The (masculine) plural genitive of both the participle victus and the derived noun victor is victorum. If I write, for example, uxores victorum infelices erant, it is unclear which wives were unhappy. The word vincere is not special; the same happens with every verb that has the perfect participle and the derived noun.
The Romans must have observed this ambiguity. Are there any known ancient word plays or confusions due to it? I do not know how to search for such a thing in a corpus. The verb need not be vincere.
What you mention is technically known as "homonymy", and it was theoretically observed and practically used by the ancients. There are different types and many examples can be given: Cicero plays with the homonymy of the proper name Verres (the corrupt governor of Sicily whom Cicero attacked in the Verrines) and the noun uerres 'male-pig'. As for the theoretical side, you can fruitfully read a passage from Quintilian's Institutio oratoria (book 7, chapter 9), where he, when dealing with ambiguity, mentions several types of homonymy and homography, e. g. Coruinum (proper name) / cor uinum (two nouns) / Gallus (man from Gaul) / gallus (cock) / gallus (priest of Cybele), in occulto loco "in a hidden place" / inocculto loco "a non-hidden place'... and many more (he also refers to the famous aio te, Aeacica, uincere Romanus...). This is not the only passage where Quintilian deals with the liguistic problem of homonymy and polysemy: in book one, he mentions the rare homonymy between fraudator ('a deceiver') and fraudator as an imperative passive form (very rarely used, indeed!) of the verb fraudo There is a lot of literature dealing with the topic of puns based on homonymy. I can try to select some references for those who may be interested-