Trying again...
Here's some equations for piston kinematics I had worked out some time ago. They might even be correct!
let c = conn rod length
let s = stroke
let r = ½s
let p = "rod ratio" = c/s
let Q = crank angle from TDC
let w = dq/dt = engine speed (radians/time, not rev/time)
let x = piston position from mid-stroke)
let y = du/dt = piston velocity
let z = d²u/dt² = piston acceleration
let a = 2p (twice rod ratio)
let b = (a² - 1 + cos²Q)^½
And first note that for infinitely long conn rod, piston motion approaches sinusoidal, where
max x = r
max y = rw
max z = rw² for constant engine speed (dw/dt = 0)
Actual motion for pistion (finite conn rod length) can then be expressed as ratio to max sinusoidal motion:
x/r = -a + b + cosQ
y/rw = -sinQ[1 + (cosQ)/b]
z/rw² = -(cosQ + [(sin²Q)/b]•[(cosQ)/b - 1])•[(cosQ)/b - 1]
These normalized motion equations approach sinusoidal for larger rod ratios, with max/min values of ±1. The amount that the normalized velocity and acceleration max/min values exceed ±1 represent the amplification due to non-infinite conn rod length. Again, engine acceleration was neglected in deriving normalized acceleration equation.
Note that maximum piston acceleration is amplified over sinusoidal max (aw²) by factor of (1 + 1/a) = (1 + ½/p). Trends pointed out by Norm above can be seen (what's with the Structurals hanging out here?....)
FWIW, I've heard max piston accelerations for near-current F1 engines as being around 10,000g, as mentioned above. Same source reported engine accelerations of 25,000rpm/sec.