Monte Carlo Method

<body onLoad="MM_checkPlugin('Shockwave Flash','','http://www.riskglossary.com/default_nofl.htm',false);return document.MM_returnValue" stylesrc="../_private/template.htm" bgcolor="#FFFFFF" link="#669933" vlink="#000000" alink="#000000"> <table border="0" cellpadding="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" id="table1"> <tr> <td width="740"> <table border="0" cellpadding="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" height="37" id="table2"> <tr> <td height="35" width="100%"> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" id="table3"> <tr> <td width="100%" align="left" valign="top"> <p class="header_gold">Monte Carlo Method</td> <td width="45" align="right" valign="bottom"> <map name="FPMap6"> <area href="http://www.contingencyanalysis.com" shape="rect" coords="1, 0, 44, 11" target="_top"> </map> <img border="0" src="../images/home.gif" usemap="#FPMap6" width="45" height="12"></td> </tr> </table> </td> </tr> <tr> <td width="100%" bgcolor="#CC9966" height="2"> <img border="0" src="../images/blank_2x2.gif" width="2" height="2"></td> </tr> <tr> <td width="0%" valign="top" bgcolor="#FFFFCC"> <table border="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" cellspacing="1" id="table4"> <tr> <td width="0%" align="left" valign="top"> <p class="header2_black">Explained:<br> <img border="0" src="../images/blank_1x100.gif" width="100" height="1"></td> <td width="99%" align="left" valign="top"><p class="word_list"> <a href="#antithetic variates">antithetic variates</a><p class="word_list"> <a href="#control variates">control variates</a><p class="word_list"> <a href="#crude Monte Carlo estimator">crude Monte Carlo estimator</a><p class="word_list"> <a href="#curse of dimensionality">curse of dimensionality</a><p class="word_list"> <a href="#importance sampling">importance sampling</a><p class="word_list"> <a href="#Monte Carlo method">Monte Carlo method</a><p class="word_list"> <a href="#stratified sampling">stratified sampling</a><br> <img border="0" src="../images/blank_1x160.gif" width="160" height="1"></td> </tr> </table> </td> </tr> </table> <table border="0" cellpadding="0" style="border-collapse: collapse" bordercolor="#111111" width="310" align="left" id="table5"> <tr> <td width="300" height="14"><font size="2"> </font><font size="4"> </font></td> <td width="9" height="14"></td> </tr> <tr> <td width="300"> <script type="text/javascript"></script> <script type="text/javascript" src="http://pagead2.googlesyndication.com/pagead/show_ads.js"> </script>  </td> <td width="9"> </td> </tr> </table> <p class="body">Many practitioners have some intuitive familiarity with the Monte Carlo method from their work. At an elementary level, it is a surprisingly simple concept, but it can be computationally expensive to use. It is easy to code Monte Carlo analyses that take hours or even days to run. To speed up analyses—to make them run in minutes as opposed to days—users need to employ techniques such as variance reduction. These techniques are easy to learn, but they are NOT intuitive. To use them, users need a sophisticated understanding of how and why the Monte Carlo method works. In this article, I introduce the Monte Carlo method from such a standpoint. I also close the article with some recommended books.</p> <p class="body">Credit for inventing the Monte Carlo method often goes to Stanislaw Ulam, a Polish born mathematician who worked for John von Neumann on the United States’ Manhattan Project during World War II. Ulam is primarily known for designing the hydrogen bomb with Edward Teller in 1951. He invented the Monte Carlo method in 1946 while pondering the probabilities of winning a card game of solitaire. Quoted in Eckhardt (<a href="#Eckhardt">1987</a>), Ulam describes the incident as: </p> <p class="quote">The first thoughts and attempts I made to practice [the Monte Carlo Method] were suggested by a question which occurred to me in 1946 as I was convalescing from an illness and playing solitaires. The question was what are the chances that a Canfield solitaire laid out with 52 cards will come out successfully? After spending a lot of time trying to estimate them by pure combinatorial calculations, I wondered whether a more practical method than “abstract thinking” might not be to lay it out say one hundred times and simply observe and count the number of successful plays. This was already possible to envisage with the beginning of the new era of fast computers, and I immediately thought of problems of neutron diffusion and other questions of mathematical physics, and more generally how to change processes described by certain differential equations into an equivalent form interpretable as a succession of random operations. Later … [in 1946, I] described the idea to John von Neumann, and we began to plan actual calculations.</p> <table border="0" cellpadding="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="345" align="right" id="table6"> <tr> <td align="right" width="9"> </td> <td width="336">  <p style="margin-top: 0; margin-bottom: 2" class="ad_legend"> <a target="_top" href="http://www.riskexpertise.com/link/advertising.htm"> <font color="#000000">Ads by Contingency Analysis</font></a></p> <p style="margin-top: 0; margin-bottom: 0"> <iframe src="http://servedbyadbutler.com/adserve/;ID=147917;size=336x265;setID=82402;type=iframe" style="margin: 0px; padding: 0px; border-width: 0px" width="336" height="265" frameborder="0" marginwidth="0" marginheight="0" scrolling="no" name="I3"> <![if lt IE 4]> <script src="http://servedbyadbutler.com/adserve/;ID=147917;size=336x265;setID=82402;type=js" type="text/javascript"> </script> <noscript> <a href="http://servedbyadbutler.com/go2/;ID=147917;size=336x265;setID=82402"> <img src="http://servedbyadbutler.com/adserve.ibs/;ID=147917;size=336x265;setID=82402;type=img" border="0" height="265" width="336" alt=""></a> </noscript><![endif]></iframe>   </td> </tr> </table> <p class="body">The <b><a name="Monte Carlo method">Monte Carlo method</a></b>, as it is understood today, encompasses any technique of statistical sampling employed to approximate solutions to quantitative problems. Ulam did not invent statistical sampling. This had been employed to solve quantitative problems before, with physical processes such as dice tosses or card draws being used to generate samples. W. S. Gossett, who published under the pen name “Student,” randomly sampled from height and middle finger measurements of 3,000 criminals to simulate two correlated normal distributions. He discusses this methodology in both Student (<a href="#Student (1908a)">1908a</a>) and Student (<a href="#Student (1908b)">1908b</a>). Ulam’s contribution was to recognize the potential for the newly invented electronic computer to automate such sampling. Working with John von Neuman and Nicholas Metropolis, he developed algorithms for computer implementations, as well as exploring means of transforming non-random problems into random forms that would facilitate their solution via statistical sampling. This work transformed statistical sampling from a mathematical curiosity to a formal methodology applicable to a wide variety of problems. It was Metropolis who named the new methodology after the casinos of Monte Carlo. Ulam and Metropolis published the <a href="#Metropolis">first paper</a> on the Monte Carlo method in 1949.</p> <p class="body">A standard way to introduce the Monte Carlo method is with the problem of Buffon's needle. This is a charming problem that leads directly to some important (not all intuitive!) conclusions.</p> <p class="body">More than 200 years before Metropolis coined the name “Monte Carlo method,” George Louis Leclerc, Comte de Buffon, proposed the following problem. If a needle of length <i>l</i> is dropped at random on the middle of a horizontal surface ruled with parallel lines a distance <i>d</i> > <i> l</i> apart, what is the probability that the needle will cross one of the lines? </p> <p class="body"><span style="font-size: 12.0pt; font-family: Times New Roman">To solve the problem, consider Exhibit 1. </span></p> <div align="center"> <center> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="300" cellpadding="0" id="table7"> <tr> <td> <p class="exhibit_header"><span class="exhibit_header">Problem of Buffon's Needle</span><br> <span class="exhibit_subheader">Exhibit 1</span></td> </tr> <tr> <td bgcolor="#CC9966" height="2"> <img border="0" src="../images/blank_2x2.gif" width="2" height="2"></td> </tr> <tr> <td> <img border="0" src="../images/ex1_monte_carlo.gif" width="300" height="200"></td> </tr> <tr> <td> <p class="body">Buffon asked what was the probability that the needle would fall across one of the lines, marked here in green. That outcome will occur only if <i>A</i> < <img border="0" src="../formulas/monte_1.gif" align="absbottom" width="30" height="13">.</td> </tr> </table> </center> </div> <p class="body">The positioning of the needle relative to nearby lines can be described with a random vector which has components <img border="0" src="../formulas/monte_carlo_A_in_[0,d).gif" align="absbottom" width="56" height="15"> and <img border="0" src="../formulas/monte_2.gif" align="absbottom" width="52" height="15">. The random vector (<i>A</i>, <img border="0" src="../formulas/monte_3.gif" align="absbottom" width="7" height="14">) is uniformly distributed on the region [0,<i>d</i>) <img border="0" src="../formulas/monte_4.gif" align="absbottom" width="7" height="11"> [0,<img border="0" src="../formulas/monte_5.gif" align="absbottom" width="9" height="13">). Accordingly, it has probability density function <img border="0" src="../formulas/monte_carlo_1_over_d_pi.gif" align="absbottom" width="26" height="15">. The probability that the needle will cross one of the lines <a name="[1]">is</a> given by the integral</p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table8"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_buffon_integral.gif" width="106" height="59"></td> <td width="5%" align="right">[1]</td> </tr> </table> <p class="body">This integral is easily solved to obtain the probability <a name="[2]">as</a></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table9"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_probability_equals.gif" width="112" height="41"></td> <td width="5%" align="right">[2]</td> </tr> </table> <p class="body">In this solution, the great mathematician Laplace perceived an amusing, if inefficient means of approximating the number <img border="0" src="../formulas/aaa_pi.gif" align="absbottom" width="9" height="13"><span style="font-family: Symbol">.</span> Suppose Buffon’s experiment is performed with the needle being dropped <i> n</i> times. Let <i>M</i> be the random variable for the number of times the needle crosses a line, <a name="[3]">then</a></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table10"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_second_probability_equals.gif" width="139" height="17"></td> <td width="5%" align="right">[3]</td> </tr> </table> <p class="body">where <i>E</i> indicated an <a href="../articles/mean.htm">expectation</a>. Note that formulas [<a href="#[2]">2</a>] and [<a href="#[3]">3</a>] represent the same probability. Setting them equal to each other and rearranging, we obtain an expression for <img border="0" src="../formulas/aaa_pi.gif" align="absbottom" width="9" height="13">:</p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table11"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_100.gif" width="90" height="44"></td> <td width="5%" align="right">[4]</td> </tr> </table> <p class="MsoNormal">This means <a name="[5]">that</a> </p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table12"> <tr> <td width="100%" align="center"> <img border="0" src="../images/monte_carlo_k95mv43.gif" width="36" height="41"></td> <td width="5%" align="right">[5]</td> </tr> </table> <p class="MsoNormal">is a statistical estimator for <img border="0" src="../formulas/aaa_pi.gif" align="absbottom" width="9" height="13">. If we drop the needle <i>n</i> times, and it crosses a line <i>m</i> times, we can substitute the result <i>m</i> for the random variable <i>M</i> in [<a href="#[5]">5</a>]. The number we obtain will be an estimate for <img border="0" src="../formulas/aaa_pi.gif" align="absbottom" width="9" height="13">.</p> <p class="MsoNormal">In 1864, Captain O. C. Fox performed this experiment three times to occupy himself while recovering from wounds. His results are shown in Exhibit 2:</p> <div align="center"> <center> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="420" cellpadding="0" id="table13"> <tr> <td colspan="6"> <p class="exhibit_header"><span class="exhibit_header">Results of Fox's Needle Dropping Experiments</span><br> <span class="exhibit_subheader">Exhibit 2</span></td> </tr> <tr> <td bgcolor="#CC9966" height="2" colspan="6"> <img border="0" src="../images/blank_2x2.gif" width="2" height="2"></td> </tr> <tr> <td align="center" width="70" bgcolor="#FFFFCC"> <i>n</i></td> <td align="center" width="70" bgcolor="#FFFFCC"> <i>m</i></td> <td align="center" width="70" bgcolor="#FFFFCC"> <i>l</i><br> inches</td> <td align="center" width="70" bgcolor="#FFFFCC"> <i>d</i><br> inches</td> <td align="center" width="70" bgcolor="#FFFFCC"> surface</td> <td align="center" width="70" bgcolor="#FFFFCC"> <img border="0" src="../formulas/aaa_pi.gif" align="absbottom" width="9" height="13"><br> estimate</td> </tr> <tr> <td height="1" bordercolor="#CC9966" bgcolor="#FFFFCC" colspan="6" align="center" width="70"> <img border="0" src="../images/blank_1x1.gif" width="1" height="1"></td> </tr> <tr> <td align="center" width="70" bgcolor="#FFFFCC"> 500</td> <td align="center" width="70" bgcolor="#FFFFCC"> 236</td> <td align="center" width="70" bgcolor="#FFFFCC"> 3</td> <td align="center" width="70" bgcolor="#FFFFCC"> 4</td> <td align="center" width="70" bgcolor="#FFFFCC"> stationary</td> <td align="center" width="70" bgcolor="#FFFFCC"> 3.1780</td> </tr> <tr> <td align="center" width="70" bgcolor="#FFFFCC"> 530</td> <td align="center" width="70" bgcolor="#FFFFCC"> 253</td> <td align="center" width="70" bgcolor="#FFFFCC"> 3</td> <td align="center" width="70" bgcolor="#FFFFCC"> 4</td> <td align="center" width="70" bgcolor="#FFFFCC"> rotated</td> <td align="center" width="70" bgcolor="#FFFFCC"> 3.1423</td> </tr> <tr> <td align="center" width="70" bgcolor="#FFFFCC"> 590</td> <td align="center" width="70" bgcolor="#FFFFCC"> 939</td> <td align="center" width="70" bgcolor="#FFFFCC"> 5</td> <td align="center" width="70" bgcolor="#FFFFCC"> 2</td> <td align="center" width="70" bgcolor="#FFFFCC"> rotated</td> <td align="center" width="70" bgcolor="#FFFFCC"> 3.1416</td> </tr> <tr> <td bgcolor="#CC9966" height="2" colspan="6"> <img border="0" src="../images/blank_2x2.gif" width="2" height="2"></td> </tr> <tr> <td colspan="6"> <p class="exhibit_legend">Results of Fox's three needle-dropping experiments are indicated. In the second two experiments, Fox rotated the surface between drops.</td> </tr> </table> </center> </div> <p class="body" style="margin-left: 0in; margin-right: 0in; margin-top: 0in; margin-bottom: .0001pt"> </p> <p class="body"> Fox’s results illustrate two important issues for the Monte Carlo method. </p> <table border="0" cellpadding="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="345" align="right" id="table14"> <tr> <td align="right" width="9"> </td> <td width="336"> <script type="text/javascript"></script> <script type="text/javascript" src="http://pagead2.googlesyndication.com/pagead/show_ads.js"> </script>  </td> </tr> </table> <p class="body"><img border="0" src="../images/bullet.gif" width="20" height="10">After obtaining a poor approximation for <img border="0" src="../formulas/aaa_pi.gif" align="absbottom" width="9" height="13"> in his first experiment, Fox modified his subsequent experiments by applying a slight rotary motion to the ruled surface between drops. He did this to eliminate any bias arising from his position while dropping the needle. His approximations in subsequent experiments were improved. This same need to eliminate bias exists today with computer implementations of the Monte Carlo method, although such biases now arise from inappropriate pseudorandom number generators instead of posture. <p class="body"><img border="0" src="../images/bullet.gif" width="20" height="10">In his third experiment, Fox used a five inch needle and made the lines on the surface just two inches apart. Now, the needle could cross as many as three lines in a single toss. In 590 tosses, Fox obtained 939 line crossings. Doing so entailed little more effort than his first two experiments, but it yielded a better approximation for <img border="0" src="../formulas/aaa_pi.gif" align="absbottom" width="9" height="13">. Fox’s innovation was a precursor of today’s variance reduction techniques. <p class="body">This example also illustrates another point. Many practitioners who first experience the Monte Carlo method on the job see it being used to solve inherently probabilistic problems, such as <a href="../articles/derivative_instrument.htm">derivatives</a> pricing or weather forecasting. They assume the use of random sampling makes it applicable only to probabilistic problems. As our example illustrates, this is not true. There is nothing probabilistic about estimating the number <img border="0" src="../formulas/aaa_pi.gif" align="absbottom" width="9" height="13">!<p class="body"> To understand the Monte Carlo method theoretically, it is useful to think of it as a general technique of numerical integration. It can be shown, at least in a trivial sense, that every application of the Monte Carlo method can be represented as a definite integral. As one example, consider our needle dropping experiment. It is nothing more than an elaborate means of estimating the integral [<a href="#[1]">1</a>].</p> <p> Suppose we need to evaluate a multi-dimensional definite integral of the <a name="[6]">form</a>:</p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table15"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_o48gfjf.gif" width="400" height="60"></td> <td width="5%" align="right">[6]</td> </tr> </table> <p class="body">Most integrals can be converted to this form with a suitable change of variables, so we can consider this to be a general application suitable for the Monte Carlo method.</p> <p class="body">The integral represents a non-random problem, but the Monte Carlo method approximates a solution by introducing a random vector <i><b>U</b></i> that is uniformly distributed on the region of integration. Applying the function <i>f</i> to <i><b>U</b></i>, we obtain a random variable <i>f</i>(<i><b>U</b></i>). This has <a name="[7]"> expectation</a>:</p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table16"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_irtutjgh.gif" width="208" height="51"></td> <td width="5%" align="right">[7]</td> </tr> </table> <p class="body">where <img border="0" src="../formulas/aaa_fi.gif" align="absbottom" width="7" height="15"> is the probability density function of <i><b>U</b></i>. Because <img border="0" src="../formulas/aaa_fi.gif" align="absbottom" width="7" height="15"> equals 1 on the region of integration, [<a href="#[7]">7</a>] <a name="[8]">becomes</a></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table17"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_i985867.gif" width="166" height="51"></td> <td width="5%" align="right">[8]</td> </tr> </table> <p class="body">Comparing [<a href="#[6]">6</a>] and [<a href="#[8]">8</a>],<span style="font-size: 12.0pt; font-family: Times New Roman"> we obtain a probabilistic expression for the integral <img border="0" src="../formulas/aaa_si.gif" align="absbottom" width="14" height="13">:</span></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table18"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_or3596h.gif" width="105" height="27"></td> <td width="5%" align="right">[9]</td> </tr> </table> <p class="body"> <span style="font-size: 12.0pt; font-family: Times New Roman">so random variable <i>f</i>(<b><i>U</i></b>) has mean <img border="0" src="../formulas/aaa_si.gif" align="absbottom" width="14" height="13"> and some standard deviation </span> <img border="0" src="../formulas/aaa_sigma.gif" align="absbottom" width="8" height="11"><span style="font-size: 12.0pt; font-family: Times New Roman">. We <a name="[10]">define</a></span> </p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table19"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_9375y9j45.gif" width="97" height="29"></td> <td width="5%" align="right">[10]</td> </tr> </table> <p class="body"><span style="font-size: 12.0pt; font-family: Times New Roman">as an unbiased estimator for <img border="0" src="../formulas/aaa_si.gif" align="absbottom" width="14" height="13"> with standard error </span> <img border="0" src="../formulas/aaa_sigma.gif" align="absbottom" width="8" height="11"><span style="font-size: 12.0pt; font-family: Times New Roman">. This is a little unconventional, since [<a href="#[10]">10</a>] is an estimator that depends upon a sample {<b><i>U</i></b><sup>[1]</sup>}of size one, but it is a valid estimator nonetheless.</span></p> <p class="body">To estimate <span style="font-size: 12.0pt; font-family: Times New Roman"> <img border="0" src="../formulas/aaa_si.gif" align="absbottom" width="14" height="13"></span> with a standard error lower than <img border="0" src="../formulas/aaa_sigma.gif" align="absbottom" width="8" height="11">, let’s generalize our estimator to accommodate a larger sample {<b><i>U</i></b><sup>[1]</sup>, <b><i>U</i></b><sup> [2]</sup>, … ,<b><i> U</i></b><sup> [<i>m</i>]</sup>}. Applying the function <i>f</i> to each of these yields <i>m</i> independent and identically distributed (IID) random variables <i>f</i>(<b><i>U</i></b><sup> [1]</sup>), <i>f</i>(<b><i>U</i></b><sup> [2]</sup>), … ,<i> f</i>(<b><i>U</i></b><sup> [<i>m</i>]</sup>), each with expectation <span style="font-size: 12.0pt; font-family: Times New Roman"> <img border="0" src="../formulas/aaa_si.gif" align="absbottom" width="14" height="13"></span> and standard deviation <img border="0" src="../formulas/aaa_sigma.gif" align="absbottom" width="8" height="11">. The generalization <a name="[11]">of</a> [<a href="#[10]">10</a>]</p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table20"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carloijcr98hjr.gif" width="144" height="54"></td> <td width="5%" align="right">[11]</td> </tr> </table> <p class="body">is an unbiased estimator for <img border="0" src="../formulas/aaa_si.gif" align="absbottom" width="14" height="13"> with standard <a name="[12]">error</a>:</p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table21"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_ojk4f9jtee.gif" width="32" height="44"></td> <td width="5%" align="right">[12]</td> </tr> </table> <p class="body">If we have a realization {<b><i>u</i></b><sup>[1]</sup>, <b><i>u</i></b><sup>[2]</sup>, … ,<b><i> u</i></b><sup>[<i>m</i>]</sup>} for our sample, we may estimate <span style="font-size: 12.0pt; font-family: Times New Roman"> <img border="0" src="../formulas/aaa_si.gif" align="absbottom" width="14" height="13"></span> as</p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table22"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_feer64.gif" width="147" height="54"></td> <td width="5%" align="right">[13]</td> </tr> </table> <table border="0" cellpadding="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="345" align="right" id="table23"> <tr> <td align="right" width="9" height="16"></td> <td width="336" height="16"> <font size="2"> </font><font size="4"> </font></td> </tr> <tr> <td align="right" width="9"> </td> <td width="336">  <p style="margin-top: 0; margin-bottom: 2" class="ad_legend"> <a target="_top" href="http://www.riskexpertise.com/link/advertising.htm"> <font color="#000000">Ads by Contingency Analysis</font></a></p> <p style="margin-top: 0; margin-bottom: 0"> <iframe src="http://servedbyadbutler.com/adserve/;ID=147917;size=336x265;setID=82402;type=iframe" style="margin: 0px; padding: 0px; border-width: 0px" width="336" height="265" frameborder="0" marginwidth="0" marginheight="0" scrolling="no" name="I4"> <![if lt IE 4]> <script src="http://servedbyadbutler.com/adserve/;ID=147917;size=336x265;setID=82402;type=js" type="text/javascript"> </script> <noscript> <a href="http://servedbyadbutler.com/go2/;ID=147917;size=336x265;setID=82402"> <img src="http://servedbyadbutler.com/adserve.ibs/;ID=147917;size=336x265;setID=82402;type=img" border="0" height="265" width="336" alt=""></a> </noscript><![endif]></iframe>   </td> </tr> </table> <p class="body">We call [<a href="#[11]">11</a>] the <b> <a name="crude Monte Carlo estimator">crude Monte Carlo estimator</a></b>. Formula [<a href="#[12]">12</a>] for its standard error is important for two reasons. First, it tells us that the standard error of a Monte Carlo analysis decreases with the square root of the sample size. If we quadruple the number of realizations used, we will half the standard error. Second, standard error does not depend upon the dimensionality of the integral [<a href="#[6]">6</a>]. Most techniques of numerical integration—such as the trapezoidal rule or Simpson's method—suffer from the <b><a name="curse of dimensionality"> curse of dimensionality</a></b>. When generalized to multiple dimensions, the number of computations required to apply them increases exponentially with the dimensionality of the integral. For this reason, such methods cannot be applied to integrals of more than a few dimensions. The Monte Carlo method does not suffer from the curse of dimensionality. It is as applicable to a 1000-dimensional integral as it is to a one-dimensional integral.</p> <p class="body">While increasing the sample size is one technique for reducing the standard error of a Monte Carlo analysis, doing so can be computationally expensive. A better solution is to employ some technique of variance reduction. These techniques incorporate additional information about the analysis directly into the estimator. This allows them to make the Monte Carlo estimator more deterministic, and hence have a lower standard error.</p> <p class="body">Standard techniques of variance reduction include</p> <p class="bullet"><img border="0" src="../images/bullet.gif" width="20" height="10"><b><a name="antithetic variates">antithetic variates</a></b>,</p> <p class="bullet"><img border="0" src="../images/bullet.gif" width="20" height="10"><b><a name="control variates">control variates</a></b>,</p> <p class="bullet"><img border="0" src="../images/bullet.gif" width="20" height="10"><b><a name="importance sampling">importance sampling</a></b>, and</p> <p class="bullet"><img border="0" src="../images/bullet.gif" width="20" height="10"><b><a name="stratified sampling">stratified sampling</a></b>.</p> <p class="body">Let's describe the method of control variates.</p> <p class="body"> <span style="font-size: 12.0pt; font-family: Times New Roman">Consider crude Monte Carlo <a name="[14]">estimator</a>:</span></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table24"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_4u398ht4.gif" width="109" height="54"></td> <td width="5%" align="right">[14]</td> </tr> </table> <p class="body">Let <img border="0" src="../formulas/xxx_xi.gif" align="absbottom" width="8" height="17"> be a <a href="../articles/set.htm">real</a> valued function for which the mean</p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table25"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_ij4f87.gif" width="107" height="30"></td> <td width="5%" align="right">[15]</td> </tr> </table> <p class="body">is known. <span style="font-size: 12.0pt; font-family: Times New Roman">We shall refer to the random variable </span> <img border="0" src="../formulas/xxx_xi.gif" align="absbottom" width="8" height="17"><span style="font-size: 12.0pt; font-family: Times New Roman">(<b><i>U</i></b>) as a <b>control variate</b>. Consider the random variable <i>f*</i>(<b><i>U</i></b>) based upon this control variate:</span></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table26"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_874rf.gif" width="219" height="31"></td> <td width="5%" align="right">[16]</td> </tr> </table> <p class="body">for some constant <i>c</i>. By construction,</p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table27"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_vrewtiu.gif" width="111" height="26"></td> <td width="5%" align="right">[17]</td> </tr> </table> <p class="MsoNormal">so we <span style="font-size: 12.0pt; font-family: Times New Roman">can estimate </span> <img border="0" src="../formulas/aaa_si.gif" align="absbottom" width="14" height="13"><span style="font-size: 12.0pt; font-family: Times New Roman"> with the Monte Carlo <a name="[18]">estimator</a></span></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table28"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carloaercr3qw.gif" width="385" height="54"></td> <td width="5%" align="right">[18]</td> </tr> </table> <p class="body"> <span style="font-size: 12.0pt; font-family: Times New Roman">This will have a lower standard error than crude estimator [<a href="#[14]">14</a>] if the standard deviation <img border="0" src="../formulas/aaa_sigma.gif" align="absbottom" width="8" height="11">* of <img border="0" src="../images/monte_carlo_9irtjug.gif" align="absbottom" width="37" height="15"> is smaller than the standard deviation <img border="0" src="../formulas/aaa_sigma.gif" align="absbottom" width="8" height="11"> of <i>f</i>(<b><i>U</i></b>). This will happen if </span> <img border="0" src="../formulas/xxx_xi.gif" align="absbottom" width="8" height="17"><span style="font-size: 12.0pt; font-family: Times New Roman">(<b><i>U</i></b>) has a high correlation <img border="0" src="../formulas/aaa_rho.gif" align="absbottom" width="7" height="12"> with the random variable <i>f</i>(<b><i>U</i></b>), in which case random variables <i>c</i></span><img border="0" src="../formulas/xxx_xi.gif" align="absbottom" width="8" height="17"><span style="font-size: 12.0pt; font-family: Times New Roman">(<b><i>U</i></b>) and <i>f</i>(<b><i>U</i></b>) will tend to offset each other in [<a href="#[18]">18</a>]. We formalize this observation by <a name="[21]">calculating</a></span></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table29"> <tr> <td width="35%" align="right"> <img border="0" src="../formulas/monte_carlo123445se.gif" width="20" height="15"></td> <td width="65%" align="left"> <img border="0" src="../formulas/monte_carlo_864jngg.gif" width="195" height="19"></td> <td width="5%" align="right">[19]</td> </tr> <tr> <td width="35%" align="right">  </td> <td width="65%" align="left">  </td> <td width="5%" align="right"> </td> </tr> <tr> <td width="35%" align="right">  </td> <td width="65%" align="left"> <img border="0" src="../formulas/monte_carlo_84ju.gif" width="144" height="17"></td> <td width="5%" align="right">[20]</td> </tr> <tr> <td width="35%" align="right">  </td> <td width="65%" align="left">  </td> <td width="5%" align="right"> </td> </tr> <tr> <td width="35%" align="right">  </td> <td width="65%" align="left"> <img border="0" src="../formulas/monte_carlo_935kgtn8fe.gif" width="175" height="31"></td> <td width="5%" align="right">[21]</td> </tr> </table> <p class="body"><font face="Times New Roman">where <img border="0" src="../formulas/monte_carlo_9iriu4mvve.gif" align="absbottom" width="15" height="13"> is </font><span style="font-size: 12.0pt; font-family: Times New Roman"> the standard deviation of </span> <img border="0" src="../formulas/xxx_xi.gif" align="absbottom" width="8" height="17"><span style="font-size: 12.0pt; font-family: Times New Roman">(<b><i>U</i></b>). Obviously, <img border="0" src="../formulas/aaa_sigma.gif" align="absbottom" width="8" height="11">* will be smaller than <img border="0" src="../formulas/aaa_sigma.gif" align="absbottom" width="8" height="11"> if</span></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table30"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_ir93kg4er.gif" width="65" height="53"></td> <td width="5%" align="right">[22]</td> </tr> </table> <p class="body"> <span style="font-size: 12.0pt; font-family: Times New Roman">It can be shown that <img border="0" src="../formulas/aaa_sigma.gif" align="absbottom" width="8" height="11">* is minimized by setting</span></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table31"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_e9kv342ges.gif" width="59" height="56"></td> <td width="5%" align="right">[23]</td> </tr> </table> <p class="body"><font face="Times New Roman">in which case, <a name="[24]"> from</a> [<a href="#[21]">21</a>],</font></p> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="0" height="36" id="table32"> <tr> <td width="100%" align="center"> <img border="0" src="../formulas/monte_carlo_983j8gtjvke.gif" width="100" height="32"></td> <td width="5%" align="right">[24]</td> </tr> </table> <p class="body"> <span style="font-size: 12.0pt; font-family: Times New Roman">Often, <img border="0" src="../formulas/aaa_rho.gif" align="absbottom" width="7" height="12"> and </span><font face="Times New Roman"> <img border="0" src="../formulas/monte_carlo_9iriu4mvve.gif" align="absbottom" width="15" height="13"></font><span style="font-size: 12.0pt; font-family: Times New Roman"> are unknown, which makes determining the optimal value [<a href="#[24]">24</a>] for <i>c</i> problematic. We can estimate <img border="0" src="../formulas/aaa_rho.gif" align="absbottom" width="7" height="12"> and </span><font face="Times New Roman"> <img border="0" src="../formulas/monte_carlo_9iriu4mvve.gif" align="absbottom" width="15" height="13"></font><span style="font-size: 12.0pt; font-family: Times New Roman"> with a separate Monte Carlo analysis. Alternatively, if </span> <img border="0" src="../formulas/xxx_xi.gif" align="absbottom" width="8" height="17"><span style="font-size: 12.0pt; font-family: Times New Roman"> closely approximates <i>f</i>, <i>c</i> might simply be set equal to 1.</span></p> <p class="body"><span style="font-family: Times New Roman">As the above description indicates, the key to the method of control variates is finding a function </span> <img border="0" src="../formulas/xxx_xi.gif" align="absbottom" width="8" height="17"> that closely approximates <i>f</i>, and for which <i>E</i>[<img border="0" src="../formulas/xxx_xi.gif" align="absbottom" width="8" height="17">(<i><b>U</b></i>)] is easy to calculate.</p> <table border="0" cellpadding="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" id="table33"> <tr> <td height="30" valign="bottom" colspan="2"> <p class="header2_black">Related Internal Links</td> </tr> <tr> <td colspan="2" bgcolor="#CC9966" height="2"> <img border="0" src="../images/blank_2x2.gif" width="2" height="2"></td> </tr> <tr> <td bgcolor="#FFFFCC" colspan="2"> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="6" id="table34"> <tr> <td width="100%"> <p class="explanatory_text"> <img border="0" src="../images/bullet_link_glossary.gif" width="26" height="15"><a href="../articles/monte_carlo_transformation.htm">Monte Carlo transformation</a> Describes the use of the Monte Carlo method in VaR measures.<p class="explanatory_text"> <img border="0" src="../images/bullet_link_glossary.gif" width="26" height="15"><strong style="font-weight: 400"><a href="../articles/closed_form_solution.htm">numerical solution</a> </strong>An approximation that can be evaluated in a finite number of standard operations.</td> </tr> </table> </td> </tr> <tr> <td height="30" valign="bottom" colspan="2"> <p class="header2_black">Sponsored Links</td> </tr> <tr> <td colspan="2" bgcolor="#CC9966" height="2"> <img border="0" src="../images/blank_2x2.gif" width="2" height="2"></td> </tr> <tr> <td bgcolor="#FFFFCC" width="75"> <img border="0" src="../images/ad_side_banner_1.jpg" width="103" height="285"></td> <td bgcolor="#FFFFCC" width="99%"> <script type="text/javascript"></script> <script type="text/javascript" src="http://pagead2.googlesyndication.com/pagead/show_ads.js"> </script>  </td> </tr> <tr> <td height="30" valign="bottom" colspan="2"> <p class="header2_black">Related Books</td> </tr> <tr> <td colspan="2" bgcolor="#CC9966" height="2"> <img border="0" src="../images/blank_2x2.gif" width="2" height="2"></td> </tr> <tr> <td bgcolor="#FFFFCC" colspan="2"> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="6" id="table35"> <tr> <td width="100%"> <p class="explanatory_text">For books on the Monte Carlo Method, see book reviews under <a target="_blank" href="http://www.riskbook.com/link_topic/math_monte_carlo.htm"> Mathematics–Monte Carlo Method</a>. In particular, Rubinstein (<a target="_blank" href="http://www.riskbook.com/link/rubinstein_r_(1981).htm">1981</a>) offers an excellent introduction. Fishman (<a target="_blank" href="http://www.riskbook.com/link/fishman_g_(1996).htm">1996</a>) is the authoritative reference. Glasserman (<a target="_blank" href="http://www.riskbook.com/link/glasserman_p_(2003).htm">2003</a>) is an advanced book on use of the Monte Carlo method in financial engineering.</p> <table border="0" cellpadding="4" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" id="table36"> <tr> <td width="100" align="left" valign="top"> <map name="FPMap285"> <area target="_blank" href="http://www.riskbook.com/link/rubinstein_r_(1981).htm" shape="rect" coords="0, 0, 83, 124"> </map> <img border="0" src="../images/book_rubinstein_125.gif" usemap="#FPMap285" width="84" height="125"></td> <td align="left" valign="top"> <table border="0" cellpadding="1" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" id="table37"> <tr> <td width="90%" colspan="3"> <p class="title"> <a target="_blank" href="http://www.riskbook.com/link/rubinstein_r_(1981).htm"> Simulation and the Monte Carlo Method </a></td> </tr> <tr> <td colspan="3"> <p class="author"><b>Reuven Y. Rubinstein</b></td> </tr> <tr> <td width="10%"> <map name="FPMap278"> <area target="_blank" href="http://www.riskbook.com/main/ratings.htm" shape="rect" coords="0, 0, 89, 15"> </map> <img border="0" src="../images/star_5.gif" usemap="#FPMap278" width="90" height="16"></td> <td width="10%"> <map name="FPMap198"> <area href="http://www.riskbook.com/main/ratings.htm" shape="rect" coords="0, 0, 89, 15" target="_blank"> </map><img border="0" src="../images/tech_3.gif" usemap="#FPMap198" width="90" height="16"></td> <td width="70%"> <p class="author"> </td> </tr> <tr> <td width="10%"> <p class="author">1981</td> <td width="10%"> </td> <td width="70%"> <p align="center"><map name="FPMap37"> <area target="_blank" coords="0, 0, 145, 19" shape="rect" href="http://www.riskbook.com/link/rubinstein_r_(1981).htm"> </map> <img border="0" src="../images/pointer_review.gif" usemap="#FPMap37" width="146" height="20"></td> </tr> </table> </td> </tr> </table> <table border="0" cellpadding="4" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" id="table38"> <tr> <td width="100" align="left" valign="top"> <map name="FPMap286"> <area target="_blank" href="http://www.riskbook.com/link/fishman_g_(1996).htm" shape="rect" coords="0, 0, 92, 124"> </map> <img border="0" src="../images/book_fishman_125.gif" usemap="#FPMap286" width="93" height="125"></td> <td align="left" valign="top"> <table border="0" cellpadding="1" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" id="table39"> <tr> <td width="90%" colspan="3"> <p class="title"> <a target="_blank" href="http://www.riskbook.com/link/fishman_g_(1996).htm">Monte Carlo</a><br> <font size="2"> <a target="_blank" href="http://www.riskbook.com/link/fishman_g_(1996).htm">Concepts, Algorithms and Applications</a></font></td> </tr> <tr> <td colspan="3"> <p class="author"><b>George S. Fishman</b></td> </tr> <tr> <td width="10%"> <map name="FPMap199"> <area target="_blank" coords="2, 1, 89, 15" shape="rect" href="http://www.riskbook.com/main/ratings.htm"> </map> <img border="0" src="../images/star_6.gif" usemap="#FPMap199" width="90" height="16"></td> <td width="10%"> <map name="FPMap202"> <area target="_blank" href="http://www.riskbook.com/main/ratings.htm" shape="rect" coords="0, 1, 89, 15"> </map><img border="0" src="../images/tech_4.gif" usemap="#FPMap202" width="90" height="16"></td> <td width="70%"> <p class="author"> </td> </tr> <tr> <td width="10%"> <p class="author">1996</td> <td width="10%"> </td> <td width="70%"> <p align="center"><map name="FPMap281"> <area target="_blank" coords="0, 0, 145, 19" shape="rect" href="http://www.riskbook.com/link/fishman_g_(1996).htm"> </map> <img border="0" src="../images/pointer_review.gif" usemap="#FPMap281" width="146" height="20"></td> </tr> </table> </td> </tr> </table> <table border="0" cellpadding="4" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" id="table40"> <tr> <td width="100" align="left" valign="top"> <map name="FPMap287"> <area target="_blank" href="http://www.riskbook.com/link/glasserman_p_(2003).htm" shape="rect" coords="0, 0, 92, 141"> </map> <img border="0" src="../images/book_glasserman_p_(2003)_125.gif" usemap="#FPMap287" width="93" height="142"></td> <td align="left" valign="top"> <table border="0" cellpadding="1" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" id="table41"> <tr> <td width="90%" colspan="3"> <p class="title"> <a target="_blank" href="http://www.riskbook.com/link/glasserman_p_(2003).htm"> Monte Carlo Methods in Financial Engineering </a></td> </tr> <tr> <td colspan="3"> <p class="author"><b>Paul Glasserman</b></td> </tr> <tr> <td width="10%"> <map name="FPMap288"> <area target="_blank" coords="2, 1, 89, 15" shape="rect" href="http://www.riskbook.com/main/ratings.htm"> </map> <img border="0" src="../images/star_6.gif" usemap="#FPMap288" width="90" height="16"></td> <td width="10%"> <img border="0" src="../images/tech_5.gif" width="90" height="16"></td> <td width="70%"> <p class="author"> </td> </tr> <tr> <td width="10%"> <p class="author">2003</td> <td width="10%"> </td> <td width="70%"> <p align="center"><map name="FPMap284"> <area target="_blank" coords="0, 0, 145, 19" shape="rect" href="http://www.riskbook.com/link/glasserman_p_(2003).htm"> </map> <img border="0" src="../images/pointer_review.gif" usemap="#FPMap284" width="146" height="20"></td> </tr> </table> </td> </tr> </table> </td> </tr> </table> </td> </tr> <tr> <td height="30" valign="bottom" colspan="2"> <p class="header2_black">Related Papers</td> </tr> <tr> <td colspan="2" bgcolor="#CC9966" height="2"> <img border="0" src="../images/blank_2x2.gif" width="2" height="2"></td> </tr> <tr> <td bgcolor="#FFFFCC" colspan="2"> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="6" id="table42"> <tr> <td width="100%"> <p class="explanatory_text"> <img border="0" src="../images/bullet_paper_glossary.gif" width="21" height="15"><a name="Eckhardt">Eckhardt</a>, Roger (1987). Stan Ulam, John von Neumann, and the Monte Carlo method, <i>Los Alamos Science</i>, Special Issue (15), 131-137.<p class="explanatory_text"> <img border="0" src="../images/bullet_paper_glossary.gif" width="21" height="15"><a name="Metropolis">Metropolis</a>, Nicholas and Stanislaw Ulam (1949). The Monte Carlo method, <i>Journal of the American Statistical Association</i>, 44 (247), 335-341.<p class="explanatory_text"> <img border="0" src="../images/bullet_paper_glossary.gif" width="21" height="15"><a name="Student (1908a)">Student (1908a)</a>. The probable error of a mean. <i>Biometrika</i>, 6, 1-25. </p> <p class="explanatory_text"> <img border="0" src="../images/bullet_paper_glossary.gif" width="21" height="15"><a name="Student (1908b)">Student (1908b)</a>. Probable error of a correlation coefficient. <i> Biometrika</i>, 6, 302-310. </td> </tr> </table> </td> </tr> <tr> <td height="30" valign="bottom" colspan="2"> <p class="header2_black">Sponsored Links</td> </tr> <tr> <td colspan="2" bgcolor="#CC9966" height="2"> <img border="0" src="../images/blank_2x2.gif" width="2" height="2"></td> </tr> <tr> <td bgcolor="#FFFFCC" width="75"> <img border="0" src="../images/ad_side_banner_3.jpg" width="103" height="285"></td> <td bgcolor="#FFFFCC" width="99%">  <p class="ad_legend" style="margin-left: 0; margin-bottom:2"> <a target="_top" href="http://www.riskexpertise.com/link/advertising.htm"> <font color="#000000">Ads by Contingency Analysis</font></a></p> <p style="margin-top: 0; margin-bottom: 0"> <iframe src="http://servedbyadbutler.com/adserve/;ID=147917;size=336x265;setID=82436;type=iframe" style="margin: 0px; padding: 0px; border-width: 0px" width="336" height="265" frameborder="0" marginwidth="0" marginheight="0" scrolling="no" name="I5"> <![if lt IE 4]> <script src="http://servedbyadbutler.com/adserve/;ID=147917;size=336x265;setID=82436;type=js" type="text/javascript"> </script> <noscript> <a href="http://servedbyadbutler.com/go2/;ID=147917;size=336x265;setID=82436"> <img src="http://servedbyadbutler.com/adserve.ibs/;ID=147917;size=336x265;setID=82436;type=img" border="0" height="265" width="336" alt=""></a> </noscript><![endif]></iframe>   </td> </tr> <tr> <td height="30" valign="bottom" colspan="2"> <p class="header2_black">Related Forum Discussions</td> </tr> <tr> <td colspan="2" bgcolor="#CC9966" height="2"> <img border="0" src="../images/blank_2x2.gif" width="2" height="2"></td> </tr> <tr> <td bgcolor="#FFFFCC" colspan="2"> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="6" id="table43"> <tr> <td width="100%"> <p class="explanatory_text"> <img border="0" src="../images/bullet_bullhorn_narrow.gif" align="left" width="27" height="22"><a target="_top" href="http://www.contingencyanalysis.com/archive/link_thread/03-1/00002350.htm">Generating joint distributions</a> <i>22 Jan 2003</i><br> Generating random vectors whose marginals are not normal.</td> </tr> </table> </td> </tr> <tr> <td bgcolor="#FFFFCC" colspan="2"> <table border="0" cellspacing="0" style="border-collapse: collapse" bordercolor="#111111" width="100%" cellpadding="6" id="table44"> <tr> <td width="100%" bgcolor="#FFFFFF">  <p align="center"><map name="FPMap0_I1"> <area target="_top" href="http://www.riskglossary.com" shape="rect" coords="0, 0, 164, 102"> </map><img border="0" src="../images/logo.jpg" usemap="#FPMap0_I1" width="165" height="103"></p> <p align="center"><span class="verdana10"> <a href="http://www.riskglossary.com/main/disclaimer.htm">Disclaimer</a></span></p> <p align="center">website: <span class="verdana10"> <a target="_top" href="http://www.contingencyanalysis.com">http://www.contingencyanalysis.com</a></span><br> glossary direct link: <span class="verdana10"> <a target="_top" href="http://www.riskglossary.com">http://www.riskglossary.com</a></span><br> copyright © Contingency Analysis, 1996 - current</p> <p align="center"><map name="FPMap1_I1"> <area target="_top" href="http://www.contingencyanalysis.com" shape="rect" coords="0, 0, 240, 88"> </map><img border="0" src="../images/logo_glossary_white.gif" usemap="#FPMap1_I1" width="241" height="89"></p> </td> </tr> </table> </td> </tr> </table> </td> <td>  </td> </tr> </table> </body>