The bane of Standard players everywhere.... or is it? If you're paying any attention to the discussion of the format, you'll know that this two-card combo has dominated all the talk. And if you aren't, let me explain - Saheeli's -2 copies Felidar Guardian, which blinks Saheeli on ETB, making it a fresh permanent that hasn't used a loyalty ability yet this turn, which lets it make another hasty Cat Beast, ad infinitum.
I, however, don't think this combo will be the format-destroying scourge most others seem to. Let's dig into why.
1. MathSo, the first thing people said was, "This is a turn 4 format now!". First of all, I suggest that good aggro decks can kill on turn 4 in most formats without interaction, but I digress. This combo will yes, sometimes be able to kill you on turn 4, but it won't be consistent at all.
The chance of naturally drawing both combo pieces by turn 4, given that you have at least 4 lands which make all the right mana you need them to, and assuming you're running 4 of each combo piece, is only 12.1% on the play, 16% on the draw. Throw in that you actually need Saheeli on turn 3, and we're down to about 11.1% and 14.7%, respectively. Not all that hot. I should, of course, point out that Saheeli's plus give a scry, which helps find the cat, but this adds only about 3.6% (a tiny bit extra on the draw compared to on the play). Still looking at a pretty low percentage in any case.
But, people have also noted the whole thing can be played on turn 6, which also reduces the opportunities for the opponent to interact. The problem here, of course, is that for that to work, you need to have 6 lands. And so even if we're assuming that you have enough lands every time, those lands take up slots in your hand. So the chances of having everything by turn 6 aren't much better - 12.9% on the play, 17% on the draw.
But this has all been assuming you just have the lands you need, which is by no means a given. Even still ignoring the color requirements, or that the last 1-2 need to ETB untapped (both of which are going to be very dependent on the precise build you use - mana for the deck can be pretty good, but getting the last land ETB untapped might be a bit tough), just having enough is a serious concern. If we look at a typical Standard land count of 25 lands, then having enough by turn 4 is only a 67.5% proposition on the play, 76.6% on the draw. When you multiply those by the existing chances we had above, and we're taking a few percentage points further than before even.
For the turn 6 scenario, it's even worse (as you might expect): only a 36.8% chance of having 6 lands on turn 6 on the play, and 47.5% on the draw. Of course, you're less likely to be color-screwed by the time you're at 6 lands, but still more unlikely to have that 6th land ETB untapped.
There are, of course, ways to make things more consistent. There are a number of cantrips in these colors, though at 1 mana you'd need a creature (though Insolent Neonate can do some amount of work for you), so effectively we have... Anticipate, Cathartic Reunion, Tormenting Voice, Nagging Thoughts, and a number of no-selection draw-1s for 2 (best of which for the deck is probably Prophetic Prism). Best case here is Anticipate, and it really does help a good amount - it takes a card slot, but lets you see 3 deeper. Of course, you need it by turn 2, and it gives you less chance to hit your taplands, so it's not without cost. But if you can get it off, it adds... several percentage points to where you'd otherwise be.
The other thing which presumably helps you, and I'm not taking account of here, is mulligans. Particularly with the scry, your ability to toss back hands which are missing too much is going to help you out. I will note that it's still harder, even with the scry, to get it all on 6 than it is on 7, and that those bad 7s do sometimes get there. So the improvements won't be huge, but they're real.
All told, these improvements may get your chance of combo-ing out up to 20% or so, and while I haven't actually simulated finding an optimal goldfish list or percentage, I find it hard to imagine you can get that to much over 25 or 30%, especially on the play - certainly well below 50%. Goldfishing turn 6 is much more plausible to be able to optimize for, but on the other hand, that's not so impressive - even midrange decks can routinely goldfish at that rate. Heck, limited decks can. If anyone has done this kind of optimization, I'd love to hear about it. I suspect the optimal goldfish-turn-4 list is somewhere around this:
That leaves the question open, though - how long does it take, typically, to get the combo together? By 'typical', I'm going to say the point at which you can expect to have a greater than 50% chance to have the combo assembled (and not the average turn it's assembled on,which is almost surely worse/later). In an unoptimized list (i.e. 4 of each combo piece, a pile of lands, maybe some cards that don't help assemble at all), we're looking at somewhere around turn 10-11 (depending on play vs draw, exact manabase and composition of irrelevant cards, etc.). In an optimized version... well again, I don't know what optimal would be for minimizing time to goldfish, but my guess is that it's probably going to be turn 6 (though turn 5 wouldn't surprise me - getting an extra turn to deploy cantrips helps a LOT).
(Pre-Post Edit: Okay, I missed Contingency Plan, but come on, let's be serious - not THAT much better than Anticipate, and hard to see it actually, you know, making the deck).
2. InteractionThere are lots of ways to stop the combo in Standard - any way of killing a 1/4 at instant speed (Grasp of Darkness, Murder, Unlicensed Disintegration, Stasis Snare, revolted Fatal Push, Warping Wail, Harnessed Lightning + an energy), any way of killing a planeswalker at sorcery speed (Ruinous Path, 4+ damage from combat and burn), 1 damage to a planeswalker at instant speed (Implement of Combustion, Shock, Fiery Temper), Counterspells (Void Shatter, Disallow, Metallic Rebuke, etc. etc.) Misc (Thalia, Authority of the Consuls, Dampening Pulse, win faster).
Most of these cards are fairly commonly played already. Moreover, virtually every deck in the format plays at least some of these already, even in the maindeck - and most have more in the sideboard. And while it's certainly possible to have plans to deal with most or all of these... well, you need to have plans to do that. Which take up slots. And time. And make your deck less of a consistent quick combo. That's not to necessarily say they're bad, but it does bring us to
3. But what about Splinter Twin?/So where do we stand?So the big comparison that gets made with the Copycat combo, of course, is Splinter Twin. Twin was so good, it even got banned in Modern. Everyone knows, of course, that this combo is worse, but Standard is also a weaker format than Modern, and is the combo really that much worse?
So, in terms of consistency at least, yes, it's quite a bit less consistent. This was especially the case when you could Preordain and Ponder - that many good cheap cantrips? You get quite a bit of consistency there. Even afterwards, though, Serum Visions is basically as good at digging as Anticipate, and it's a whole mana cheaper, which means you can use both turns 1 and 2. Furthermore, you got to play with more pieces than 4 of each - typically you played 6 Exarch/Pestermite, which is a 50% increase. That 50% doesn't translate to 'having it' 50% more of the time, of course, but it's not that far off. So this is really significant.
The bigger thing, though, is that Twin got to play a different kind of game. Turn 3 Exarch, tap down your land, untap kill you. It's only vulnerable to instant-speed interaction. It can kill out of nowhere. And realistically, represent doing other good things as well. Lots of flexibility. And lots of generally other good cards - the snapcaster/bolt/remand/cantrips Blue Moon kind of game. The deck was way more consistent, way more efficient, and played a really good game even when it wasn't comboing off.
At this point, I want to make a point about the other shell a lot of people have discussed for the combo, and that's using it as a finisher in a control deck. In some sense, I can see that - there's really not a reaosn it wouldn't work - but I am not terribly convinced by this, either, for one big reason: Torrential Gearhulk. Gearhulk already provides that deck with quite a quick clock to finish the game off, it takes fewer slots, it comes at instant speed, it helps support the control aspect of the deckmuch better. So it's a bit tough for me to think that such a deck is going to go for the combo over Gearhulk, or take up enough slots to go with both.
Having said allllllllll of that, I don't think the combo is just terrible. I could see it still being good. I wouldn't be entirely stunned if it was strong enough that a banning needs to happen (though I kind of doubt it). My main point is, you need a really good shell around it that plays to its strengths. You can't just throw it anywhere and have it be busted. It's not like pre-ban Eldrazi was in Modern where every flavor was effectively busted, and it was all about optimizing for the mirror. You need to build the right shell for it to be good, and my guess is that it will be good in that shell, but not busted.
What is that shell? I think a lot of people are reasonably close - you play Jeskai, you play a lot of good ETB creatures, with a mix of disruption and a bit of selection. Not too far from the Panharmonicon decks we started seeing last Standard.