I show you something:
this is a graph about the efficiency of three my tops.
This efficiency is calculated as the percentage of RPM retained after one minute of spinning.
For example, these are the data of the first five minutes of spinning of my new record:
Minutes RPM Efficiency
0 2917
1 2759 94.6%
2 2607 94.5%
3 2469 94.7%
4 2339 94.7%
5 2220 94.9%
The calculation is made in this way:
2759 : 2917 = 94.6, percentage of RPM retained after the first minute of spinning.
2607 : 2759 = 94.5, the same for the second minute of spinning,
and so on.
These data, (efficiency vs. RPM), plotted in a graph help comparing the performances of different tops.
I start comparing these three old tops, (then I will add the data of the new tops):
the most efficient one of these three is the biggest one, the Nr. 9.
The least efficient one is the littlest one, the Nr. 14.
At parity of starting speed and toppling down speed the most efficient top would be the one which spins longer.
But heavier tops are more difficult to spin fast so they are started at a lower speed; the Nr. 9 is so heavy that I couldn't even reach 1000 RPM starting speed with it, and the best spin of this top is only 41 m 25 s, in spite of its excellent efficiency, (this top also has an external tip so it topples down soon, at about 135 RPM, this too shortens the spin; the external tip does not deteriorate the efficiency of the top, in any case).
On the other hand, the Nr. 14 is the lighter top, the easiest to spin fast.
I can start this top at 2760 RPM by multiple twirls, many more than with the Nr. 9.
But being littler/lighter/with less AMI, its efficiency is lower,and slows down more rapidly. Its best spin is 46 m 15 s.
The Nr. 23 has an efficiency in between. In spite of being quite heavier/bigger than the Nr. 14, I can start it at almost the same RPM, (2717); so it has the best compromise between efficiency and starting speed, and its best spin by multiple twirls is 58 m 19 s.
The problem with the Nr. 14 is that there is a limit in the speed of the fingers. Whatever the weight of the top, I never have been able to spin a finger top to much more than 3000 RPM. So, reducing the weight/size/AMI of the top, over a certain limit, will reduce its efficiency without the advantage of a significant faster starting speed, and the spin times will be reduced.