Processing Power to Chess Rating Ratio

50 here is what the Deep Thought team wrote about the relationship between search depth and chess strength in a 1989 article:

The ascent of the brute-force chess machines back in the late 1970s made one thing crystal clear: there is a strong causal relationship between the search speed of a chess machine and its playing strength. In fact, it appeared from machine self-test games that every time a machine searches one extra ply, its rating increases by about 200-250 rating points. Since each extra ply increases the searched tree size by five to six times, every two-fold increase in speed roughly corresponds to an 80-100 rating point gain. Ratings obtained by machines against human players indicate that this relationship holds perhaps all the way up to the Grandmaster level where Deep Thought currently resides. The presence of this causal relationship was the reason the project was started in the first place.

In other words, faster means deeper and deeper means stronger, d that was all that really mattered. You can chart the progress of chess machines with rating as the y axis and the number of positions jarched per move as the x, and it makes for a nice diagonal line. StartJ with Chess 3.0 in 1970 at around a 1400 level, to Chess 4.9 at 2000 in 1978, Belle breaking 2200 in 1983, HiTech at 2400 in 1987, and Deep Thought at a Grandmaster-level 2500 in 1989. The chips get smaller and faster, the search goes deeper, and the rating rises.