beta分布：1）这是单一方式;二）有有限的，非负结束点;三）不一定是对称的。 The Normal distribution does not satisfy qualities b) and c).正态分布不符合质量b）和c）。 The Normal curve is used for calculating Probability of Completing a Project by a Given Date.正态曲线用于计算完成某一日期一个项目的可能性。 This is possible because the Project Length T e is calculated by simply adding the t e 's along the Critical Path.这是可能的，因为工程长度Ť e是通过添加吨 é'计算拧沿关键路径。 Since the t e 's are all random variables, so is T e . 由于 T é's的所有随机变量，因此是 T 的位置 。 'But the interesting (and fortunate, from a statistical standpoint) result is that T e does not have the same distribution as the t e 's but follows what is called a normal distribution…' Please refer Wiest and Levy, 'A Management Guide to PERT/CPM', Prentice-Hall International, Inc., Englewood Cliffs, NJ USA, 1969, Chapter 4. '但是，有趣的（幸运的，从统计学的角度来看）的结果是那件T e不具有与吨 é'，但相同的分布拧如下所谓正态分布...'请参阅威斯特和利维，'A管理指南在计划评审技术/ CPM的'，普伦蒂斯-厅国际公司，联合利华，美国新泽西州，1969年，第4章。 The PERT Model.”在PERT的模型。“
[3] [3] Of course, whether the company can maintain and/or grow this normal earning power in the future without dilution is the key.当然，是否能够保持和/或在未来成长未经稀释这是正常的盈利能力是关键。
[4] [4] http://en.wikipedia.org/wiki/Normal_distribution http://en.wikipedia.org/wiki/Normal_distribution
