\documentclass{report}
\setlength{\parskip}{10pt}
\setlength{\parindent}{0pt}
\usepackage{amssymb}
\begin{document}

Joey, and other co-authors on the jet paper,

Here are the values of $\chi^{2}/N$ for the
D\O and CDF jet data, for the cteq6m PDF's
and the up/down PDF's of eigenvector 15.

\begin{center}
\begin{tabular}{|c|c|r|r|r|r|}
\hline \hline
file name & PDF set & \multicolumn{2}{|c|}{D0 data}
 & \multicolumn{2}{|c|}{CDF data}
\\ \hline
 & & $\chi^{2}/N$ & N.F. & $\chi^{2}/N$ & N.F. 
\\ \hline
cteq6m0	& cteq6m & $0.718$ & $0.974$ & $1.47$ & $1.007$
\\ \hline
jet100 & cteq6m & $0.724$ & $0.974$ & $1.47$ & $1.007$
\\ \hline
jet129 & EV29 & $1.25$ & $0.938$ & $1.43$ & $0.999$
\\ \hline
jet130 & EV30 &	$0.389$ & $1.035$ & $1.72$ & $1.008$
\\ \hline\hline
\end{tabular}
\end{center}

The calculations were done by Liang.

Comments
\begin{itemize}
\item{Fits cteq6m0 and jet100 are really the same.}
\item{The normalization factors (N.F.) were recalculated,
i.e., {\em optimized} for each PDF set.
So each fit is the best fit to the given PDF set.
This seems to be the right comparison to make
if we regard the 40 EV sets as alternate
hypotheses for the PDF's.
In other words, we are testing whether that hypothesis
will fit the data, allowing the overall normalization
of the data to vary.}
\item{EV29 is the displacement along eigenvector 15
in the direction such that the gluon is soft,
i.e., has smaller density at large $x$ than cteq6m.}
\item{EV30 is the opposite displacement along eigenvector 15,
such that the gluon is hard,
i.e., has higher density at large $x$.
[In fact, it has a big bump for $x \gtrsim 0.5$
(for low $Q$).]}
\end{itemize}

This is the question:
Are the up/down eigenvector sets 29 and 30
reasonable alternate fits?

Looking at the $\chi^{2}$ values may produce
more heat than light in this case.
Maybe we need to go the next step and look at the
comparisons graphically.

Dan 

\end{document}