@Vincent BergOkay, first off, I think you have me confused with DarkKnight, who, despite the similar name, is a completely different person.
Secondly โ and you would already know this if you had actually read my previous post rather than playing your favorite game, "CW leaps to conclusions and then vociferously defends his unwarranted assumptions in the face of all fact and reason" โ when I was talking about "strong" and "weak" verbs, it had nothing whatsoever to do with active and passive voice. I was talking about a system for classifying Old English verbs that describes how they were conjugated. This is still relevant in Modern English because some of those differences in conjugations have carried through, and, as LupusDei correctly observed, a lot of the verbs that get described as "irregular" really aren't. They're regular by another scheme that's buried under a thousand years of linguistic evolution, which your high school English teacher knows nothing whatsoever about.
So if we look at "dive", because we have both strong and weak versions of it โ dufan and dyfan, respectively:
The -an (or -ian for type 2 weak verbs, or occasionally -on, as in beon, mentioned above) is the Old English infinitive verb ending. If we lop that off, that leaves us with the verb stem.
So we lop the -an off dyfan, that leaves us with the verb stem dyf. Dyfan is a type 1 weak verb, so that verb stem doesn't change when it's conjugated. All that changes is the endings. Present singular 1st is dyfe, 2nd is dyfest, 3rd dyfeรพ, present plural is dyfaรพ, past singular is dyfde, past plural is dyfdon, present participle is dyfende (which is then further declined as an adjective), past participle gedyfed, and so on.
Dufan, on the other hand, is a type 2 strong verb. Lopping off the infinitive ending leaves us with the verb stem duf. However, because it's a strong verb, that verb stem morphs when the verb is conjugated. Present 1st singular is dufe, as you would expect, but then the 2nd person is dyfst and the 3rd dyfรพ, and the plural is back to dufaรพ. Then in the past tense, the 1st and 3rd singular are deaf, while the 2nd is dufe and the plural is dufon. Present participle is dufende, but the past participle is gedofen.
Note the conspicuous lack of the "-d" that marks "regular" PDE past tenses, which the weak dyfan has.
(Sorry about the unclear formatting here; I don't have access to tables. You can look them up on Wiktionary if you'd like them laid out more clearly.)
So to bring us back on topic, "sneak" derives from Old English snican, which was a type 1 strong verb. That means that the verb stem, snic, transforms as it's conjugated. In particular, the past singular, rather than being snicede ("sneaked"), as you would expect if it were a weak verb, becomes snac ("snuck").
So "snuck" is actually an older "correct" form than "sneaked", because I trust Clark Hall over Merriam-Webster for ancient usage. I suspect that what's happened is that prescriptive grammarians attempted to "regularize" it, but the older usage continues to leak through, whether because it's actually persisted continuously in certain populations, or because those "irregular" forms actually aren't arbitrary and people are instinctively attempting to follow inflection rules that they don't consciously understand. (Much like people will instinctively follow the Royal Order of Adjectives even though most of them have never even heard of it.)
In any case, "snuck" and "dove" are both perfectly legit, and have ancient pedigrees โ in the former case, actually older than "sneaked".
Tangentially, Old English had a vestigial passive mood rather than just using [to be] [participle] like Modern English does. However, when I say "vestigial", I mean, "there's like two examples in the entire surviving OE corpus". Like the dual number, it was almost gone even then.