I believe that's what some people do.

The 4050 has a higher output sink (logic low) current capability than the 4514, but that's not important in this application.
The output source (logic high) current capabilities between the 4050 and 4514 isn't that much different (4050 spec is a little higher). So likely someone at Bally/Stern realised the 4514 decoder was capable of driving the SCR gate on its own and they removed the 4050 buffer chips from the circuit.

I think I would leave the 4050 in there.
Way back when these boards were made - the 4514's were dirt cheap and adding an output driver which provides a layer of protection to them had little return.
Today - 4514's are obsolete and getting expensive.
Add that 4050, it has a chance that it may protect the 4514 in the event of a catastrophe.

fwiw the 4514 decoders almost never fail in this circuit. Back when I did repair work I probably replaced less than 5 decoders out of hundreds of boards yet would go through tons of blown SCRs. 4050 are common and cheap tho. If just one circuit in the 4050 is bad could snip the bad legs off and stradle a zero ohm resistor or do something like the data buffers on WMS sys 6 where the entire chip is just jump'd input to out.

Odd thing is the aux lamp driver boards with 4013/4555 was pretty common to have those logic chips go bad.

TI is still making CD4514BM which is a SO24 surface mount version in the same pinout. You can use the SO24 to DIP24 mod boards in a pinch if the dip versions go completely obsolete / OOS.