Yeah, having tried this out and found it to be impossible and unnecessary I reckon it's best to leave these ancient, conservative, well-established old lineages unsplit family-wise. That being said I would like to see some deep divisions recognised at subfamily level with 2 in Formicariidae and maybe 4 in Rhinocryptidae. Not sure if there are any suitable names available and to be honest I haven't had enough coffee yet to embark on that kind of hunt!
Is it worth making two monotypic subfamilies for Formicarius and Chamaeza? The latter probably needs a new family name to be described, as it's previously been in Formicariinae and Formicariini before all the splitting.
The morphological analysis of Rhinocryptidae by Maurício et al (2012; pdf), while not consistent with the molecular analyses (according to TiF), defines a number of subfamily, tribe and subtribes that could be useful.