The IN-4 & IN-18 can be used in the same socket. Physical pin spacing is the same (14-pins, on a 18mm diameter pin circle, 24.5 degrees apart). Pin-out assignments is different, between the two.
Pin # IN-4 IN-18
==== ==== ====
. 1 . . . K4 . . N/C
. 2 . . . K6 . . K4
. 3 . . . K8 . . K5
. 4 . . .SCR* Anode
. 5 . . . K9 . . K6
. 6 . . . K7 . . K7
. 7 . . . N/C . K3
. 8 . . . K0 . . N/C
. 9 . . . K2 . . K8
.10. . . A2* . K2
.11. . . K3 . . K1
.12. . . K5 . . Anode
.13. . . A1*. . K0
.14. . . K1 . . N/C
*the IN-4 has two different configurations. One has two anodes (A1 & A2), and separating screen, allowing it to be operated as a bi-quinary nixie, in addition to a "traditional" nixie. The other configuration omits the screen and an anode. It only has one (A1), and can only be operated in the traditional mode.
I just plugged an IN-4 tube into a IN-18 socket, and it fits without issue. So this is NOT a guess, but CONFIRMED.
Though, I would suggest against the IN-4 for a nixie clock, as it does not contain mercury, and therefore has a short life. I found that out the hard way. My first nixie clock used two IN-4s paired with two ZM1032 (biquinary) nixies:
The 8051 uC drove all 4 as a multiplexed biquinary group. 2x4=8 anode drives, and 5-cathode pairs. 8x5. More complicated than it needed to be if I just used regular nixies. Biquinary only has an advantage when run by descrete circuitry. ICs made them obsolete.
Well the short life of the IN-4, caused the zero to disappear in less than a year. Mercury is a good thing in nixies (IN-8-2, IN-12, IN-14, IN-16, IN-17, & IN-18).