### Table 1: The multiplicity of heavy quark pairs at the Z mass in percent according to the various methods discussed in the text. The errors shown on the Monte Carlo results are purely statistical.

"... In PAGE 7: ... Also shown is the result of varying by a factor of two, which makes a 25% di erence, roughly constant with energy, and the bottom quark mass by 5%, which makes a 10% di erence at the Z mass, reducing with energy. Table1 lists the predictions for e+e? annihilation at the... ..."

### TABLE X OVERALL STATISTICS OF THE DIFFERENCE BETWEEN THE ADSL BIT RATE COMPUTED WITH NEXT SUMMATION METHODS AND THE ADSL BIT RATE COMPUTED BY THE MONTE CARLO METHOD ACROSS ALL SIMULATIONS

2002

Cited by 6

### TABLE X OVERALL STATISTICS OF THE DIFFERENCE BETWEEN THE ADSL BIT RATE COMPUTED WITH NEXT SUMMATION METHODS AND THE ADSL BIT RATE COMPUTED BY THE MONTE CARLO METHOD ACROSS ALL SIMULATIONS

2002

Cited by 6

### Table 3: Dimuons forward-backward asymmetries measured in Monte-Carlo dimuons for generated and selected events having qs0=s gt; 0.9, with the statistical error. The results are estimated using two methods : (1) a counting method and (2) a t method with likelihood maximisation

### Table 2. Monte Carlo Statistics

"... In PAGE 9: ...nclination is between 44.7 and 45.3 degrees. Table2 summarizes some statistics from the Monte Carlo simulation. 44.... ..."

### Table 4. Monte Carlo method

"... In PAGE 5: ...4615 0.4342 7175 430 500 Table4 . contains results obtained by standard Monte Carlo method with 10 million points, presented in [2].... ..."

### Table 1: Prices computed by alternative methods under the 2-factor GBM model

2000

"... In PAGE 13: ... 4.2 Computational Results Table1 documents the spread option prices across a range of strikes under the two factor Geo- metric Brownian motion model [22], computed by three di erent techniques: one-dimensional integration (analytic), the fast Fourier Transform and the Monte Carlo method. The values for the FFT methods shown are the \lower quot; prices, computed over , regions that approach the the true exercise region from below and are therefore all less than the analytic price in the rst column.... ..."

Cited by 5

### Table 2: Systematic errors for the impact parameter method. The above measured value for the lifetime was corrected for background which was taken to have a zero lifetime. This was checked by merging events from the various sources that compose the background with the Monte Carlo sample that was used in the t, with the proper normalization. The change in the lifetime was as expected within the statistical error. Including the systematic error, the lifetime measured with the impact parameter method is: = 293:7 8:2 4:6 fs:

1996

"... In PAGE 7: ...3% due to the beam energy measurement uncertainty and beam energy spread. These systematic errors are listed in Table2 , and when combined in quadrature result in an overall systematic uncertainty of 1.6%.... ..."

### Table 3. Bias and Monte Carlo variation in conventional PEA Parameterizations

1997

"... In PAGE 39: ....2.1. Conventional PEA Table3 provides information on the performance of conventional PEA in approximating the conditional expectation, exp [e(k, O)] , that is the focus of Marcet (1988) apos;s analysis. The results in Marcet and Marshall (1994) indicate that conventional PEA is arbitrarily accurate for sufficiently large M and N.... In PAGE 39: ... The question that interests us here is how well the algorithm works for the values of M and N used in practice. For the results in Table3 , we set M = 10,000. By way of comparison, to solve the growth model, den Haan and Marcet (1990) use M = 2,500, den Haan and Marcet (1994) use M = 29,000, and den Haan (1995) uses M = 25,000.... In PAGE 39: ...aphson method to solve a* - S(a*; N. M) = 0. When this method is successful at finding a solution, we found it does so more quickly than does the successive approximation method. The first three terms in each cluster of four numbers in Table3 provide information about bias. The unbracketed term is the value of the statistic, s, indicated in the first column implied by the dynamic programming solution.... In PAGE 40: ...The results in Panel A of Table3 pertain to various second moment properties of consump tion, investment, and output. Here, aj , j = y, c, i denote the standard deviation of gross output, consumption and gross investment, respectively, and p(y, j), j = c, i denote the correlation of gross output with consumption and gross investment, respectively.... In PAGE 40: ... Here, aj , j = y, c, i denote the standard deviation of gross output, consumption and gross investment, respectively, and p(y, j), j = c, i denote the correlation of gross output with consumption and gross investment, respectively. The results in Panel B of Table3 pertain to first and second moment properties of Tobin apos;s q and asset returns. The results in Panel A indicate that, at least for parameterizations (1)-(6), the conventional PEA performs reasonably well.... In PAGE 41: ...results are based on I = 50). These are reported in column 2 of Table 4 (column 1 simply reproduces the results from Table3 for convenience.) TO diagnose the reasons for the poor performance of conventional PEA for model (7), consider the results in Figure 4.... In PAGE 43: ...Approximating Marcet apos;s Conditional Expectat ion Function by PEA Collocation We applied PEA collocation to approximate e in all seven models, and obtained acceptable accuracy with N = M = 3 for models (1) to (6). By apos;acceptable apos;, we mean that all statistics analyzed in Table3 and 4 are within 10 percent of their exact values. We only study bias for this method, since Monte Carlo uncertainty is not applicable.... In PAGE 58: ...ith a value for a) to construct mz, ..., mso,ool in the manner described in step #l. These data were then used in the nonlinear regression specified in step #2. collocation N=3 N=5 (ii) The entries in the first column are reproduced from the last column in Table3 . There, I = 500, though 48 of these had to be discarded because capital converges to zero in simulation.... ..."

### Table 1: Monte Carlo Results

in Bo Honor'e

1998

"... In PAGE 17: ...he Buckley-James estimator is inconsistent when the errors are t(1)(i.e., Cauchy) distributed or heteroskedastic. The results in Table1 indicate that the estimation methods proposed here perform almost as well as the Buckley-James estimator under normality, and that the superiority of the latter disappears when the errors are nonnormal. As might be expected, the procedures proposed here, which do not impose homoskedasticity of the error terms, are superior to Buckley-James when the errors are heteroskedastic.... In PAGE 40: ...Table1 : Monte Carlo Results (continued) Standard Normal Buckley-James CLAD ( = 0:50) STLS True Values -1.000 1.... ..."

Cited by 5