Brave Sir Robin (Bant Knight Retreat)

[anwei]

I just did a deck tech with Bant Knight+Retreat combo at SCG STL, and was in a couple feature matches, losing to Shardless BUG, beating reanimator, and losing in T4 to storm. (None of the matches are particularly interesting, and none were combos.)
I ended up 6-3 Day 1, 6-0 Day 2, and lost in Top 4 to Storm.

The List I ran
Lands (22)
3 Trop
2 Tundra
1 Savannah
1 Taiga
1 Forest
1 Dryad Arbor
4 Misty Rainforest
4 Wooded Foothills
1 Wolf Run
1 Sejiri Steppe
1 Karakas
2 Wasteland

Guys (14)
3 Hierarch
1 Birds of Paradise
1 Qasali Pridemage
4 KotR
2 Vendillion Clique
1 Sigarda
2 JTMS

Spells (24)
4 GSZ
4 StP
4 Brainstorm
1 Sylvan Library
4 Retreat to Coralhelm
4 Force of Will
3 Daze


Sideboard (15)
2 Envelop
1 Qasali Pridemage
1 Crop Rotation
1 Bojuka Bog
1 Tormod’s Crypt
1 Pithing Needle
1 Gaddock Teeg
1 Scavenging Ooze
1 Krossan Grip
1 Engineered Explosives
1 Submerge
1 Path to Exile
1 Izzet Staticaster
1 Council’s Judgment

Playing
Some of this is obvious, but a few tips:

• The basic combo uses Knight to get a land that untaps Knight, and you repeat until you have a big Knight, netting mana along the way (or tapping guys). End the cycle with Wolf Run, and trample in.
• More specifically: Hierarch+Knight+Retreat on Turn 3, you Knight a land for a fetch (untap Knight), fetch something up (untap Hierarch or tap their guy), then tap Hierarch and the land for mana. This nets 2 cards in the yard and 2 mana and resets. You do this like 5-6 times and you have a 12/12+ Knight and like 10 mana, and their team is tapped down. End by pulling out Kessig Wolf Run and giving Knight +8/0 and trample and swing lethal.
• When comboing, try to leave a fetch uncracked and a target and Sejiri Steppe untouched in your deck. You can respond to instant-speed removal while Knight is tapped by fetching with this land (untapping Knight), then activating Knight on the last to go get Steppe and counter the spell (even Abrupt Decay). It's best to play (or have) this land first (before you cast Retreat), but that's obviously not always possible. Use Knight activations to start with 2 fetchlands, then combo off of one and leave the other aside, if necessary. The deck is new enough that people hold off on removal sometimes, and waiting 1 Knight activation means it's already too late. Similarly, if they go to remove the enchantment, you can crack the fetch to start over, in response, and finish comboing before their removal resolves. If you think they have instant-speed removal, you don't have a fetch out, and you're not under pressure, it may be best to pass the turn, since any fetch beats their removal, and you'll probably find one before they find a second thing. Obviously, there are risks.
• Ask to shortcut when doing the activations (after getting your first fetch, as sneakily as possible), and order the lands in your hand so that you don't run out of forests/plains to fetch (especially with your leftover fetch). I stack these as: Fetch (backup for Steppe), Fetch (combo), Dual, Fetch, Dual, Fetch, Dual... Fetch, Taiga, Kessig Wolf Run, Dual, Steppe. The Dual/Steppe at the end of the pile sit in reserve, waiting until (if) they're needed. You should decide initially which lands you want to keep around, in case something goes wrong (i.e., don't get your combo ruined and be stuck with Dryad Arbor, Steppe, Wolf Run, and Forest in play). In some cases, you need to waste yourself to boost Knight.
• Gauge whether you should combo in the attack step or your main phase. Sometimes you'll only have the necessary mana in your main phase. Examples: I had to fetch out wastelands for Maze of Ith before combat once, and had to generate enough mana to GSZ for Qasali Pridemage on Ensnaring Bridge a couple times. Sometimes you need to do it during the attack step, e.g., when flash creatures or Aether Vial mean that potential blockers could show up, and you want to tap them down. Do this by main-phase Knight out 2 fetches, then attack, then crack a fetch to untap Knight and start again). This can also be helpful for situations where they might wait to try to disrupt you mid combo and leave you stranded for future turns (e.g., Stifle on the very last Retreat), so you're already attacking and can't end up tapped accidentally. It's also a good safety net so you don't screw up and miss a trigger.
• If you have to give Knight pro-blue to dodge blockers, either do it after attacking or make sure that Steppe is the last land you fetch out, and that you stack the Steppe trigger first, to let the untap ability resolve. If you screw up, she can't untap. I did this a few weeks ago. It was mortifying. I passed the turn with a 25/17 trample Knight, in turns, in game 3.
• When you have the combo, there are some obvious ways you lose (like double removal or Ensnaring Bridge or fog or whatever). Some ways you lose fall under the heading of "things that normally happen in games." It's best for us if our graveyard is untouched, we have a mana dork, they don't have tons of blockers and/or protection blockers (TNN), they haven't gained much life, we haven't lost too much life, and they don't have removal. We can usually absorb 1 of those with no problem, and 2-3 of those (depending) if we're willing to waste ourselves out of lands as part of the combo. When any are true, you may need to plan carefully. Pay attention to the life you're able to pay for fetches and the number of "dead end" lands in your deck, which will inhibit how big Knight can grow or determine sequencing.

My brief thoughts/changes
I play Bant a lot, and really like Aggro-Combo-Control shells like NO Bant. I'd say ~40% of the wins are through the combo; slamming a fast Jace, exalted Clique, or normal Knight is more common. Here's the biggest problem with the deck, and why I still don't know if it's any good after T4'ing: In NO Bant, Knight (or Clique/Jace) beats are plan A and NO'ing a dork into a hydra is plan B (or vice-versa). In this deck, Knight (or Clique/Jace) beats are plan A and Knight is plan B. And Clique and Jace are fewer and less supported. Whether the combo is good enough or not, too much hangs on Knight.

As asked below, Decay in particular can be a problem - my 3 losses in the swiss were all to Abrupt Decay decks (though I was like 3-3 against them total). Crop Rotation counters it (and active Knight protects herself), and I'd probably go up to 2. I don't actually think Decay is that much more problematic than Swords, and Knight is pretty much already a must-kill card. With Zenith, you get a lot of them - Zenith recycling means, midway through a grindy game, you're drawing like you pack 10ish Knights in your deck. That's pretty good, and they show up a lot.

The options to address this are to protect Knight better or to have more threats, and I think the latter is better. I tried Sylvan Safekeeper a lot, and it was often disappointing. When I had the land to spare and hit GSZ, I often just wanted to grab another Knight. I think, if I play Legacy again anytime soon, I'd seriously consider moving FoW to the sideboard for a maindeck Crop Rotation (2nd in side), a 2nd Pridemage, a Scavenging Ooze or 3rd Clique, and Maze of Ith. Maybe I'd try Safekeeper again. A pair of Mother of Runes might be great, though lacking oodles of creatures is tough. For this version of the deck, I'd plan on swapping Cliques for True-Names, as they provide a very resilient Plan B and can be quite fast with exalted.

One of the more difficult decisions is whether to keep Kessig Wolf Run. Cutting it for a Wasteland (and Taiga for Tundra or a 9th Fetch) makes the mana much better, lets us go to 4th Hierarch, and the 3rd Waste is probably quite good, especially if cutting a Retreat. The problems are ensuring we can hit lethal and beating blockers. The blocker problem is mostly solved by Retreat, particularly if we don't need to make mana off fetching lands. The biggest widely-played issue is True-Name, and Steppe can allow one attack through them (which does cost our ability to avoid removal on the combo turn). Double Mother of Runes, to protect a blocker against Knight and Retreat, or Mirran Crusader with Sword of Fire/Ice, are much less likely.
Hitting lethal is a real issue. As above, comboing without KWR is best when our graveyard is untouched, they don't have tons of blockers and/or protected blockers (TNN), they haven't gained much life, we haven't lost too much life, and they don't have removal. If our graveyard hasn't been touched, a non-KWR deck can get Knight up to a 24/24 (non-exalted) by starting with 5 lands in play and putting everything in the yard, including wastelanding Karakas and Sejiri Steppe. If we want to keep the Steppe plan up (a good idea) and combo smartly from 5 lands, we can get Knight to a 22/22. StP on Goyf or a couple DRS activations (lands or life gain) are enough to stop this, and that's not great.

So, the "corner" cases where KWR is necessary - 4+ DRS land exiles, StP'ing a guy or getting hit with batterskull and not doing other damage that game, TNN+removal, a single GY hit like Relic or Crypt, etc. - come up often enough. It also doesn't put you all-in if you want to win and not just get to 12/12, swing, and pass the turn: if they have double removal or whatever, you didn't just waste yourself out of the game permanently. I think in my win-and-in at the Open, my Painter opponent slammed T2 RiP and I untapped and won, ramping into GSZ for Pridemage. By the time I cast it, I had 5+ lands in exile and only 3 in play: Knight would have maxed out at 17/17 without it. Obviously, we're a big favorite to win that, but painter+blast leaves us with 0 lands in play or the deck, which means actually stopping earlier and not going all in.

That all said, the cost of the red splash is real, you can still win most combo games without it, and most actual games are not fast combo games, but they run long enough that you get damage in along the way and both ways of building leave them with similar outs.

The Name
Several names have been floated around for the (Modern) deck, and I like "Brave Sir Robin" (bc it's a retreating knight, get it?) with honorable mention to "Helm's Deep." I've read that an SCG commentator finds this suggestion (and the conventions of nomenclature in Legacy) inhospitable to new players.

This gets discussed a lot. Obviously, as a side issue, SCG are singlehandedly responsible for the largest mass of Legacy name discrepancies in the past couple years. For good or bad, and eventual clarity or not, using Khans names means renaming 50%+ of the meta.
The one thing I would add is, deck names highlight internal unity and thematic coherence. "Bant" or even "Bant Tempo" lists colors and maybe a strategy. "New Horizons" references a very specific over-the-top tempo deck in a specific meta that knew what it wanted to be and why. "Esper Stoneblade" and "Jeskai Stoneblade" sound like they both run random piles of cards matching an arbitrary color identity, rather than being two interrelated but distinct strategies. Those are broad enough that each list probably shouldn't have it's own name.
This is a brew. "Reliquary Retreat" may become the publishing name of any any number of Modern and/or Legacy UGWrb/GWBu/GWBur decks that play Brainstorm/Force/GSZ or Collected Company/Geist/Chord or whatever else, and I don't think that's great. If there's anything Creative about building a deck, then dignifying the focus of that process with a name seems like a good thing. I doubt whether John Steinbeck could have published "Indebted Family Moves to California During the Depression and Struggles to Survive" with the same punch as "Grapes of Wrath." Building a deck is orders of magnitude less creative than writing a novel, but I hope deck names share more with novel names than with the labels on sorted recycling containers, which is what "Reliquary Retreat" feels like.