Wednesday, December 21, 2016

Live Booster Draft

Love drafting? Love playing Magic? Ever wish you could do both at the same time, rather than one after the other? I have. For ten years. I got a game published trying to make it happen. (Here's a diary of how that led somewhere else fruitful.) Last month, the idea struck me again, this time actually for Magic:
I don't hype my own stuff much. Today is different. This is the good stuff. You want to try Live Draft. Really.

A critical detail that makes the whole thing click: You also have to draft your lands.

I've done this a bunch of times now, and it's fantastic. You cannot get a better game out of 2 booster packs, nor as much play out of 6. I might well be the top expert on two-player draft formats, and this is The Best Two-Player Draft Format.

The highlights:

• Every game is a real game. Mana screw and land flood can only happen if your opponent puts serious effort into it and you allow it. Not only that, you're crafting your starting hand and every draw thereafter. Every game is real… good.

• Needing to draft land affects your card valuations considerably. Basic land goes from being a guaranteed and limitless resource to both vital and contested. Since your starting hand is whatever seven cards you pick first, you have to draft land early just to make your first land drop. If you're on the draw, you can spend your first several turns taking the land you need to cast your 7 spells. but you're risking that your opponent doesn't scoop you.

• We only add one of each basic land to each booster; so if you want to cast any black spells, you better get a swamp before your opponent takes them all; and if you want to cast a double-black spell, you've got to get two swamps. That means the card that you've been tossing out of booster packs since you started drafting—that annoying basic land—is kind of a big deal; having three of a basic land in the draft makes a huge difference for casting spells of that color.

• In a two-player Live Booster Draft, you see every single card in the draft except your opponent's first pick. That's a lot of information with which to make your plays and your picks (more than I can use). It also gives each player one potentially game-shaking surprise, which is nice drama for an otherwise unusually strategic experience, plotting out how you'll answer the threats you know your opponent has taken, and how you plan to win.

• Nearly every game has played within a turn or so of exhausting both booster packs, or a bit less. I had originally guessed we'd have to open new boosters as games went on, but you only need one pack per player.

• Live Draft begets an amazing league format that I'll detail further below. It's also not confined to traditional booster draft (though that's the only mode I've tested so far); it'll work with any draft where players take a single card at a time.

The rules for Live Booster Draft:

0. This rule is new and optional, but recommended: Shuffle up an equal number of each basic land and give one to each player randomly and privately. This is your first pick. Read in the comments below why you'd want to do this.

1. Each player opens a booster pack and adds one of each basic land to it.

2. Players draft their first seven cards and then the game begins. Roll for the choice to play or draw first as always (or use the last loss to decide).

3. Whenever you would draw a card, instead draft a card from the next pack, the one to your right. If you would draw multiple cards, draft one card from that many packs. Every time you draft, pass that pack left as normal.
When the game begins between Ajit and Betty, pack A is on Ajit's left side (because that's the pack he took his seventh pick from) and Betty's right; pack B is between Betty's left side and Ajit's right. After Betty draws from pack A on her turn, she passes it to her left, queued up behind pack B. After Ajit draws from pack B on his second turn of the game, it goes on his left (while Pack A remains on his right).
Ajit casts Tezzeret's Ambition. He draws card 1 by drafting from Pack A and passes the pack left. He draws card 2 from Pack B and passes the pack. Finally, he draws card 3 from Pack A (again) and pass the pack (again). In this way, the order of packs can be altered; Betty would have drawn next from Pack B if Ajit hasn't cast that spell, but instead she'll be drawing from Pack A.
4. Whenever you would interact with a library other than drawing a card—like searching it—treat the closest booster pack to your right (the one you would draft from if you drew a card right now) as your library. Shuffle it first if the order matters. Don't pass the pack afterward (unless you also draw from it). Effects that solely re-order the library are irrelevant; you can skip shuffling and scry effects—unless they immediately use the library after, like Llanowar Empath.

5. Players never lose the game for being unable to draw a card. If one pack is exhausted, draw from the next. When all packs are exhausted, play continues without new cards. If all players agree—because the game is stalemated or for any other reason—they may replace a pack when it is exhausted with a new booster pack. (You might only agree to replacing the first pack emptied on the basis that the second pack is also replaced when emptied. Such agreements are binding unless all players agree to cancel them.)

'Milling' effects like Tome Scour put cards from the pack you're currently treating as your library into your graveyard or exile, adding them to your card pool. 'Shuffling' effects like Gaea's Blessing put cards from your card pool into the pack you're treating as your library, removing them from your card pool.
We were quite amused to discover that Perpetual Timepiece is a very strong card in this format, because it adds cards to your pool (removing them from your opponent's possible picks) and then when you're ready, you can crack it during your upkeep to put the best card in your graveyard (whether it's one you drafted or milled there) back into the pack that's currently your library, then drafting it for your draw step.
I recently played LBD with Eternal Masters and opened Diminishing Returns. We elected to remove it from the pack because it would require us to open two more boosters and miss out on 10-24 cards, which seemed wasteful for a premium pack like EM. In a cheaper format, sure; It'd be chaotic and strong, but not broken. Technically cards you exile from 'your' library would become part of your pool, but I'd skip that if we did play Returns because that's supposed to be a drawback, not a huge advantage toward your league deck. Speaking of which…
Live Booster Draft League

Live Draft is the best way to get play value out of opening one booster per player. It also lends itself to the best way to get play value out of opening three (at least for groups too small for a regular booster draft, 5p-). First, you Live Draft; then Live Draft again; then build 30-card decks using your combined pool from those two drafts and play a best-of-three match; then Live Draft a third time; and finally, build 40-card decks using all the cards you drafted and play a best-of-three match.

There is a hidden greatness to drafting during your second and third Live Draft of a league: The game you're playing is entirely its own and will be determined by your picks and plays, but the cards you're taking are being adding to your previous picks for the 30/40-card decks you're going to build.
If you went {G}{W}{B} the first time, but the new packs have a lot of good {R} and {U} cards, do you switch completely to ensure victory in this game at the expense of building a consistent deck afterward? Do you ignore them and solidify your Abzan affinity, improving your future chances but costing you this game? Or do you walk the line, taking {R}{W}{B}, perhaps, knowing some of your cards will make the cut next round, and some won't?
Sometimes you see cards you want for your pool but know you can't play this game. How highly do you pick them? They're dead cards in your hand (unless you've got looting effects) and won't affect the game, so taking them highly might cost you a game-changing pick, but if you wait too long your opponent might just hate draft them. Speaking of which, when do you hate draft? It's definitely correct sometimes, but it's expensive when it costs you a pick and a draw.
We opened a Demon of Dark Schemes in a second Kaladesh live draft, and even though there was a swamp in one of the boosters, giving us three all told, there were also two good {B}{B} removal spells, so I took one swamp and my opponent spent two consecutive early picks taking the others (which was costly but brilliant, because it ensured him those two removal cards as long as I had relevant picks). That made the Demon literally unplayable for either of us. But my first draft had been {G}{B}, so I wanted it for my card pool and ended up taking it 13th maybe (out of 20, remember). 
In a single Live Draft, it doesn't matter if there are cards left undrafted in the booster packs when the game ends. In LBD League it does: When a live draft ends, continue drafting cards from the booster packs in the order they come to you until all cards are drafted. These are part of your pool for deck construction.

The deck construction rules are simply that you must play a minimum of 30- or 40-cards, depending on the round, and that you may (in stark contrast to the live drafts) add any basic lands you choose and exclude any cards from your pool that you like.
I've done this league numerous times now, and it's super fun. It also chunks well across short play sessions. I play 2p over lunch with my co-worker, and we'll often play our first two LBDs day one, build and play our 30-card decks day two, Live Draft the third time and build our 40-card decks day three, and then play our final decks day four.
It's dangerously easy to go five-colors in a single Live Booster Draft (because you can just take that fifth basic land when you need it) but that hurts your deck-building later on; for league play, try to focus on two or three colors even if you splash more.
For a tournament, I recommend giving every round equal weight: Each live draft is worth 3, 0, or 1 points (win, lose, or draw), and each match is too.
With four players, round 1 (live draft): random pairings. Round 2 (live draft): pair winners, pair losers. Round 3 (30-card decks): Yet unmatched pairs. Round 4 (live draft): winners, losers. Round 5 (40-card decks): winners, losers.
For a longer 4p tournament: Play each live draft against a different opponent. Play each deck you build against every opponent.
With six players, round robin.
Thanks for reading. I hope you get a chance to try this out. If so, let me know how you liked it in the comments below.


  1. This sounds awesome. I'm really looking forward to trying it out!

    I have one suggestion for the league play: instead of 30 and 40 card decks at the two stages, I suggest trying 20 and 30, particularly if you are only going to play single matches/games.

    This is an armchair recommendation in this context so far, so take with a pinch of salt, *but* it is based on both theory and experience in finding good deck size requirements for different formats.

    There are two main pressures on deck size for a format: producing enough variation between games, and providing players with an appropriate amount of selection power over cards (in some formats it's easy to tweak this directly without changing deck sizes, but that's not easy here).

    On variation: 20-card decks happily provide enough variation for maybe three games, and start to feel kind of repetitive after. The 40-card decks which are standard in limited have much more longevity -- which makes sense, as they're often used for in the vicinity of five rounds of matches.

    The other factor is how much selection power people get. This depends on how big the total card pool is, what proportion each player ends up with, and potentially in some fiddly way on the actual selection mechanism. As a motivation for thinking about this, note that sealed generally produces less powerful decks than booster draft, despite having twice as many cards-per-player. When splitting a card pool between two people, I think the number of cards-per-eventual-deck-slot you want is something like 30% higher than in 8-player booster draft (that's an empirically judged number; I'd believe anything between 20% and 50%). That fits pretty well with 20-card decks after two rounds of live draft. If you're feeling conservative you could try 25 and 35 instead.

    1. I hope you try the 20- and 30-card decks, and let me know how that goes. I've been playing best-of-three matches with my decks and 30- and 40- have felt right so far (but I've only done those three times each, so I can't guarantee that's optimal).

      It's a good point that decks will be better the more players are drafting / packs being opened. I've only done this two-player so far and I can imagine an 8p experience being significantly different. (I had been thinking a group of 4, 6, or 8 would do simultaneous unlinked two-player LBDs, but you certainly could keep the packs flowing around the whole table, with just a bit of awkwardness. Worth trying.)

  2. One thing that I am worried about is that having 0 unknown info makes this a bit prone to analysis paralysis. One option would be to only draft out of the top five(?) cards of a pack, starting with the basic lands. That way you're always seeing at least one new card on your turn and you feel less compelled to have perfect knowledge of you're opponent's hand. It also makes shuffling the pack more exciting, because it can totally change the options. I look forward to trying it out!

    1. There's one piece of unknown information per opponent—the first o cards they draft. But yes, having almost all of the information makes the game much chessier than usual. Originally, my buddy and I were using the public-knowledge-stays-face-up courtesy but we stopped and it was just more fun to forget every last pick they'd made.

      I actually tried an access draft (not live) where players have very little information about the content of packs and gain more over time. It was not good. But please do try your idea and without, and let me know how it goes.

  3. We tried this out last night. As things worked out, we played a 3-player game, live-drafting 3 packs of Kaladesh.
    We wondered what to do with landsearch spells (in this case Attune with Aether). We decided to play it that you just get a basic land out of the box, not that you search a 10-card booster for whatever lands might be left in there, but I'm not sure that was the right choice. Between one player having two Attune with Aether and someone else having a Prophetic Prism, colour wasn't much of an issue.

    I first-picked Panharmonicon, and was able to draft combos with it such as Hightide Hermit plus Architect of the Untamed :) The format lends itself to drafting combos rather nicely: you can preemptively pick up the more-useful-on-their-own pieces, and then draft the final pieces the moment that you want them.
    I still died first, due to tapping slightly too many of my big guys attacking, and not playing around the RW player having Hijack for my other big blocker (and not leaving mana up for my Dramatic Reversal).

    The game was very interesting and cool, and we're quite looking forward to playing the second live draft in a week or two (hopefully with Aether Revolt prize packs from the prerelease) and then playing the first prebuilt-deck league game. I'm more inclined to tell people to build a 20- or 25-card deck from 2 packs' worth than a 30-card deck.

    1. Awesome!

      Attune with Aether searches the booster pack that you would draft from next (rule 4).

      If the three-player-ness of it was alright, cool. If that felt off, you could try Three-Headed In-Fighting (but I have no idea how that does with Live Draft).

      A 30-card deck is usually 17 spells and 13 lands; it shouldn't be hard to draft 17 spells from two booster packs. I suggest you aim for 30-card decks and if everyone agrees while building that's too much, then try smaller ones.

  4. Yesterday was the first time I opened an additional booster pack. We were drafting RtR and my opponent went turn 1 Urban Burgeoning (I know, right?), turn 2 Chronic Flooding, allowing him to put 6 random cards from our packs into his graveyard (and thus card pool) each round. Fortunately, I got a turn three Keening Apparition to stop him, but he'd already nabbed 9 cards.

    I had Inspiration, Runewing, and Righteous Authority, so I really wanted to keep drawing cards when the packs ran out. He agreed, and I opened the pack to find Vraska the Unseen. And then I started drawing even more cards from it. I won that game, but…

    It felt a bit off that I got first crack at our collective third booster pack. The rules above give players the tools to ensure that either everyone gets a new first pick/pack or no one does, and that's good enough for now, but I'll be keeping an eye out to see if we shouldn't codify it so that's the only way new boosters can be introduced.

    Let me know if you open a single extra booster and see a problem with that.

  5. Great idea, Jay. I hope it catches on.

  6. A tweak:
    Shuffle up five (more) basic lands and give two to each player (hidden) to start their hands.

    The existing land allotment is very tight, and we've found that there's room to loosen it up a bit. This also makes cutting a color of land more of a gamble, which allows players to focus on drafting spells a little more.

    Try out his tweak, and let me know how you like it.

    1. We've moved to just one basic land each. Two drastically reduced the value of land, but starting with one random basic keeps land-drafting relevant (but still not quite as cutthroat as without).

  7. I found this format on Reddit a couple of months ago and it's a blast. When I tried to look up this page again, though, it took nearly an hour of Googling! "Live draft" usually brings up Pro Tour drafts and NFL events.

    I'd like to suggest you rechritsen this format to a more Google-able name, so that when it takes the Magic world by storm and has a huge following, folks like me can have a very specific search term when we try to find footage on YouTube or Twitch.

    How about Stream Drafting? (No, definitely not that.)

    1. Thanks, Chris. I'm glad to see people are hearing about the format.

      As of today, I'm finding that "live booster draft" puts this page on the top in Google and in Bing; quite possibly because of all the redditors searching for it.

      That said, I'm open to suggestions for a better name.

    2. Well, there's Merchant Draft, after the game you mention at the top :)

      Other suggestions from Reddit:
      "Concurrent Booster Draft"
      "Rotating Booster Hand"
      "Sushi Go - MTG Style"
      "2 Player Draft 'n' Play"

  8. I just tried this with Duel Masters / Kaijudo.
    The first 9-card pack fed our shields (which we randomized the order of) and then then first 4-cards of our hand. The next 9-card finished our 5-cards hands so we could start.
    We ended up opening one more booster for each of our drafts because we ran out of cards before the game was fully decided.
    It was fine. Very different from Magic LBD, since half the cards you take end up upside down in your mana zone, and you have to wait for your opponent to attack to get access to your first picks.
    The first combined pool we just played all the cards we drafted. The second we had enough to pare down to 40 card decks.
    Ultimately, the games with shuffled decks were better. Because threats and answers are very binary in this game, the chessier version was too chessy. Just being able to double-block, much less cast auras or instants gives Magic the depth to thrive in LBD, where DM sputters out.
    Ah well.

    1. While I haven't tried the official method of drafting DM, this experience leaves me skeptical. Colors and archetypes matter much less and you almost always take the rare. Too many automatic or inconsequential decisions.

  9. Props for coming up with this really cool format! Name suggestion: Hybrid Draft Format.

    A question i have is would this work with my cube list? There's a lot of dual lands in there, which are prob less common with boosters.

    1. You can totally Live Cube Draft.
      Dual lands will be of slightly less value given the basics we add in, but still relevant.

  10. First things off, I really, really like the idea of this format. I think it's fun and filled with lots of interaction. I'm only bothered by one small detail, and that's that scrying has no effect during a draft match.

    Scrying is, by all accounts, an important mechanic, and for many key reasons. By existing, it passively minimizes the effects of "bad" variance while increasing the effect of player skill, since a player is making decisions about what to keep on top and on bottom- sometimes this is as trivial as getting rid of two lands in the lategame, while sometimes you're stuck thinking about which of two similarly powerful cards you'd rather have at the moment (or how good they are versus the value of a new unknown draw from your remaining library). It occupies a valuable design niche that isn't as strong as card draw but is strong enough to be useful; this, in turn, allows it to be used by colors that aren't usually given raw card draw, such as green and red. In Limited, it's a fine way of giving some power to cards that would be too broken with draw but unexciting without it, which gives skilled game designers options and tools to work with. Lastly, it has deckbuilding synergies with effects that care about the top of your library- these effects are more commonly seen in powerful eternal formats, but we had one in a Limited set recently with Dryad Greenseeker in Core 19, and that card was widely evaluated as being powerful and exciting for draft.

    Now that we've established that scry is important and popular in the current era of design philosophy (after all, it was made evergreen after Origins), we see the problem: Live booster draft straight up cuts-out a key mechanic from many Limited environments. This argument seems a bit silly at first because of course live booster draft is a distortion of normal Magic: for example, heavy color requirements are much worse than they would normally be, and card draw is much stronger than it normally is. But the key difference here is that these changes are still a reflection of what Limited magic fundamentally is. In normal magic, these downsides are downsides and these upsides are upsides. There's a difference between having flying be unimportant in a draft set versus not having flying at all in a draft set. Does having no flying creatures make that format more fun to play, or has something that we love and know about Magic been completely removed? I'd argue the latter.

    Now you see where I'm going with this: Scry should have some sort of replacement effect. It's just too ubiquitous and useful as "that effect that gives soft card advantage" to be straight-up removed from live draft, since that means a not-insignificant number of cards from modern-design sets will be unnaturally weaker than usual. I don't know what this effect should ideally be, but I know that I wish it existed. To design this effect, it should be a reflection of what scry is: it involves potentially important decisions, it's weaker than draw but still very useful, and there's a meaningful difference between scry 1 and scry 5.

    1. Scrying does phenomenal work in normal Magic, mitigating a huge issue. It's not as necessary in LBD because the format erases that issue almost entirely. Crutches are vitally important to a person with a leg injury, but aren't needed at all once the leg is healed. The argument that LBD needs scry to always do something because scry does so much in other formats is like arguing that we should keep the crutch after we're better.

      It is fair to say, though, that a card with scry loses value in LBD. Fundamentally, that's fine because everything's value shifts in LBD. But if we were designing cards for LBD, or a new game around LBD, it's certainly true that we'd never put a mechanic on them that rarely does anything. In that light, it's understandable to want scry to do something.

  11. Therefore as a starting point to think about, my first proposal is as follows: to scry x, shuffle both packs, then you look at that many cards from the top of each pack. You may exchange any number of cards from the "top-pile" of one booster pack with a card from the "top-pile" of the other pack.

    This is a significant increase in complexity and interaction from having no effect at all. Now you're interacting with information by thinking of hatedrafting and what your opponent might prefer or what you might prefer, or how the packs will get picked away from by the time the cards you rearranged will be the strongest picks left in the pack- for instance, will my opponent have the stronger of the two, assuming the order doesn't change? Sure, there will be many meaningless scry decisions when you reveal two chaff cards that are equally useless, but an analogous situation happens quite frequently with normal scry. In that sense it can be said to faithfully recreate the spirit of scry. It's not a perfect solution by any means, but I wanted to start somewhere.

    If you got this far then thanks for reading my insanely long essay.

    1. edit: I worded my description of "scryswap" badly. The same number of cards is moved into and out of each pack, so for each card you put from one "top pile" one card from the other "top pile" replaces it.

    2. I love the intent behind this suggestion. Being able to shift a card or two between packs could make a big difference and is an impressively faithful analog to what scry does elsewhere. The details of the execution sounds pretty cumbersome and would hurt the pace of play, but if an iteration of it were fast to execute, I'd try it out in a heartbeat. Nice.

  12. tl;dr

    scrying is cool because reasons
    I wish scrying did something in live draft :/
    Scrying should do: to scry x look at x many cards from the top of each pack, then swap any cards from one pack with an equal number of cards from the other
    now scrying is cool again :)