I agree the superior measured edge retention of N690 vs 154CM is somewhat unexpected. The test is largely the result of hardness of the steel and the hardness and volume of its carbides when the other factors are kept constant. It's weird that the chart lists both steels as 61+. Does that mean almost 62? Why the plus? While cobalt can strengthen the matrix, this difference would likely be irrelevant when the steels are tempered to the same hardness.
To look at the potential difference in carbide hardness and volume, I did thermodynamic calculations with JMatPro of the carbide content of N690, N690 without cobalt, and 154CM. Here are the results:
N690 without cobalt:
2000F - 12.47% M23C6
1960F - 13.3% M23C6
1920F - 14.28% M23C6
N690:
2000F - 12.72% M23C6
1960F - 13.54% M23C6
1920F - 14.28% M23C6
154CM:
2000F - 11.07% M23C6
1960F - 12.18% M23C6
1920F - 13.2% M23C6
A couple of caveats: 1) These are models not experimental measurements. There is no guarantee they correlate with reality. 2) These are calculations for thermodynamic equilibrium, something we never actually reach in reality. Often the carbides present in these high alloy steels are formed at high temperatures during casting and never dissolve. In other words, they are not the carbides that would be calculated thermodynamically.
With that out of the way, we can analyze the results. N690 has around 13% more carbide volume than 154CM at the same austenitizing temperature. Furthermore, since N690 datasheets recommend lower austenitizing temperatures than 154CM, it is possible that it has even more carbide volume. The 13% volume difference with 154CM is pretty close to the 17% increase found in the CATRA testing. The thermodynamic calculations predict very little difference in carbide volume with the cobalt addition. Therefore, we would expect little difference in CATRA testing if N690 was made without cobalt, and differences between N690 and 440C are probably primarily due to differences in Cr and Mo. To test this conclusion, I also ran the thermodynamic calculations for N695 (440C):
2000F - 6.16% M7C3
1960F - 6.92% M23C6, 3.37% M7C3
1920F - 11.65% M23C6, 0.88% M7C3
As expected, the carbide volume of 440C is significantly lower.
In conclusion, thermodynamic calculations support the result that N690 has superior edge retention to 154CM. There are always concerns with testing, of course, we don't know how much statistical variation there is in the CATRA test or if multiple runs were used to give us an average result. Superior edge retention does not always mean a superior steel, of course. And superior CATRA results do not always mean greater edge retention depending on the cutting application.