I know Uemura reported success with the dibromo->bromohydrin->epoxide->ketone route, but did anyone ever get the Yuki method of treating the dichloro or dibromo compound with 15% KOH to work?

Yuki et al, Japanese Patent 69:10,776 Chemical Abstracts 71, 30220e (1969)
23.2 g (0.1 mole) of 1-(3,4-methylenedioxyphenyl)-1,2-dichloropropane is refluxed for 10 hrs with 75g of 15% KOH (potassium hydroxide) solution. The mixture is cooled extracted with benzene. The benzene is removed in vacuum and the residue distilled to yield approximately 15.2 g of piperonylacetone (bp 149-151?C/10 mm Hg). 85% yield from the dichloro compound.

Ritter reported only tar with aqueous KOH but suggested ethanolic hydroxide would work and WizardX has suggested 50% aqueous/alcoholic or ethanolic KOH, but I don't think there were actually any reports of success with alcoholic KOH.

It always seemed like such a simple procedure (aside from working with demon chlorine gas or demon bromine liquid).

I'm inclined to believe Ritter's multiple tar reports; he had proven credibility.  I can see the dihalo compound forming the 1,2-propanediol upon KOH treatment, but I don't see how it would undergo a pinacol rearrangement (the typical acid hydrolysis) to the ketone or otherwise form the ketone under basic conditions.  Maybe the patent forgot to mention that the benzene extraction and distillation actually gave the 1,2-propanediol and an acid hydrolysis is necessary to produce the ketone?

Halohydrines are formed from the reaction of a Halogen in a basic media acting on an alkene, I wounder why anyone would chose to form the Dihalo compound first instead of proceeding directly to the halohydrin.