According to my data a 4/1 chance has an average win strike rate of around 17%(no other factors taken into account). This would mean that the probability of backing 3 consecutive winners would be .17x.17x.17 which is about 0.5% or 1 in 200(assuming independence between results).

My data suggests that 125/1 chances win with a probability of .0025 which equates to 1 in 400(very small sample however).

So in answering your question i would say that the former is more likely to happen than the latter remembering the assumption has been made that 4/1 chances win less than 1 in 5 and 125/1 chances win less than 1 in 126 times.