General Fixing the Die Roll

[Eric Chan]

I'm gonna start with the premise: the die roll matters in cube, and it matters quite a bit. Without any hard statistical evidence in my back pocket, I can't tell you exactly how much it matters. But I've seen enough anecdotal evidence over the years to tell me that it's not an insignificant factor. If your cube is like mine - where aggro decks frequently have the other player within lethal range by turn four, even if they then take a little while longer to actually close out the game - then going first is pretty durned important. Running your Loyal Cathar into a Rune Snag is pretty disheartening; on the flip side, knowing your Bitterblossom is all but immune to countermagic because you're a master at rolling sixes is a great feeling.

How much your actual match win percentage is increased by your winning the die roll, I couldn't tell you. It could be 2%; it could be 5%. It can and almost certainly varies by cube. But I'm just going to move forward with the assumption that, with the speed of cubes in the here and now, you want to win the roll and you want to be on the play.

So, having said all that: is there a way we can compensate the player on the draw for their disadvantage in tempo, aside from handing them one extra pity card?

I've been thinking about this topic for quite a while, and I'm not really a design wizard like Mr. Waddell, so it's a nut I've never been able to crack. But then last week, Hearthstone came out on the iPad. I installed it and dicked around with it for a bit, and while the actual merit of the game itself remains in question, I'm intrigued by how they chose to handle the die roll situation. It's clear that Blizzard has been thinking about this problem for a while, too, as they address it with two creative rules.

1. The player on the draw gets to see their extra card before they make mulligan decisions.
   
   
2. The player on the draw starts with a Lotus Petal in hand. (The game doesn't enforce a maximum hand size.)

The first rule is subtle, but having that extra nugget of information is obviously helpful when it comes to making the call on a questionable opener. The second rule is a more blatant attempt to even out the tempo disadvantage, as it lets the player on the draw play a two drop immediately on turn one (!), or just save it up and accelerate a key five drop one turn early.

I'm wondering if one or both of these rules could be applied to cube. The main downside is probably one of complexity creep; both of these kind of fall outside of the realm of typical house rules, and might be easy to forget, or to misinterpret. I can see the first rule being confusing if players are forced to peek at the top of their decks. I suppose I could just tell the player on the draw to always draw up to eight cards, and then to skip their first draw step as well. The second rule feels like it has a greater chance to 'break' something fundamental about Magic - WHAT DO YOU MEAN HE GETS A FREE LOTUS PETAL?! - but I also feel like it directly addresses the problem better than any harebrained scheme I've ever come up with. I could print out a bunch of those new Gold tokens, and have one available for each match.

So this is where you guys tell me if I'm off my rocker. Is this a problem you've ever thought needs addressing? How far-fetched do these gimmicks sound? Lay it on me, I can take it.

 

[Changling Bob]

This is totally me hypothesising out loud, but I kinda feel like this is the side effect of making agro 'good', when comparing it to control and midrange decks that wouldn't look super out of place in constructed. Of course, this isn't constructed, and draft control decks typically don't have the same sort of early game defenses together that a constructed deck does, due to both what comes around in the packs, but also the sorts of things we include in cubes.

On the one hand, we provide control with varied tools for varied situations, barter in bloods, cryptic commands, terminus, entreat the angels, planeswalkers and so on. On the other hand, we jam the meanest, fastest agro cards (and lots of them!) because otherwise agro decks are 'bad'. So when the agro deck comes together, it's hyper efficient compared to the control decks, because you could draft *experiements* six one drops that can have more than one power, fifteen proactive two drops, and a curve that only tops out at 5 because zealous conscripts is great and I didn't draft any fours. (nb: I probably misbuilt the deck I'm referring to).

Lets go the other way now and draft control/midrange. So I drafted some kind of UG control-y stuff. It's not very good because I don't know how to draft properly, but the same is true of the agro deck, so hopefully they modulo out. But if we examine this, we have seven creatures that can come down turn 2, three bounce spells, and a turn three counterspell. I drafted another 3 creatures to put in, but one is wild nacatl, that well known 1/1 for G in UG. The only thing that'll save this against my agro deck is a small miracle (and me building with less top end).

Now, again, I suck at drafting complex things, so it could well be that you could draft something with much more early presence despite being a mid-game, controlling or combo deck, but you bleed out a certain number of deck slots to doing that instead of doing the control-y thing.

So if you look at my cube list (which I do have to stress isn't actually built or played with), I don't have the punchy agro decks that everyone says they want, but neither does it have unstoppable top end, runaway card advantage planeswalkers, or super-dangerous midrange (although I bet you could build that last one in cbob's cube constructed). In theory, my cube plays out like good WotC block limited, rather than a proxy for constructed, but given it is a limited format, I feel like that's probably ok. Agro creature decks are still going to exist, they're just going to have to interact and have contingency plans against having their early game neutered.

Pinch of salt, YMMV, etc.

 

[CML]

I'm gonna challenge your premise. What sort of statistical edge do you think you get from playing first in, say, Theros limited, or RTR limited?

 

[Dom Harvey]

If you push aggro and (in/)directly punish control by cutting down on cheap removal, games will naturally come down to the die roll a lot more.

Planeswalkers are the biggest offender here, since it's much easier for the advantage they create to 'snowball' into something insurmountable.

 

[Eric Chan]

Well, for the first two years of my cube's existence, I was a pretty savage power-maxer. All of the good cards, and multiples of them! Even then, I had a nagging feeling that the die roll was skewing results more than I'd like. Interestingly, I wonder if the midrange attrition mirrors, including planeswalker on planeswalker battles - where tempo is less of a concern than being the person with the last permanent standing - are a place where you'd want to draw, so that you can win the late-game topdeck wars.

CML, I know that you're going to bring up mtggoldfish.com - which I'll agree is an excellent resource - but I'm not sure if random Limited formats are the best comparison for cube. There are a couple of formats in recent years where I recall choosing to draw, though these days all the Limited formats glom together in my head, so I couldn't tell you which formats those were. What do the numbers look like for, say, Block? The last couple seasons of Standard? Even Modern and Legacy?

I suppose what I'm doing is challenging the assumption that the advantage of going first is exactly offset by drawing an eighth card. I know this rule has been around since forever, and at this point it's commonly accepted as fair. But I want to peel back the cover and figure out why one extra card is more or less equivalent to half a turn in tempo. What evidence do we have that this leads to balanced games? If we figure that the games are actually balanced, can we assume that this is true across all formats - limited, constructed, and otherwise? How about across the myriad matchups from decks of various archetypes?

If there's any amount of inequity - however slight it may be - can we fix it?

Maybe a simpler question to start off with is how often you choose to draw when you win the die roll in cube. If our revised premise is that everything is balanced down the middle, you should be choosing to draw roughly half of the time. Your opponents should be doing the same. Would you say that this is a better indicator of whether the die roll matters?

 

[Dom Harvey]

The last bit might be true in ideal world where the win percentage is exactly 50%, but if it's slanted even 1% in favour of playing first it's basically never correct to draw.

 

[Eric Chan]

Yeah, that's kind of what I'm getting at. Even if there's only a 1 or 2% advantage, I'm wondering if there's something minor we can do for the player on the draw to try and recalibrate.

Maybe my strong language in my original post kind of exaggerated the degree to which this problem affects my cube environment. I don't think the advantage gained by going first anywhere close to double digits. It's definitely closer to the 1-3% range. But it's still just perceptible enough to make me want to fix it.

 

[Eric Chan]

Calvin just mentioned that if I were to actually implement the Lotus Petal / Gold rule, he'd pretty much choose to be on the draw every time as a control deck. Turn three Wrath of God, here we come!

Yeah, okay. You got me there. That is clearly too over the top. What works for Hearthstone can't translate directly to Magic, I suppose.

 

[ahadabans]

I think without hard numbers, trying to rely on whether guys choose to go first or second is not going to be accurate information. Perception is more important in that case and perception isn't always reality. I saw a study a long time ago made with emergency responder types (I don't remember if it was EMT's or cops or whatever) and most of them felt that there were crazier things that happened on full moons than at other times of the month. Statistically though, there was no variance. As human beings, we look for patterns even if there aren't any because chaos is not something we readily accept.

Not saying that is necessarily true for Magic. I agree that there are certainly match-ups that benefit more from the half a turn of tempo versus the extra card. And I think the aggro scenario you described in your first post is a solid example. But I think the better way to combat that is to either slow your cube down just a touch and/or add more removal so that the guy on the draw has more options.

I think I've mentioned this in previous posts, but I don't particularly care for hard aggro. I think it's boring to play and boring to play against because there just aren't a lot of play decisions. If you only get to see 9 cards in your deck, either you have an answer for the aggro player or you don't. It's not like you have a ton of game play choices and how you make them determines the outcome. It's really about do you draw enough answers before the other guy kills you. For me personally, I think the game is better when it develops a bit and the outcome then becomes more about how you use your resources.

It's a double edged sword though. If you slow the game down too much, you make the best strategies more control oriented. And it becomes just about how can drop more bombs. It's a hard thing to balance.

My 2 cents is more removal is almost always good for your meta because it gives guys answers and options. Even if you stifle some of the more fragile arch types, I think games just play better. There's too much broken stuff in cube. If you can't consistently interrupt things, then it just becomes who can assemble the broken synergy first. I personally don't find that fun.

 

[FlowerSunRain]

I'm perfectly ok with decks losing to aggro if they fail to interact at all with the opponent for three straight turns. That seems fine. Even just nuking one of the decks threats will slow their clock by a couple of turns, which in turns gives the deck more opportunity to develop and answer. Its the cards that remove of the other player's ability of meaningfully interact (Vortex, Winter Orb, Tangle Wire) that bother me and these cards are probably also the ones that make the die roll most important.

If the first turn advantage is meaningful, then giving the 2nd player a meaningful counterbalance seems fine. I think the free lotus petal is probably too far. A free scry 1 before turn 1 might be appropriate. Or is it too much to give the second player an extra card AND an extra filter? I have no idea, but I think whatever the bonus is it needs to be worth significantly less then a card.

 

[Diakonov]

Hearthstone also addresses another "flaw" inherent in the game of Magic that is drawing lands. I've been starting to pay closer attention to how many games I win simply because I draw fewer lands than my opponent (within reason, of course). I don't know the statistics but I imagine there is a noticeable correlation.

But, eh. That's how Magic goes. Unless it turns out that the die roll is really affecting our cube games by as much 5%, I would probably take the same perspective here as I do with drawing lands. If it IS by 5%, then I agree with what's been said about finding solutions within the card pool first.

 

[ahadabans]

Mana flood/screw is definitely a weakness in the game. And I think it's why I really like cycle effects and scry because both really help in this area (some of my favorite lands are the cycle lands).

I had a friend I used to play with regularly and we hated mana screw so much that we came up with our own resource land rule. I think I posted it before, but here it is again.

Any card in your hand can be played face down as a "resource land". It has tap: add 1 to your mana pool. Playing a resource land counts as your land drop. If a resource land would leave play for any reason, remove it from the game instead. If you control more resource lands than non-resource lands, sacrifice a number of resource lands until you control the same number.

Those last two rules were there to prevent bounce or sacrifice shenanigans and to make it so you couldn't run zero land artifact decks. This rule worked well and we had way fewer game decided by mana screw at least.

 

[Spiderdoofus]

Hearthstone is not the first game to remove lands, every CCG Richard Garfield designed after Magic was this way. Some solutions work better than others. In general, I think the crystals in Hearthstone don't really solve the problem at a gameplay level. Sure, it feels worse to lose to mana screw, but getting guaranteed mana means curving out and tempo become more important, and so winning becomes dependent on drawing in a different way. A similar problem exists in Sol Forge, which has no resource system. If you draw weaker cards (i.e., lower level) than your opponent for one or two turns, you will lose. Although it's better psychologically to be able to play your spells and removing lands helps, these games tend to just move variability to another place. FWIW, I think Netrunner has one of the more interesting solutions to removing lands.

Personally, the die roll thing doesn't bother me that much. I mean, someone has to go first, and I think the statistical edge is small. I wish it were more of a choice, but I don't feel bothered.

I think a possible solution might be to give players a choice. If you go second, you draw seven cards and begin the game with a Lotus Petal in the command zone. During your first draw phase, you can chose to replace your draw by taking the Lotus Petal and placing it in your hand.

 

[Changling Bob]

Netrunner's great at this, due to having actions per turn, and each side gets click for a card or a credit (~mana) by default, so economy cards (~acceleration, except that it's much more necessary) end up being about effort compression, rather than 'you didn't get it so you died', it just takes a bit longer and you're drawing gas instead.

Netrunner is great, play Netrunner. Also a Garfield design, fwiw.

 

[CML]

i don't disagree, the higher power a format is, the more important it becomes to go first. so it'd be correct to draw into the t8 (with a worse seed) for basically every limited PTQ ever, but in Legacy you'd sometimes play (with a possibility of missing).

my own suggestion is best-2-out-of-3 Rock-paper-scissors (a concession to the constructed h8rs), one-two-shoot west-coast style not this one-two-three-shoot atlantic BS

 

[Jason Waddell]

Eric, why are you rolling 6-siders and not D20s?

 

[ahadabans]

D6's are so much cooler than D20's. I don't like D20's.

In fact, it's my least favorite polyhedral. I'd rank them in the following order from best to worst:
1. D6 (old school classic - can't beat just a plain old cube with dots on it).
2. D12 (looks badass and in every system I've played, when you roll a "12" you did something gnarly).
3. D10 (percentile dice are fly)
4. D8 (not much to say - it's better than the next two)
5. D4 (sucks balls in actual use because you can't really roll it - though it does look like a pyramid so at least it looks cool)
6. D20 (too many damn sides and it's been made cliche by D&D. Fuck. This. Die).

 

[Eric Chan]

We actually roll two D6s and add up the sum. I bring a pile of six-siders because they're more handy than larger dice for +1/+1 counters and such. The few D20s I have are usually snapped up for use as life counters.

And hey, even MTGO uses D6s - not that they should be seen as the gold standard for anything, but there you have it.

 

[Miasmir]

We roll D12 to see who goes first and I'm very happy with it.

I am, however, very near to implementing the alternative mulligan rule where you go to 6 scry 1, then to 6, then 5 scry 1 etc. Just wary of introducing any more complexity when we're trying to add some newer players into the fold.

 

[Jason Waddell]

We actually roll two D6s and add up the sum. I bring a pile of six-siders because they're more handy than larger dice for +1/+1 counters and such. The few D20s I have are usually snapped up for use as life counters.

And hey, even MTGO uses D6s - not that they should be seen as the gold standard for anything, but there you have it.

MODO also never has players tie and reroll. The dice are rigged! And the shuffler too!

 

[CML]

i swear modo's coin-flips actually are rigged, though

 

[Changling Bob]

d12 is best die, and anyone who disagrees is objectively wrong.

 

[Jason Waddell]

 

[Alex Blanton]

I received a d30 as a present a few weeks ago, and I've been using it for die rolls as much as possible. I usually end up using 2d6 so I don't waste time searching through my sea of dice, but the novelty breaks ice with people you don't know well.

If you really want to be absurd though, you bring that d100 sitting in your LGS for God-knows-what.

 

[ahadabans]

I had a D30. I could go make coffee and eat a bagal while I waited for it to finish rolling. That thing was a practical joke.