I have had the same problem with a Jonsered 2050 and confirmed that it was NOT a timing issue. Timing at pull over speed was 23deg BYDC (same as a swapped coil from a similar problem free Jonsered) and that should not present a kickback when starting. Only conclusion I could come to was the problem might be related to excessive carbon build up so I gave up trying to figure it out and just retarded the timing by 5deg to 18deg BTDC. Problem went away and still seems to run normal at high rpm under load.
Beware when you start modifying the timing, you have to use a timing light to find out where you are before making any changes. Guessing and trial and error is not good. When you start relocating the position of the FW on the crank you will lose the relationship between it and the piston, so you will have to do the timing from the clutch side. You need a sector from a degree wheel (easy to make) and you have to find piston TDC accurately.
If you are modifying the timing to make a replacement coil work and the coil has way too much retard for starting purposes as compared to the existing coil, any advance you make to the timing at starting speed will also show up as the same amount of advance at top rpm. For stock chainsaws the maximum advance is 30deg, and your original coil is probably mounted to provide close to that, so if you advance the timing at starting speed by 5deg, you might be setting the high rpm by 35deg. Check with a timing light before doing any serious cutting or engine damage could result.