Overfull hbox when inserting math

Question

I have tried to make a simple table for showing my confidence intervals and p-values (Default just inserted for now). However, when I insert math I get the error for overfull hbox. Can anyone help me how to deal with this?

\begin{table}[!htp]
\centering

\sisetup{detect-all}
\NewDocumentCommand{\B}{}{\fontseries{b}\selectfont}

\begin{tabular}{
  @{}
  S[table-format=1.2]
  S[table-format=-1.2]
  S[table-format=1.2]
  S[table-format=1.2]
  S[table-format=1.2]
  S[table-format=1.2]
  @{}
}
\toprule
\multicolumn{2}{c}{\textbf{Baseline vs. ANN}} & \multicolumn{2}{c}{\textbf{Baseline vs. Logistic Regression}} & \multicolumn{2}{c}{\textbf{Logistic Regression vs. ANN}} \\
\cmidrule(lr){1-2} \cmidrule(lr){3-4} \cmidrule(l){5-6}
{p-value} & {95 \% conf. int.} & {p-value} & {95 \% conf. int.} & {p-value} & {95 \% conf. int.} \\
\midrule
{$1.0 \cdot 10^{-17}$} & {$[0.31, 0.44]$} & {$1.0 \cdot 10^{-17}$} & {$[0.31, 0.44]$} & {$1.0 \cdot 10^{-17}$} & {$[0.31, 0.44]$} \\
{$1.0 \cdot 10^{-17}$} & {$[0.31, 0.44]$} & {$1.0 \cdot 10^{-17}$} & {$[0.31, 0.44]$} & {$1.0 \cdot 10^{-17}$} & {$[0.31, 0.44]$} \\
{$1.0 \cdot 10^{-17}$} & {$[0.31, 0.44]$} & {$1.0 \cdot 10^{-17}$} & {$[0.31, 0.44]$} & {$1.0 \cdot 10^{-17}$} & {$[0.31, 0.44]$} \\
\bottomrule
\end{tabular}

\caption{Models and metrics (triples of data)}

\end{table}

Answer 1

The overfull lines are not caused by the mathy material. Instead, they are caused by the header cells, which simply contain too much material. The following screenshot contrasts the looks of the OP's table and of a possible alternative which does fit within the default-width text block.

\documentclass{article}
\usepackage{booktabs}
\begin{document}

\begin{table}[!htp]
\centering

%% first, the OP's code
\begin{tabular}{@{} cccccc @{}}
\toprule
\multicolumn{2}{c}{\textbf{Baseline vs. ANN}} & \multicolumn{2}{c}{\textbf{Baseline vs. Logistic Regression}} & \multicolumn{2}{c}{\textbf{Logistic Regression vs. ANN}} \\
\cmidrule(lr){1-2} \cmidrule(lr){3-4} \cmidrule(l){5-6}
{p-value} & {95 \% conf. int.} & {p-value} & {95 \% conf. int.} & {p-value} & {95 \% conf. int.} \\
\midrule
$1.0 \cdot 10^{-17}$ & $[0.31, 0.44]$ & $1.0 \cdot 10^{-17}$ & $[0.31, 0.44]$ & $1.0 \cdot 10^{-17}$ & $[0.31, 0.44]$ \\
\bottomrule
\end{tabular}

%% now, a possible alternative
\bigskip\bigskip
\begin{tabular}{@{} cccccc @{}}
\toprule
\multicolumn{6}{c}{Pairwise Regression Comparisons}\\
\midrule
\multicolumn{2}{@{}c}{Baseline vs.\ ANN} & 
\multicolumn{2}{c}{Baseline vs.\ Logistic} & 
\multicolumn{2}{c@{}}{Logistic vs.\ ANN} \\
\cmidrule(r){1-2} \cmidrule(lr){3-4} \cmidrule(l){5-6}
$p$-value & 95\% CI & 
$p$-value & 95\% CI & 
$p$-value & 95\% CI \\
\midrule
$1.0 \cdot 10^{-17}$ & $[0.31, 0.44]$ & $1.0 \cdot 10^{-17}$ & $[0.31, 0.44]$ & $1.0 \cdot 10^{-17}$ & $[0.31, 0.44]$ \\
\bottomrule
\end{tabular}

\end{table}
\end{document}