My feelings to start:
The formula for an effective hate card is complicated. The best hate cards are typically broad enough that they can be used effectively against fair percentage of the metagame, thereby justifying their SB slot
I want to call out, I think anything that could be called a hate card is really the wrong approach entirely and rarely works. (Grafdigger's Cage and Containment Priest probably worked, credit where credit is due).
When I think of good printings that solved real vintage problems without restrictions, I think of the aforementioned Dack Fayden "solving" Tinker, Delver of Secrets dethroning the glut of Jace, the Mind Sculptor + Dark Confidant decks, and the evidence today that the best single card against Workshop decks isn't Energy Flux, it's Oath of Druids. Gifts Ungiven got people to stop arguing for the restriction of Goblin Welder (believe it or not a LOT of people argued for that VERY loudly), but there are people who probably preferred playing against Slaver. None of these cards single handedly made a bad metagame a great one, but I think they all did more good than harm.
This hasn't really happened too many times in vintage, but when it has, it happened without trying, and I think it could be engineered. What all of these cards have in common is that they're not hate cards at all, they're generally useful cards that contribute to a deck's "Plan A", and just happen to have some form of strategic superiority over another deck.
Hate cards basically all have the same set of problems. I'm dipping into deckbuilding strategy here, and this might seem obvious, but bear with me.
In order for a card to help you win a game of magic, it has to generate an advantage.
In order for a card to generate an advantage, it has to cost you less than it costs your opponent.
I don't mean this in the strictly literal mana sense (though that is important), but rather your spend across all of your resources, both in the game and in the meta-game of deckbuilding. If you Lightning Bolt a Young Pyromancer after they've made 2 tokens, you're behind on that trade. That trade can still be correct, but only if you have a deck that can so sufficiently handle 2 Elementals on the board that it's worth losing a card over -
that advantage doesn't come from nowhere, it has to be somewhere else in the deck, you have to have already spent the cost in card-slots to run it.
The typical hate card is a permanent that your opponent cannot win through, using his Plan A. The default game plan of a deck facing down hate is to run removal. This is not universally the case, but much of the time, hate cards are specialized where the removal is more general, which means right off the bat, the cost for you to run 4 hate cards is more than the cost for your opponent to run 4 removal spells. Within the Oath matchup, an Abrupt Decay for an Aegis of Honor is an even trade, neither player is ahead - but your deck ends up worse against the field, and in aggregate, Oath wins in the metagame on that trade overall, even if it doesn't end up ahead against your deck. This can be mitigated with a hate card that does better than trade with removal (like Rest in Peace) or a hate plan involving a mix of cards that require different types of answers from your opponent.
This is exacerbated by decks that have non-removal answers to hate, immortalized in the magic phrase "There are no wrong threats, only wrong answers". Two current examples are Hollow One out of Dredge and the Inferno Oath deck which is designed to pretty consistently hard-cast its Oathable creatures. If we assume that Leyline of the Void perfectly counters Dread Return and Ingot Chewer perfectly counters Hollow One, you're already on the back foot. If you draw a mismatched Leyline of the Void when they draw Hollow One, you're behind ... but if they draw a mismatched Nature's Claim when you draw your Ingot Chewer ... you're still behind.
What all this comes down to is that in the metagame at large, a hate card doesn't incentivize an opponent not to play a deck. It costs you more to run the card than it costs them to run an answer, so iterated over a tournament in any mixed metagame, the hateful decks get paired agains the hated less and less.
I guess this has gotten pretty far afield from card design ... but probably the most common thing I see when someone suggests a card design for vintage is a specialized hate or answer card. There are already a ton of hate cards that don't see play in vintage, and even the best hate cards don't seem to be metagame shifting in any noticeable way.
I think there's a lot more room for impact in vintage card design if we consider two things:
Look for designs which cost less to play than they cost to answer. Low mana isn't enough. Cards that are flexible in other matchups, cards that replace themselves or leave something behind when they're removed - that's more important than the disruptive effect being more powerful.
Consider the inherent drawback of specialized cards as a tool to weaken decks. I've long believed that strong hyper-specialized mirror cards are a good tool to keep overpowered decks in check. Picture an artifact that was very powerful in a Workshop mirror, and weak elsewhere (I imagine something along the lines of Kill Switch but I suppose it would have to be better). If you can incentivize some Workshop pilots to run that card over something more general purpose, their position relative to the field weakens if they want to maintain an even matchup against other Workshop decks. Consider how important Steel Overseer is to Workshop mirrors, and how important Mental Misstep is to Mentor mirrors, and then consider how good Overseer is against Mentor compared to Misstep against Workshops