Swindler is a card which I've long rated very highly. It is cheap, and in the best case, incredibly devastating. However, recently it has lost a lot of its luster in my eyes. Let's break it down.
Swindler is terminal. This should give you some amount of hesitancy in opening Swindler/Swindler. On the other hand, the attack can be very powerful, so it's not crazy to think that it might be worth risking the collision. There are 2 main general factors here: how much does the collision hurt you, and how good is the attack?
Checking the opening percentages, we can see that compared to Swindler/Silver, by opening double Swindler you're losing 19% chance to hit 5 at least once, and a couple percent to go 5/5. You're also nearly 25% less likely to hit $6 (all of these figures are for turns 3/4 and assume no funny business from opponent giving you cards or stealing them from your deck, or spy attacking you, etc. - though it's important to note that without some kind of spy effect, their Swindler attack doesn't change these numbers for turns 3 and 4). You have a 37.9% chance to attack them twice, and a 98.5% chance to attack them at least once, up from 0% and 83.3% respectively. This means that by opening double Swindler, you get slightly more than .2 fewer 5-costs and slightly less than .25 fewer 6-costs (if you want those), and in return, you get .53 more Swindler plays.
To figure out whether or not that's worth it, you need to gauge how important hitting 5 (or 6) is, and how much the attack is worth. Let's go after the attack first: 7/12 (58%) of the time, you're hitting a Copper, which you will always be turning into curse. Copper into Curse is very good in the early game, because coppers are a huge part of the economy. 1/4 of the time, you're hitting Estate. This is really bad - worse than silver since it skips their estate for them. Indeed, you're always cycling early game, which is not good for you, though the appeal of the good cases is definitely enough to overcome that. Still, a quarter of the time it's actively bad. The last 1/6 of games, we are hitting their buys. How good this is varies a LOT from board to board, as when there's something really bad (used to be true more than it is now, but note that it's effectively always true for 5-costs by going into Duchy), this is very good. But sometimes you're e.g. turning Swindler into Silver, which is a downgrade, but a minor one.
So though I've gone for double Swindler in a huge percentage of games I've played where that was an option, I now think that this is usually incorrect, that Swindler/Silver is often better. The cases for this are when there's a key 5 or 6 you want to get early. The cases where doubling down on terminals is better include those times where there's something bad at every price point, and of course when the 4s are not much worse than the 5s. I also want to note here that Swindler/Swindler is a decent amount better (in a relative sense) 'on the draw' (i.e. as 2nd player), since you are more likely to hit a key card (on turn 4, your opponent has more total bought cards in their deck). Since hitting these cards is the best thing you can usually do, that is a noteworthy if not huge boost.
Swindler is a fine card later in the game, but it tends to not be special. Hitting estates is still bad. Hitting Copper turns from being very good (denying a good fraction of economy) to being very mediocre, as one copper into a curse doesn't matter so much at the mid-stage, and the fact that you skip the copper for them starts being a non-negligible fraction of that penalty. The simple way to look at it is, there are just fewer shuffles left for the junk to be junk, so one shuffle of skipping the junk is a bigger fraction of the total times that your opponent has to draw it for the remainder of the game.
Hitting important cards can still be quite good, but very often the opponent's deck has a lot of redundancy, so that messing one thing up is not too terribly difficult to correct with future buys. This is particularly true of a well-established engine, where they can just re-buy whatever component you're denying, and given that they're drawing lots of cards almost every turn, they get to do so quite quickly. If you play many Swindlers in a turn, it can still be very good, since it's much harder to deal with something like 4 extra terminals all-of-a-sudden, but if you are really having the chance to play 4 of a terminal payload card in a turn, there's often going to be an alternative that's better.
Swindler is a card which really wants you to be keeping track of knowing your opponent's deck. Knowing what you're going to hit can be very nice at times (though unless you have a Spy effect, you usually only have some probability cloud of a guess). More importantly, though,you want to know what card is going to be worst for their deck. Do they have too many terminals? Not enough payload? Are they lacking +buy? Something I don't see often enough from other players, but which I feel like I do often, is giving my opponent a card and then buying the exact same card. Sure, sometimes there's a Thief on the board, and you just give them that every time because it's terrible, but especially as with more expansions we've had fewer 'bad' cards, it's more often the case that you want to give them a marginal card, and you want to make sure it's the most marginal one for their specific deck.
One place where a Swindler-heavy strategy can really work out is in an Aggro Pile deck. It's a disruptive card that slows the opponent down, and it can also take chunks out of piles. It's usually pretty hard to make this your game plan from the start, because you have little control over what you hit, but it's something to be aware of as the course of the game progresses. If piles get low enough, especially depending on your position in terms of luck and deck quality has unfolded, you might want to start angling to get that pile ending, applying pressure to either seal up a short-term advantage (most likely in points) or hoping to get lucky because you're in a very bad spot. If you get in this situation, you need to be very careful about how much you lower piles and what you do in terms of points. If you start spinning the points angle too early, then your opponent's better deck will have time to crush you. But if you continue to slam your opponent with Duchies as you are also depleting piles, this may well turn out poorly for you as well.....
The most obvious defense you can take is to buy cards only on values where there's nothing really bad to Swindle into. You usually can't/shouldn't really make a go of this, as 5 always has Duchy, but 5-costs also tend to be really good and important to get. You should not take a big hit in card quality just to try to avoid something getting swindled, especially since it's far more likely for you to draw and use the card than for the card to get swindled. What you should do, though, is be willing to take some very marginal downgrades. The obvious kinds of examples here are something like Village over Walled Village when Thief is on the board and you aren't relying on the Walled bonus (which you almost never are). In general, 3s tend to be safer, as there aren't tons of very bad 3s, though 4s, 7s, and potion-costs are potentially actually the safest, in those cases where they're the only card at that price point available (though you do have to worry about running the pile out). Super marginal cantrips are things you can think about avoiding altogether. Pearl Diver is really the poster boy here - the chance to get turned into Estate is sometimes quite bad. On the other hand, having these means other things are less likely to get hit, so you need to figure out whether you're more in need of protecting your few key cards ("Good Stuff" decks often fall here) or staying clean and not getting yourself junked up (engines hate having an extra estate).
More importantly as defenses come other cursing attacks. It's certainly possible to get both these AND Swindler, but when the curses run out, you risk turning them back into coppers, which is very bad - at this point in the game you usually would much prefer Silver to Swindler, and because this point will come, you need to seriously consider whether the early spurt of attacking effectiveness from a 0th shuffle Swindler is worth it.
There are a number of other minor defenses. Moat and Lighthouse work about as well as they ever do (though they can get turned into Estates I suppose - this usually isn't a deciding factor though). Silver flood decks shrug Swindler off fairly well - often they give you a silver back, and worst case you're getting like a Chancellor, which really isn't that bad when you're silver-flooding. Jack of All Trades and Hermit are particularly good, since getting Coppers turned into Curses actually makes them trashable, when otherwise they wouldn't be. Against those cards, you usually skip Swindler, though I will note that if you do Swindle a copper in those cases, you should almost certainly go ahead and give the Curse. Cards with on-trash benefits help defend a fair amount (usually), but it's quite a situational thing. Be on the lookout for them, though.
Swindler is a strong card. When cheap options are good compared to expensive ones, and especially when there are bad things to stick people with, going for multiples can be good. This can lead to pile outs, but you need to have lots of game awareness.
More often, you will get one as an annoying attack that helps you ramp up early on. There's quite a bit of variance in the card, but that doesn't mean it's bad.