Suppose for example the gains between bottoms are:

91.31x, 16.11x, 4.97x

We divide each by the consecutive to get the rate of change in growth:

( 91.31/16.11, 16.11/4.97 ) = ( 5.67, 3.24 )

This allows us effortlessly to find the next number (n=2), assuming the net growth will approach zero (i.e. a factor of 1) eventually:

1+4.67*(2.24/4.67)**n = 2.07

Now we can calculate the growth rate backwards:

4.97/2.07 = 2.40x

Those steps can then be repeated ad infinitum.

The chart is not absolute. You can drag the historic top and bottom points around to match them to whatever you believe to be a more accurate measurement than the selectable SMA range. You can match the points to the power law function (drag the first bottom lower) and mirror it fairly accurately. See what it does for the growth of the bottom-to-top gains. See how solid the next bottom prediction is, or how easily the model fails. There are many possibilities here.

I hope you have fun working with the graph and it answers some questions.

Please note that the date of bottoms and tops is NOT predicted, but simply repeated from the previous, or current all-time high.
The points only drag vertically, because the date has no effect on the calculation.

SMA 14 days for bottoms has been preset, because there seems to be a more stable valley at that value, and it seems reasonable regardless.
SMA 55 days has been preset for the tops, to maximize alignment of the predicted to the observed SMA-55 top value of $115,975.04 on 2025-08-02.

The date of the predicted tops is now offset, according to the offset produced by the observed SMA top value in 2025.

Naturally this is not financial advice, and no guarantees for correctness.