Hello,

First post here!

(TL:DR - ADHD isn't perfect. Can we make adjustments to the values assigned to cards? Read below for examples.)

Like many of you, I have found the ADHD (ABCE) rating system quite useful as a baseline comparison and evaluation tool for Keyforge decks. (Don't know what ADHD is? Read this ) I want it to be understood from the outset that I am not trying to be disparaging of this resource, I think it is very helpful for what it is. However, I am trying to be critical. Having myself reviewed the Google Sheet ( here ) used by the creator to generate these ratings, I think the ratings given to cards need to be reevaluated. I want this system to be even better at accurately supplying baseline analysis of Keyforge decks, and it is to that aim that I raise the following questions and scenarios of accuracy and, what I perceive to be, inconsistency:

Example A

According to the last known database I was using, examine the following:

Here we can see two almost identical cards, with identical ratings. However, the cards are not identical. Coward's end is conditioned upon the creatures destroyed being damaged. Considering the kind of nuances other card ratings try to incorporate, I think Coward's End should be reflected at a lesser Board rating than Gateway to Dis. Objectively, Gateway to Dis is a better card, and the rating should reflect that.

Disclaimer : I understand that trying to assign objective values to subjective cards is not what the ADHD system is trying to accomplish. However, I think that even within the system of logic it uses to assign ratings there are inconsistencies.

Example B

Another example of inconsistent ratings:

This argument is a little more subjective. Ganger Chieftain and Gauntlet of command have the same effect, except that Ganger Chieftain is upon play, and GoC is an artifact (so only usable on subsequent turns). However , how is it that Ganger Chieftain is rated a 1 in Efficiency, and GoC a 0.67 for the same effect? My suspicion is in consideration of Ganger Chieftain being immediate, and GoC being delayed, but I don't think I've ever played a game of Keyforge where I didn't ready one creature with my GoC. It's unreasonable to put this at less than 1. An Efficiency rating of 1 is to the effect of drawing a card, or using another creature (for instance, Inspiration has an E rating of 1). Giving GoC a rating of 0.67 is saying that you only get to ready and use a creature with it 2 out of 3 games. Which is absurd. You will at least get one use out of this artifact. This is also supported by the fact that it is hard to remove artifacts, so it is unlikely that after playing a GoC that you will not have it on a later Brobnar turn. Now, an argument could be formed that Inspiration reads "ready and use", whereas GoC is "ready and fight", thus without targets to fight it is less useful, so a less rating. But if we were to concede that point, then Ganger Chieftain should not be on par with Inspiration.

Example C

Also, the logic used to grant GoC a lesser E rating than Ganger Cheiftain in the previous example is reversed here:

I don't know why these two cards should have different Efficiency ratings, except that Invasion Portal stays in play, meaning you may get multiple uses out of it. However, as compared to the above example, where the Artifact got the worse Efficiency rating as compared to a creature, I don't understand why the Artifact here gets a better rating than an Action. Theoretically an action is recycled even faster than a creature. So if we use the previous example's logic that the creature's Efficiency is perceptibly higher than an Artifacts, surely an Action's would be even higher? I also disagree (as with GoC) that Sound the Horns should be less than "1" efficiency. Sure there are times when you play it without Brobnar creatures in your deck, but I think it makes sense to assume you play it once per game to draw a Brobnar creature. And if not, then why should Invasion Portal not have a less than 1.00 Efficiency rating?

Example D

Another example:

Why does Tocsin have a Board rating, but Succubus does not? They have the same body-size, 3 power, so the only consideration could be towards their different effects upon the board (Tocsin discarding a card when reaping, Succubus reducing hand-size by one). However, Succubus' value actually occurs sooner than Tocsin's (upon your opponent's draw card phase, as opposed to the next time you ready and reap with your Tocsin), which makes it a more immediate threat on board. I would argue these creatures have very comparable board presences. For Tocsin to be rated a full value of 1 (when average Board ratings for a deck are 18.25, so Tocsin constitutes almost 5.5% of your deck's Board Control), and for Succubus to have a rating of 0 doesn't seem to follow logically. Especially when you examine how other creatures with a B rating of 1 control the board, Tocsin is quite underwhelming.

Example E

This example is more direct:

Again, Library Access has an E rating of 1.00, assuming you play a Logos card after playing Library Access and draw a card, and Battle Fleet has an E rating of 0.67, assuming that, 1 out of 3 times that it's played, the idiot playing his deck didn't have another Mars card in his hand...this is just absurd. If it is assumed that Library Access draws you a card upon playing it (i.e. that you had one other Logos card in your hand), then it should be assumed that Battle Fleet draws you a card (i.e. same assumption, that you had one other Mars card in your hand).

If anything, Library Access should have an Efficiency rating of 2, and Battle Fleet something like 1.67, or even 2 itself. Library Access having an E of 1 is putting it at the same level as a Silvertooth . I think anyone who's played Keyforge gets the sense that resolving a Library Access generates more Efficiency than a Silvertooth. The scenarios where you play Library Access with only one other Logos card in hand are so narrow (does not ADHD try to stay away from such narrow situations?) that I think it warrants a higher E value.

Example F

There are not only inconsistencies amongst like-comparisons between cards, but also just blatant mistakes. For instance:

Technically speaking, as often as Neuro Syphon steals aember (the instance that gives it a Control rating), it should also be drawing a card. yet it has a 0.33 for Control, and a 0.00 for Efficiency. I think it makes sense that if the card steals 0.33 aember, then it should draw 0.33 cards.

Example G

There are other instances of questionable ratings, or ones that don't make sense. And this not even for cards that are overly complicated, synergistic, or hard to assign a value to. For instance:

Blood Money has an Aember rating of 1...when the card generates 2 aember. So, in effect, ADHD is saying that half the time that Blood Money is played, it has no effect, assuming there are no creatures to play it on. Well where is that logic coming from? That logic is not applied across the board. We don't see other cards having their abilities halved as a baseline assumption. For instance, Dust Imp , another card that grants 2 Aember upon a creature's death, has an Aember rating of 2. The same goes for Truebaru (a very contingent card, more contingent, I would say, than Blood Money). Why can't we assume that upon playing Blood money (as all of the ratings are assumed that we've played the card) that it will eventually return to us 2 aember? We do so with both Dust Imp and Truebaru...

Why do we care?

I care because when the numbers are this inconsistent, it feels like a disservice to the users of this system. Especially when people are basing purchases on ADHD ratings, or price points for selling their decks. Or even just using ADHD to evaluate what a strong deck is. Whether people should be doing this or not is another question, but the fact remains that they are. I would like to propose submitting an edited ADHD list to the group/creator of the system. I would create my own edits, and would welcome any one else who'd like to submit edits or raise questions to do so here in this post. Of course, I'd first have to get into contact with said group. I would also welcome any dissenting arguments to what I've laid out. I have no doubt that other people will not agree with me. I tried to remain as object as possible in my analysis when considering such subjective matter, as I know the creator(s) of this system must have likewise endeavored to do. I simply think that now that we've had time to evaluate and play more games, we have a better understanding of the actual value of cards.

Thanks for reading, and if anyone who is affiliated with the creator(s) of ADHD reads this, thanks for your work in creating and maintaining such a great resource!

Gateway vs End

Cowards End kills UNdamaged Creatures, not damaged ones, it is entirely possible to end up with no creatures on your opponents side and several remaining on your side. conditionally better but not as easy to set up so them being even is a fair representation.

Chieftan vs Gauntlet

Brobnar wants to smash with dude as fast as possible, Cheiftan does this, even allowing use of a creature just played the same turn, speed is the key for Brobnar as it has few other tricks I find other than big dudes smashing face. additionally chieftan is a creature itself so it can also reap when needed giving it more versatility. again id go with the ADHD assessment.

Portal vs Horns

Mars is a slower house so long term incremental gains are far more powerful in this house than a 1 shot. especially when considering a lot of Mars creatures play abilities require ready Mars creatures in play to trigger, so activating this the turn after you play it might be preferable in a lot of cases. again Id go with the ADHD assessment.

Tocsin vs Sucubus

random discard is considerably more powerful that limiting hand size, even if it is delayed. if you hit power cards in their hand like bait & switch etc. it can turn a game in your favor. admittedly Sucubus is a passive effect that works even when you don't call house Dis for your turn giving it some weighting but I think ADHD assessment is perfectly fine here too.

the remaining assessments are more abstract rating issues so I wont comment on them.

these are of course my own opinions, as your assessments are your opinion, and all are up for debate. I'm sure the person who wrote the program will consider your discussion points and make changes if they feel they are warranted.

I know this isn't the answer you are looking for but I think you are overthinking the ADHD system.

I will admit that even I like to look up a new deck just to see what it says, but that's where it ends. I don't give it too much weight when I start looking at a deck I got trying to figure it out. The more I play this game the more I think there are several factors it doesn't account for and in a unique game where each play is essentially unique there is just not a good solid way to analyze a deck and figure it out at a glance.

All the analysis in the world will not replace experience. Each deck is a puzzle and the only way you're truly going to figure them out is by putting in a lot of sessions with them.

My favorite deck is still the first one I got. According to ADHD it's a slightly below average ACE deck. However its resourcefulness in finding ways to get me aember while delaying my opponent is fantastic and I have a lot of fun playing it. I figured out how to play it very quickly and when I loooked at its ADHD (well after 10 games with it) it was very clear to me that the strategy of the deck was not reflected in that analysis at all.

My "worst rated" deck is my second favorite to play and boasts all kinds of tricks to keep an opponent on edge. This means nothing but on a cool note it has 6 rares! Had I gone by the ADHD (mostly negatives) I would have put it back in the box and never used it, and now there's a good chance it could staple itself into the 3-deck format tourneys that I go to.

My "best" rated deck is inconsistent despite being rated a mid/high-end ACE. The most recent deck that I've opened and played with has a lone positive (B rating of +21) yet one look at the deck list and any opponent has no trouble preventing that board control. So even then its biggest strength according to analysis is easily negated.

All that to say that the system is a nice starting point, but in most cases after I had played games with the decks I found the analytics to be an inaccurate gauge compared to what I had learned through playing with them.

So I wouldn't be too worried about how it rates individual cards and I don't think the system needs revamping, but take that with a grain of salt as it's just my opinion. For those that use and enjoy it, keep on, and I think the creator of it has a lot to be proud of in creating a resource that many love and many more (like me) still utilize even if briefly.

tl;dr - It's fine as is .

@Madgremlin

It demonstrates that the author of this system was fallible, and the system is too flawed to trust.

8 minutes ago, Derrault said:

@Madgremlin

It demonstrates that the author of this system was fallible, and the system is too flawed to trust.

And therefore, each person who wants to use it needs to do one (or more) of these three things:

1. Accept it for what it is, and make adjustments based on gameplay.
2. Reject it and develop their own system.
3. Tweak it and publish the results of that tweak for others to criticize similarly 😏

10 minutes ago, Derrault said:

@Madgremlin

It demonstrates that the author of this system was fallible, and the system is too flawed to trust.

That is a slightly negative way of looking at it.

I think what it does demonstrate is that the mechanics, cards and the entire game are more complex than what the ADHD Algorithm extrapolates, and that is a good thing.

There are clearly differing views on what cards are better than others, and certain cards become considerably more powerful when paired with other specific cards and that is what makes this game so wonderfully complex.

Also as mentioned above, you are more than welcome to ignore ADHD entirely, that is your choice. Me I choose to view it as just one more tool to assess decks strengths and balances critically and independently above and beyond my own biased gut feelings and initial impression of a deck and to try to understand them better and to see where I might be playing the deck in a different manner to optimize its performance.

If an analytical model gets too complicated, the numbers become as hard to read as the system it attempts to model. In choosing the 4 variables to rate, and what those numbers mean, you necesarily have to abstract some elements that wont be represented. I applaud the second look.

I dont know the original thought behind the system, but I can make up reasoning for your points. Which is how any system of analysis is going to be.

Both cowards end and gateway have roughly equivalent board control - you will toss 2 enemy cards on average in one cycle. (I think 3 might be a little more realistic). Cowards end is better in that you can damage yourself to take away some of the sting. There isn't enough nuance in abce to capture that. When you need a board clear, whether you end up with a gateway or a cowards, you are going to play it.

For invasion portal vs gauntlet of command, the portal runs you through your deck, which can have value. Having that on an artifact may be more efficient than ready and fight, where you dont always have something worth fighting. Sound the horns you are just playing a card to draw one. It might have been better to just have a creature in that spot, other than pushing for the reshuffle. If it were me, these would all be 1s though. I think e is slightly undervalued overall.

Succubus has no board rating imo because she dies before you use her for anything. Too weak to be a priority. Then again, Tocsins reap ability isnt really captured. Discarding your opponents cards in hand could be a kind of board control.

Library has more efficiency than battle fleet because the card you draw might be logos, and if it is, you get to library access again. I dont know how good the numbers are. In either case, you might draw into the wrong house and then it got you nothing, so...

Neuro syphon draws .33 cards, but as with library access, that card will only be efficient .33 times. Too low to get an e value in this system.

Blood money is going to give you 2 if you can put it on something you can kill. Probably better than 1, but probably lower than 2 (adhd is on average... what does that mean when matchup is so important?) Maybe 1.66?

13 minutes ago, saluk64007 said:

If an analytical model gets too complicated, the numbers become as hard to read as the system it attempts to model. In choosing the 4 variables to rate, and what those numbers mean, you necesarily have to abstract some elements that wont be represented. I applaud the second look.

I dont know the original thought behind the system, but I can make up reasoning for your points. Which is how any system of analysis is going to be.

Both cowards end and gateway have roughly equivalent board control - you will toss 2 enemy cards on average in one cycle. (I think 3 might be a little more realistic). Cowards end is better in that you can damage yourself to take away some of the sting. There isn't enough nuance in abce to capture that. When you need a board clear, whether you end up with a gateway or a cowards, you are going to play it.

For invasion portal vs gauntlet of command, the portal runs you through your deck, which can have value. Having that on an artifact may be more efficient than ready and fight, where you dont always have something worth fighting. Sound the horns you are just playing a card to draw one. It might have been better to just have a creature in that spot, other than pushing for the reshuffle. If it were me, these would all be 1s though. I think e is slightly undervalued overall.

Succubus has no board rating imo because she dies before you use her for anything. Too weak to be a priority. Then again, Tocsins reap ability isnt really captured. Discarding your opponents cards in hand could be a kind of board control.

Library has more efficiency than battle fleet because the card you draw might be logos, and if it is, you get to library access again. I dont know how good the numbers are. In either case, you might draw into the wrong house and then it got you nothing, so...

Neuro syphon draws .33 cards, but as with library access, that card will only be efficient .33 times. Too low to get an e value in this system.

Blood money is going to give you 2 if you can put it on something you can kill. Probably better than 1, but probably lower than 2 (adhd is on average... what does that mean when matchup is so important?) Maybe 1.66?

You meant to say Coward's end is "conditionally" better. It certainly isn't after an AoE damage card like Poison wave or Ammonia Clouds has been played. And certainly not better when you have no creatures on the board. In my opinion, Gateway is better because it does not have conditions on it at all.

To answer the question: Can we make adjustments to the values assigned to cards?

The spreadsheet is out there, we certainly could each copy it and modify it to fit our own ratings. But then we lose the part where it's comparative to other decks as more get analyzed using it.

There's already so much that the system can't take into account (like combos and synergies). There's not going to be a perfect analysis machine that's going to give everyone what they specifically want. You either accept it or reject it, and/or find or create something else.

I wondered about the rating of Nimquid the Fair (A:0, B:1, C:0, E:0) vs Champion’s Challange (A:0, B:2, C:0, E:.33). I thought they should be more closely rated to each other.

Thanks for starting the thread. I was going to message the creator of KeyForge Compendium, but forgot.

11 hours ago, msieder said:

You meant to say Coward's end is "conditionally" better. It certainly isn't after an AoE damage card like Poison wave or Ammonia Clouds has been played. And certainly not better when you have no creatures on the board. In my opinion, Gateway is better because it does not have conditions on it at all.

Sure, sometimes its better (you can protect your board) and sometimes its worse (their board is protected). In a system that tries to rate the average case (without getting into the weeds of calculating a true average), it makes sense to see them rated the same.

I think my biggest quibble with the system is efficiency. The stat needs to be there to round things out, but feels more subjective as so many types of thing have to be tracked there. The scores usually end up pretty low, even for highly efficient decks. I see people complain about a bad adhd score on thrir good deck, then I see a 9 or a 10 for E and do a "well, actually..."

I do think things like library access or recursion should probably get a higher E.

55 minutes ago, Starbane said:

I wondered about the rating of Nimquid the Fair (A:0, B:1, C:0, E:0) vs Champion’s Challange (A:0, B:2, C:0, E:.33). I thought they should be more closely rated to each other.

Thanks for starting the thread. I was going to message the creator of KeyForge Compendium, but forgot.

Numquid is capped to the size of your board. I don't know that a full 1 bonus to champions is fair...

11 minutes ago, saluk64007 said:

Numquid is capped to the size of your board. I don't know that a full 1 bonus to champions is fair...

Champions also gets E:.33, where Numquid doesn’t.

Numquid remains on the board to reap or fight to in future turns as well as removing your oppenent’s only creature if they have just 1.

Sometimes Champions is better, sometimes Numquid is better, that seems very KeyForge to me. I just question why Champion’s rating is so much better.

Another thing that I think ADHD does wrong is to evaluate some card power whose role is dependent of the deck composition.

A clear example: Full moon gives you an increase in the aembar rating, but it entirelly depends on the ammount of, mainly, wild creatures you have.

For example, I get two copies of full moon in a deck with only 2 creature of that house. That is reflected as a good aember value. In real life, that two cards are mostly useless and antisinergistic.

As opposite, if you obtain two full moon in a deck with 5 or 6 creatures of that house, the cards become much better.

12 hours ago, Polenesio said:

Another thing that I think ADHD does wrong is to evaluate some card power whose role is dependent of the deck composition. 

A clear example: Full moon gives you an increase in the aembar rating, but it entirelly depends on the ammount of, mainly, wild creatures you have.

For example, I get two copies of full moon in a deck with only 2 creature of that house. That is reflected as a good aember value. In real life, that two cards are mostly useless and antisinergistic.

As opposite, if you obtain two full moon in a deck with 5 or 6 creatures of that house, the cards become much better.

Yeah, knowing what you know about your deck, you might want to subtract 2 from the aember value as those cards may be dead most of the time. Although you still might play full moon to get an aember or 2 a few times. If you ever have 2 full moons and a single creature, you are getting the full value adhd would assign. A card-based rating system will never be able to handle this type of situation. I've been working on a synergy analyzer, but it's complex and will need to compare every card to every other card in your deck. I'm using heuristics, so full moon will give you a point for every untamed creature for instance, but it's still going to take ages before I can come up with something for every card. And I don't know if the final numbers will be useful for anything.

But the ultimate folly is that so much depends on what other cards there are in your deck, and how reliably they combo.

Take one of the other mass creature-destruction actions, Save the Pack. Is it better or worse than Gateway to Dis/Cowards End? I have no idea. But in a deck where it reliably combos with Co-operative Hunting (allowing me to selectively damage creatures before Saving the Pack) it is clearly better than it would be in a deck without the ability to damage opposing creatures. On the other hand, it is clearly objectively worse than Gateway/Cowards when I am playing against a deck containing lots of Sanctum creatures - but how does a spreadsheet capture THAT?