Deck Measuring Metric
General forum
Posted on June 30, 2016, 3:47 a.m. by Postmortal_Pop
I was inspired by a post in the EDH section about whether or not a player way of describing his deck matched the metric other players used. This for me thinking, could a set metric be designed what would ideally measure the strength of any deck in comparison to other decks of the same or similar format?
All we would need is a list of factors, based off of the cards used in the deck, that could be easily checked and graded in a 1-10 scale. Then you could average the scores out for a grade and use it to compair with an opponent.
Before I go further, while it would seem easier to grade on a win/loss ratio, I feel that this is more of a grade on the pilot. An amazing deck can bomb if the pilot doesn't understand it, and I've seen plenty of budget decks destroy matches with the help of an experienced player.
So far I've come up with this.
Mana base - this measures the consistency of land in your deck. 1 being all basic land with no filters and 10 being a perfect balance of multicolored or utility lands
Consistency - counts the number of cards integral to your win and/or the amount of like minded cards. (If your deck Mills, how much of it Mills, if it's an infect deck, how many ways can you get poison Snag through?)
Perfection Clock - if you had the perfect draw, how many turns would you take to win?
Perfection Chance - what is the statistical chance of you getting a winning combo in your opening hand?
Survivability - how easy is it to break your combo? (weird paradise combo can win on turn 2 but it does to Natural State.)
I also believe a list of grading curves should be used to balance out inherent shortcomings.
Color Saturation - until more cards are printed, more then 2 colors will always have a less perfect mana base that 2 or less colored decks. Try as you like, there will always be a day when you draw 
land and green creatures.
Intended Speed - aggro decks are designed specifically to make the game short, Lantern of Insight decks are designed specifically to make the game long. This intentional difference shouldn't place one above the other if both can be reliable.
This is far from a finished list, and I would love to hear if I've missed anything or if you have a suggestion for other points that could be graded on.
Souljacker says... #3
If you wish to compare different archetypes, you need to find the scale they balance on.
For example, mono vs multicolored decks can be measured on the scale of manabase <> power level. A multicolored deck usually has a higher power level, but a worse mana base (slower/hurtful). A monocolored deck has a good manabase but usually lower power level because it has less cards to chose from. So if you rate decks on both manabase and power level, you can compare the average of those ratings with eachother.
Another scale could be speed vs consistency. A combo/aggro deck can be very fast, but less consistent (not having a combo piece, or having infect creatures but no pump spells). A control deck is slow of course, but super consistent.
I think these 2 scales give you a great starting point to compare different decks in a format, but to use it effectively you need to calibrate the grading levels by using different decks you have a good idea about their relative rating, and see if you input the parameters into your model you'll get good representations of that expected rating.
June 30, 2016 7:28 a.m.
I agree with Boza in that points 4 (and 5) are better evaluators than the others, but disagree that this is "devolving". If you create decks solely based on this, it would, but having an aid at the time you finish a deck and have not tested it... that could be helpful/useful.
What concerns me is that repeated use of "win" and "winning". We forget that EDH is a social format, where winning is of course the ideal result, but not-winning sometimes means a lot more fun (if you have played with your girlfriends you'll get what I mean).
I support the notion of working on a way to evaluate and score the consistency and survivability of a deck. The other parameters are sometimes too subjective, while consistency's not. I'll crunch some numbers and ideas, maybe we manage to create something here.
Cheers people!
June 30, 2016 7:38 a.m.
I think that a collection of metrics that measure different things would be the most helpful.
These could always be weighted and summed for single number comparison based on the number of players, group play style and values, or other particulars of a situation.
Different people and situations emphasize these in different ways. These are all fairly separate from each other and combining them could lead to loss of clarity in interpretation outside of a specific context or situation.
Some useful measurable things might be:
Consistency (mana base, synergy density, tutors and combo redundancy)
Interference (counter spells, steal effects, removal, wipes, impact on board and game States of others)
Early game presence (combo risk turn clock, density of aggressive moves, super ramp cards like Sol ring and mana crypt)
Maximum Power level (ability to close a game out / lack thereof, number of players that could be taken out in a single turn, amount of damage generated in a single turn)
Resilliance (recursion, protection of permanents, turn clock to recover from a board wipe, redundancy)
Flavor (density of theme, uniqueness of cards and strategies, potential for stories worth telling to come out of matches)
June 30, 2016 7:53 a.m.
See, I can get behind this - general areas of interest and deck choices. I do not think there is any feasible way to assign points or grading to these and they can be highly subjective, but it looks a lot better than the previous version.
June 30, 2016 7:57 a.m.
The main problem with any of these scales comes when somebody simply says, "ok can you define what a 6 out of 10 means consistently, and how it differs from a 7?"
Arbitrary scales like these like interpersonal reliability. What I may think is a 7, you may think is a 5. It may also lack temporal reliability. Ask me if something is a 5 out of 10 on one day and I may agree. Ask me the same question in 3 months time and I might disagree.
Arbitrary scales like these are just pretty bad for a whole host of reasons but mainly because no one can agree with what they actually mean.
June 30, 2016 8:30 a.m.
Souljacker says... #8
You could have people here vote on the scales, then you get an average with a large sample size. It doesn't mean it's objective, but might be fun and gives some insight.
June 30, 2016 9:13 a.m.
I think a single number (including TO voting - look at what some of the most popular decks are) holds little clear meaning. Sure, it can be fun.
I do think something like consistency based on quality of land base, % of cards that contribute to strategy synergy, and number of tutors can be measured and understood. Each of these could also be separate too.
June 30, 2016 9:21 a.m.
For the 'perfection clock', are you including perfect plays from other players? I have a Merieke Ri Berit commander deck that would take about 15 turns to win if nobody played anything, but could otherwise end it all by turn 4 or 5. Also, how about a category based on price?
June 30, 2016 9:35 a.m.
But how does quality of the land base (subjective, but not everyone can run ABUR duals to score highest), % of cards that contribute to the strategy (is a deck with 60% cards that contribute to the strategy necessarily better than one that has only 50% of cards contribute?), number of tutors (says nothing really, since any deck running black can outperform everything else in that regard) tell you anything?
Decks do not need to pass SATs and go to college, they can do just fine without being graded! The point of the exercise is the same as the tiering of EDH generals - gives you no overall value.
June 30, 2016 9:49 a.m.
I had more in mind to characterize than to rank, judge, or grade.
Land base quality: A high quality land base allows a deck to be played out with more predictability and reliability with less variation from color screw, especially in early turns.
Strategy density: Decks that are built so that every single card is required to contribute to the deck goal are more likely to win in more or less the same way every time, hence being described as "consistent". This is not intended to mean "better" since there are cases where a divergent set of win conditions is more strategic.
Tutors: since edh is a singleton format, all other factors being equal, decks and games tend to have far more variation than the typical 60 card constructed decks. Tutors reduce this variation significantly by providing a higher probability of key cards (especially combo pieces) seeing play in any given game.
June 30, 2016 10:26 a.m.
So, something a deck would have for example high land base quality, medium strat density and low tutorability?
June 30, 2016 10:30 a.m.
Something like that, yeah. Then for people that want to, they could combine more detailed descriptors into more general ones like "consistency", "predictability", "pressure", " interference", "resilliance".
I think it would have a feel like aggro/control/combo/midrange descriptions more than tiers.
I also think it would be important to make sure the refinement of the scale is not misleading. High/medium/low might be enough. Certainly a scale out of 100 would be unlikely to meaningfully differentiate a 76 from a 77.
Boza says... #2
Well, most of these seem to relate to combo decks, which is not bad, but it limits observation on other decks.
Mana base - bad metric for monocolored decks - if a mono-green aggro deck has 20 Forest, it scores a 1. But if it has 17 forests and 3 Pendelhaven, it scores almost perfectly?
Perfection clock - not valid for anything but combo and/or hyper aggro.
Perfection chance - same as above.
Consistency - finally a metric that is useful, at least for non-control decks. This is a bad way to measure consistency. It should be effiency of wincons compared to turn they want to be cast. Ie a Standard Esper Dragons deck does not want to win before turn 10, for example, so having 4 Dragonlord Ojutai as its only win con is A-OK. But a Modern Burn deck will not score highly if it relies on Hidetsugu's Second Rite as a win condition.
Survivability - once again, only combo decks are concerned and even then, it is not universal. Natural State can break up the Voltaic Key + Time Vault Vintage combo, but does this make it a bad combo? Every single combo can be broken by one card (storm loses to Mindbreak Trap, Dredge loses to Leyline of the Void, etc.), so that is not realiable at all.
Color saturation - that rolls up into manabase.
Intended speed - you yourself state how despite having different intended speeds, different archetypes can score the same, making this not useful at all.
So, not to be the bearer of bad news, but I suggest to start over, as this really does not do anything. I suggest to take 2 decks, think of NEW metrics to apply to it, then score those decks against each other. Start empyrically on this one.
IMO, I do not see this idea bringing anything exciting to the table and actually developing this into something meaningful will make deckbuilding meaningless by devolving it to numbers.
June 30, 2016 4:34 a.m.