A Formula for Determining your Deck's Power Level

[Aetherium Slinky]

Problem is, that formula only works for a deck with an infinite combo that wins the game.

It only adds points to the strongest archetypes, that was the point. If you're running a consistent combo win con you deserve points for that. We see most cEDH decks follow this trend. You get a lot of points for efficient combos but less so for other types of combos.

You might be right about the compensation part 4->5 because we deduct points from the tutor category so you lose points in there anyway. That was just an afterthought that put more emphasis on the CMC of the deck and how fast it can threaten a win through the strongest archetype in the game; combo.

But all in all I think I neglected powerful deck archetypes like elfball that don't rely on that many tutors or combo in general. Your Ezuri deck sounds a lot faster than my Merieke so it would make sense to remove that category altogether. I calculated a 9.4 for Merieke using 4 as the denominator and cutting the last category about cards needed to win. Then 10.1 vs your 9.28 with the denominator 4 and keeping the category about cards needed to win meaning any way you count it you get a higher value for a lower powered deck in this case. I don't like it. The formula isn't probably very good at judging deck power.

I like the commander compensation, though. It accounts for the turn you're able to cast it through CMC considerations and it adds a lot of points for being a 'free card' when you start the game. Tutor and draw commanders definitely get an advantage here.

[WizardSpartan]

I like the commander compensation, though. It accounts for the turn you're able to cast it through CMC considerations and it adds a lot of points for being a 'free card' when you start the game. Tutor and draw commanders definitely get an advantage here.

Definitely. With Kumena being included as draw in my list, the total number goes up to a very reasonable 9.

But all in all I think I neglected powerful deck archetypes like elfball that don't rely on that many tutors or combo in general.

I think Ezuri elfball is hard to quantify with a formula, like you said. Ezuri is (somewhat) unique in that he takes mana and turns it into a win directly. Only 20 mana on a board of 4 elves likely results in an opponent dead, and it's extremely easy to get a board state just a little bit larger. I could honestly remove the 4 noncreature tutors I have from my list and replace them with more elves, and consistency just probably wouldn't go down. That's the advantage when all you need is fat mana to win and 8 cards in your deck provide fat mana.

[Mynus]

Well, I guess this makes all my decks super casual, which in turn means my playgroup is super casual, as I win more than the 25% expectancy of a 4 person game. I have put in a few decks, but I don't think I have any over 8, and most not over 6-7.

I tend to not run tutors for the very reason why they are so good. They tend to make each game the exact same; go get the 1-3 cards you need for a very specific win. I don't find this fun, I like the variability of the 99, I like having each game feel different.

I know 2cmc ramp is preferred, but I have no problem building in 3cmc ramp into my games. I also tend to play budget cards, which probably limits some things.

[WizardSpartan]

I tend to not run tutors for the very reason why they are so good. They tend to make each game the exact same; go get the 1-3 cards you need for a very specific win. I don't find this fun, I like the variability of the 99, I like having each game feel different.

I have so much trouble with this. I love the variability of not having tutors, but then I play like 3 games in a row and just don't draw the cards I need despite having statistically decent amounts of them. Therefore, I include tutors to make sure I get the cards I want, but then I'm seeing several cards from the same small pool of best cards because why would I tutor for any other piece of card draw when Yawgmoth, Thran Physician & Skullclamp do whatever all the other card draw do but just better?

[Slyvester12]

That's why I include silly win conditions in my decks. If I'm behind, I can always tutor for a board wipe or a counterspell, but if I'm ahead, I go for the silly win. In Yuriko, I can always tutor for Scroll Rack and make my opponents eat 11 damage every combat from the same Temporal Trespass, but going for a 40 damage Doomsday pile is more fun.

As I mentioned recently, my myr deck can go for infinite mana into a win very quickly, or I can compile a Rube Goldberg that shoots my opponents to death with three Myr Battlespheres and infinite Myr tokens.

My favorite is my Omnath, Locus of Rage deck. If I have 15 mana,  Boundless Realms,  Scapeshift, and Splendid Reclamation, I can do over 100 damage in one turn. Optimal? No. Memorable? Yes.

[Morganator 2.0]

We play casual commander because we like the subjective fun of the experience.  Not because we want to quantify it.

Speak for yourself. I play casual commander, and I still have an unhealthy obsession with numbers and stats.

If we're only aiming for a ballpark estimate this is something that's doable. I don't have time right now, but I might be able to throw together a rank-based equation, using the scale that Judaspriester and dexflux put together.

https://deckstats.net/forum/index.php/topic,49777.0.html

[Loggiu]

This formula doesn't work my korlash score only 7.7 with more than 25 discards and removal cards that means you fight only with your general against my infect creatures probably without enough lands. Is too much /20 for interactions if you play like smallpox....

[Slyvester12]

This formula doesn't work my korlash score only 7.7 with more than 25 discards and removal cards that means you fight only with your general against my infect creatures probably without enough lands. Is too much /20 for interactions if you play like smallpox....

No, I think that's fine. If anything, 7.7 seems high for a mono black voltron deck relying on targeted discard effects. Those don't translate very well to EDH, usually.

[Loggiu]

If it wasn't voltron the vote was lesser cause i've probably had more creatures. I can equip another creature with infect and I got sheoldred, call to the grave and smokestack that gives me only 3/20 plus all the other f***ing sacrifice cards, no problem with hexproof or indescrutible, for me the formula doesn't work. Doesn't count big cratures with infect or card that can take the table like sheoldred, that give points for me, a lot of cards that can take the table alone or an efficent strategy, and what about the elfball bug??? got lesser lands and more mana source means nothing....

[fire5167]

Hmm. I tried to make a more data-oriented version of the original equation. The more values a formula takes acknowledges, the more accurate it should be, so I added a few more. I based the equation on the Witch of Agnesi graph which avoids breaking the system when someone says their deck is composed of 99 ramp cards.

I'm not sure how to paste an equation here, so I have the LaTeX for it. But LaTeX is a pain to read on its own. If you Copy into Desmos Graphing Calculator, add sliders, and then set your data on the sliders, it should produce a rating for the EDH deck.

Here is the LaTeX:
y = \frac{2(10 + R_{RampCount})}{(L_{LandCount} + \frac{R_{RampCount}}{2}-37)^2 + 10(1 + L_{TapLandCount}) + 10(1 + R_{RampAvgCMC})}+\frac{10 + R_{RemovalCount}}{(R_{RemovalCount}-10)^2 + 10}+\frac{10(W_{WinConCount}+T_{TutorCount})}{(W_{WinConCount}-3)^2 + 1}+\frac{10 + C_{CardDrawCount}}{(C_{CardDrawCount}-10)^2 + 10(1+C_{CardDrawAvgCMC})} + 2

Obviously, it's not perfect, but it's slightly more comprehensive.

[Soren841]

My Sliver Hulk cEDH deck (sadly illegal because of Flash but it was my baby) scored 15.965. I think it's undervalued in the formula tbh. It doesn't take into account how many pieces the combo is, how much mana it is to carry out the combo, and the fact that the outlet is in the command zone.

^ this was with the 1st formula

[Soren841]

With the new formula I'm assuming that the wincon variable is 1 for Sliver Hulk because it's literally just Food Chain combo. That gives it a score of 39.7778529926 with the new one.

[fire5167]

Yeah, you're right about the Number of Win Conditions. Originally I was going with "how many ways can you go infinite" like if you run both

Demonic Consultation + Thassa's Oracle and
Godo + Helm of the Host
then the WinConCount = 2

So I guess the number of WinConditions in Sliver Hulk would be Food Chain + however many tutors you run. I think...

Obviously, that logic kinda gets destroyed when you take into account that not all combos have the same number of pieces, and clones like Lab Man can replace Thassa's Oracle and so on. Also, that's true about the differing mana cost of Win Conditions as well.

[Aetherium Slinky]

Yeah, you're right about the Number of Win Conditions. Originally I was going with "how many ways can you go infinite" like if you run both

Demonic Consultation + Thassa's Oracle and
Godo + Helm of the Host
then the WinConCount = 2

So I guess the number of WinConditions in Sliver Hulk would be Food Chain + however many tutors you run. I think...

Obviously, that logic kinda gets destroyed when you take into account that not all combos have the same number of pieces, and clones like Lab Man can replace Thassa's Oracle and so on. Also, that's true about the differing mana cost of Win Conditions as well.

There's something interesting about the numbers in this formula. Kinnan scored 99. It's the winconcount-term that gives immense power to combo decks with a few lines and tons of tutors. In Kinnan's case it's (10(17+2)) / ((2-3)^2 +1) = 190 / 2 = 95. That needs some adjusting.

[Soren841]

Well obviously an equation like this is always going to favor certain numbers. For example I tried 11 for wincons because that's how many total cards I have for winning including redundancy. My score went up when I lowered it to 1 (because there's really just 1 combo). But if u lower it further it goes down, and if u increase it to 2 it goes up a little. It assumes there's optimal levels for everything and the deck that's closest to that will always get the highest score. For some things this works (I'd say 28 lands with no taplands is pretty optimal objectively) but the amount of ramp in a Flash Hulk deck for example would be way lower than an Ad Nauseam/Storm deck.