It's an ease of use issue. Try jumpering your drives the same. Maybe the twist works ?
Actually, different standard.
Shugart had four drive select lines and one common motor line. If you notice, all motors spin on the TI when one does.
IBM didn't think the PC would handle four motors going at once, so it made different motor enable lines - in doing so, it changed the drive select lines from a quantity of 4 to 2, added two motor enable lines, and in the process, changed the pinout. Using the twist won't work on Shugart compliant controllers due to the pinout change. The IBM pinout keeps the drives jumpered as DS1, and uses the twist to make the second motor select and the DS0 signal controller go to the second drive's drive-enable. The drive thinks it's DS1, but it's really getting a DS0 on it's DS1 input pin, and it's getting the second motor select, not the first. Thus the 'motor 2' line after the twist will never activate the drive.
Another interesting tidbit is that IBM did offer two floppy headers so you could get all four drives if you really wanted, but then cut down to one header, thus, everyone on the PC side using two floppy drives max.