Never thought about that and never done that. Interesting. I guess that your desired full-length band will not be generated if a regular gene is an aim. However, it can also depend on what DNA fregment you are amplifying. For example, if you have two primers flanking a 200-bp 'A' region (AAAAAAAAAAAAAAAAAAA......200A), then you should still be able to amplify this fragment if dTTP and dATP are included in your added 3 dNTPs.

I would expect the polymerase to try hard to incorporate the wrong bases and for the reaction to fail if there is any reasonable length amplimer between the primers ( except for Yuan- Yeu Lau's example which should work)

The only other amplimer which might work is a single missing base between 2 primers ( excluding the 3 in your mixture) when I think that a wrong base will sometimes be incorporated at the missing base position but I have no evidence for this other than when doing denaturing HPLC of a single pcr product we see a mound of pcr misincorporations in the amplimer so it is clear that polymerases are able to make mistakes

there may be no amplification except if what you have repeating nucleotides of what you have added in the flanking regions

never do this test. It will waste your time and money.

You won't get your fragment amplified except for the short stretch until polymerase runs into the missing nucleotide. You might get occasional disincorporation of a base at the first or even the second missing nucleotide position but the probably of elongating beyond that is miniscule.

You will not get any amplicon