Strategy branchpoint

We can continue both strategies by disconnections at the branchpoint, each needing simple aryl ketones 16 and 18, but 15b requires a homoenolate reagent for the d3 synthon 19 and we should rather avoid that, while 15c needs a simple enolate 17 and we prefer that. 

But this strategy is doomed too as the allylic bromide will almost certainly react at its less hindered end on the vinyl group. However, we might choose the alternative branchpoint disconnection 38a and consider conjugate addition of some vinyl-metal (copper ) derivative to the enone 41 that could be made by some aldol process from the ketone 42 and an enol(ate) of acetone 43. 

Disconnection of the bicyclic diketone 21 starts reasonably to reveal 22 but the two carbonyl groups are now 1,6-related suggesting a reconnection strategy (chapter 27). But this is impossible as the bridgehead alkene 23 is too strained to exist. However, if we extrude the carbon atom in the seven-membered ring between the ketone and the branchpoint 22a, we get a new ketoester 24 with a 1,5 relationship that can be made by conjugate addition (chapter 21). 

An optimal strategy will have to balance the cost of taking unnecessarily short steps against the cost of having to backtrack to correct for an overly long steps that jump past multiple branchpoints. 

---