How do I decide two signatures match?

Question

This image shows two signatures made by "a" man named George Cox. One signature is on an Army Attestation Form when a boy of 15 signed-up in 1900. The other comes from a 1922 Migration Application by a man of 36.

Addition: I chose not to date the signatures because I thought it might bias your view. If I was wrong, see the comment below.

I have a WWI Service Number for the 1922 Cox which led to a Medal Card but his other records have not been found. (Many WWI records were destroyed in WWII bombing.)

My interest in the 1900 Cox arises from the fact that the 1922 Cox went from being an insurance salesman on 3 August 1914 to being part of the first mobilisation on 14 August and then the front line of the Battle of Mons on 23 August with the Royal Scots Fusiliers. I infer that this was unlikely unless he was a former Regular soldier. I then began a search for pre-war service records.

The 1900 form shows an age (in years and months) but no precise date of birth. That age is close, but not an exact match for a calculation from the d.o.b given in 1922.

I need to decide if the two signatures are sufficiently similar to justify further investigation of the 1900 Cox.

(As an aside, the 1900 Cox trained as an army bandsman and the 1922 Cox had strong connections to the Salvation Army. Proves nothing, but ...?)

Update

I have taken up the suggestion of @Rob Hoare to make a collection of signatures from the 1911 Census. Here are six of the approximately 1100 George Cox recorded in England and Wales.

The impact of rigid schooling on how people wrote their name is quite evident. Clearly matching the signature on the back of a credit card would have been much easier 100 years ago.

Answer 1

They certainly could be a match, but since the two signatures are from a person (or people) of a similar age from a similar place it may just be the way handwriting was taught in the schools of that place and time.

As this is a common name, you can use that to your advantage by collecting a large number of samples of the signature. An Ancestry (or Findmypast) search of the 1911 census for George Cox of around the right age turns up of many of them, and signatures are available for the ones who are head of households. You can also get signatures from further military records.

From that much larger set of signatures, you can get a better idea of what differences are possible. The less the differences exist in your two signatures above, the more likely it is to be the same person.

The way the capital G and C are written seems to be very common for the era, so that doesn't help much. The size of the loop on the lower case g, a gap (or not) between C and o, and the shape of the x, do vary more. That's just what I noticed with a small extra set of signatures, with more of them you could see more traits.

The link between the O and the X, and the shape of the X, are the main features of your two signatures that would make me think they are different people, but a larger sample would narrow that down (you might even find the 1922 George!).

Answer 2

I immediately thought they were different signatures, based on the lower-case g, r and x. However, I am by no means an expert. There is some basic advice here and here but I suspect you need a wider range of samples to distinguish features likely to be unique to an individual and features likely to derive from how handwriting was taught at a particular time.