circle with snake decoration is not centered

[Firestar-Reimu]

Brief outline of the bug

Minimal working example (MWE)

\documentclass[tikz]{standalone}
\usepackage{amsmath}

\usetikzlibrary{arrows.meta,decorations.pathmorphing}

\begin{document}
\begin{tikzpicture}[>=Stealth]
\draw[->,thick] (-3.5,0) -- (3.5,0) node[right] {$x$};
\draw[->,thick] (0,-3.5) -- (0,3.5) node[above] {$y$};
\draw[gray,very thin] (-3,-3) grid (3,3);
\draw[dashed,red] (0,0) circle (1.5cm);
\draw[red,thick,decorate,decoration={snake,amplitude=1pt,segment length=1cm}] (0,0) circle (1.5cm);
\end{tikzpicture}
\end{document}

you can change the amplitude or segment length, but overall the snake path is slightly higher than the circle

[muzimuzhi]

Looks like the snake decoration is direction sensitive:

\documentclass{article}
\usepackage{tikz}

\usetikzlibrary{arrows.meta, decorations.pathmorphing}

\begin{document}
\begin{tikzpicture}[decoration=snake]
  \draw[postaction={draw=red, decorate}]
    (0,0) rectangle (3,2);
\end{tikzpicture}

\begin{tikzpicture}[decoration=snake]
  \draw[postaction={draw=blue, decorate}]
    (0,0) -- ++(3,0);
  \begin{scope}[xshift=4cm]
    \draw[postaction={draw=blue, decorate}]
      (0,0) -- ++(3,0) -- ++(0,2);
  \end{scope}
  \begin{scope}[xshift=8cm]
    \draw[postaction={draw=blue, decorate}]
      (0,0) -- ++(3,0) -- ++(0,-2);
  \end{scope}
\end{tikzpicture}
\end{document}

[muzimuzhi]

Nope, for line-to subpaths, the "height" of a snake decoration is sensitive to the previous subpath length. But for circles, it's not affected by radium.

\documentclass{article}
\usepackage{tikz}

\usetikzlibrary{decorations.pathmorphing}

\begin{document}
\foreach \i in {0, 0.05, ..., 0.3} {
  \begin{tikzpicture}[decoration=snake]
    \draw[postaction={draw=blue, decorate}]
      (0,0) -- ++(1+\i,0) -- ++(0,2);
  \end{tikzpicture}
}

\foreach \i in {0, 0.05, ..., 0.3} {
  \begin{tikzpicture}[decoration=snake]
    \draw[postaction={draw=blue, decorate}]
      (0,0) circle[radius=.5cm+\i cm];
  \end{tikzpicture}
}
\end{document}

[hmenke]

This looks like a pretty tricky one. In the meantime I can suggest a solution using a parametric plot.

\documentclass{article}
\usepackage{tikz}
\begin{document}
\begin{tikzpicture}
  \draw[->,thick] (-3.5,0) -- (3.5,0) node[right] {$x$};
  \draw[->,thick] (0,-3.5) -- (0,3.5) node[above] {$y$};
  \draw[gray,very thin] (-3,-3) grid (3,3);
  \draw[dashed,red] (0,0) circle (1.5cm);
  \draw[red,thick] (0,0) plot[variable=\t,domain=0:360,samples=100,smooth]
    ({(1.5 + 0.1 * sin(5 * \t)) * cos(\t)}, {(1.5 + 0.1 * sin(5 * \t)) * sin(\t)});
\end{tikzpicture}
\end{document}

or using the complete sines decoration from the tikz-feynman package

\documentclass{article}
\usepackage{tikz}
\usepackage{tikz-feynman}
\begin{document}
\begin{tikzpicture}
  \draw[->,thick] (-3.5,0) -- (3.5,0) node[right] {$x$};
  \draw[->,thick] (0,-3.5) -- (0,3.5) node[above] {$y$};
  \draw[gray,very thin] (-3,-3) grid (3,3);
  \draw[dashed,red] (0,0) circle (1.5cm);
  \draw[red,thick,decorate,decoration={complete sines,amplitude=1pt,segment length=1cm}] (0,0) circle (1.5cm);
\
\end{tikzpicture}
\end{document}