HICKMAN, Rickmansworth, 1780-90
I'm afraid I haven't got a magic answer.
Clearly one has to be careful, but as you go back you will eventually reach a point where either there is not information, or 100% proof is impossible, and one had to work with a degree of probability. (For instance if a marriage takes place and 25 years earlier someone is baptised in the same village with the same name it is only very likely that the two references apply to the same individual - see Right Name, Wrong Body?) I have a "gap" in my own family tree where someone turned up in Wiltshire could be a child born in Lancashire. There is no direct evidence - but some very strong clues such as very uncommon name and right age, same occupations, similar family Christian names, the Lancashire son was alive but unable to execute his mother's will (because he was too far away?), Lancashire family definitely mobile, and marriages in Westminster. My guess (because I haven't yet tried to find if suitable records exist) is that the common factor may prove to be employment by the Duke of Devonshire. Because of the strength of the circumstantial evidence I suspect it is far more reliable than anything simply based on parish registers.
There are two related approached, and how well they work in your case depends on the individual, their social status, and whether relevant records have survived for the towns or villages of interest.
The first is to build a picture of all the individuals who could be ancestors from available resources. Did they leave wills? Were they on poor relief? Are they mentioned on the militia lists? Were they copyhold tenants in the local manor, etc. This is likely to show you how the individuals fitted into society in the places they lived. The search may come up with some definite statements about relationships - or at least suggest which individuals are the most likely ancestors,
The second is to explore related branches of the family because cousin marriages were common and can suggests possible liaisons. In addition it may turn out that a cousin with no heirs has left money to relatives in a will - and hence provide information about the identity of individuals on your ancestral tree. See Who is Related to Who? which gives some examples of such links.
If you can add to the information given above tell me.
Link updated January 2008