I searched it first and it seems like it has been a while since this topic has come up here. I have played both cards a lot over the years, but have mostly stuck to Serum Visions over the other. One of the few times that I actually played Sleight of Hand in Modern in a deck other than Storm was a grixis control deck that abused flip-Jace, where they were extra Serum Visions copies.
Something that I learned from experience in formats other than Modern is that card-selection is highly relevant when in a deck with lots of shuffle effects.
In particular situations and decks, Serum Visions is undoubtedly a bit stronger, but it is worth considering that it is a turn before scrying pays off, which devalues it considerably if fetches start getting involved.
While it goes against habit, I'm going to be trying Sleight of Hand in Snapcaster decks more often in decks where I find myself wasting the scrying on Serum Visions...URW Spell Queller Control is a current example.
I still think you generally want SV over Sleight. Yes, there are times where Sleight is better. If you need an answer immediately and that answer is the second card from the top, Sleight beats out SV. But in many cases, SV's digging is more likely going to beat out Sleight. If you're looking for a specific card for a later turn, SV sees one more card which can make all the difference. SV also avoids awkward situations where you wanted both cards. This extra digging, and selectivity, is frequently more valuable with Snapcaster decks. For example, I love the T1 SV, T3 Snap/SV play in certain scenarios, especially against an aggressive deck where I'm digging for something like T4 Verdict. Sleight does not facilitate that well as I'm digging one less on T1 and then one less again on T3. Sleight also tends to be a little less valuable in the late game where I'm not looking for an immediate play so much as I am sculpting an endgame plan.
In general, I like to look at results to inform my decisions. There are very, very few decks I can think of that ran Sleights before the full SV playset. This could be netdeck groupthink, but I think it's more likely an indicator about how much more powerful the extra dig and selectivity is. Having run both cards in Ad Nauseam, I'm often happier with SV than Sleight. Having played SV a lot in UW Control, I'm much happier seeing SV than I would be seeing Sleight in that slot.
One thing that helped my SV utility was just fetching before I played SV. This is doable in decks like UW Control and UR Storm because it's hard to fetch wrong in a two-color deck, but obviously becomes harder with Jeskai, Esper, or Grixis. This really optimizes probability because you remove a land from a deck (unless you're digging for a land, in which case you fetch after) and then know your scrys stay on the bottom and don't get shuffled in. Again, I know it's not feasible for all decks that run SV, but it helped me improve my SV usage in at least UW Control.
I don't think I'm ballsy enough to swap out SVs for Sleights in a deck like what you listed, UW/x Spell Queller, but in playing Storm recently similar thoughts have crossed my mind. Being able to "pocket" one of two as opposed to one of one scratches the "immediate payoff" itch. But at the end of the day Sleight only ever lets you see 2 cards while SV lets you see 3. Or if you're drawing it out to the next draw step, Sleight lets you see 3 and SV 4. But in the scenario where you're extending it to the next draw step, SV is MUCH better due to the ability to keep one of the cards seen if it's what you need.
If you never get to that same top card in your next draw step because you're constantly shuffling, then the difference gets tighter. But how often does that actually happen? In a typical 8 fetch list it's not actually as often as you might think.
Something that I learned from experience in formats other than Modern is that card-selection is highly relevant when in a deck with lots of shuffle effects.
In particular situations and decks, Serum Visions is undoubtedly a bit stronger, but it is worth considering that it is a turn before scrying pays off, which devalues it considerably if fetches start getting involved.
While it goes against habit, I'm going to be trying Sleight of Hand in Snapcaster decks more often in decks where I find myself wasting the scrying on Serum Visions...URW Spell Queller Control is a current example.
In general, I like to look at results to inform my decisions. There are very, very few decks I can think of that ran Sleights before the full SV playset. This could be netdeck groupthink, but I think it's more likely an indicator about how much more powerful the extra dig and selectivity is. Having run both cards in Ad Nauseam, I'm often happier with SV than Sleight. Having played SV a lot in UW Control, I'm much happier seeing SV than I would be seeing Sleight in that slot.
One thing that helped my SV utility was just fetching before I played SV. This is doable in decks like UW Control and UR Storm because it's hard to fetch wrong in a two-color deck, but obviously becomes harder with Jeskai, Esper, or Grixis. This really optimizes probability because you remove a land from a deck (unless you're digging for a land, in which case you fetch after) and then know your scrys stay on the bottom and don't get shuffled in. Again, I know it's not feasible for all decks that run SV, but it helped me improve my SV usage in at least UW Control.
If you never get to that same top card in your next draw step because you're constantly shuffling, then the difference gets tighter. But how often does that actually happen? In a typical 8 fetch list it's not actually as often as you might think.
Standard: lol no
Modern: BG/x, UR/x, Burn, Merfolk, Zoo, Storm
Legacy: Shardless BUG, Delver (BUG, RUG, Grixis), Landstill, Depths Combo, Merfolk
Vintage: Dark Times, BUG Fish, Merfolk
EDH: Teysa, Orzhov Scion / Krenko, Mob Boss / Stonebrow, Krosan Hero