Most efficient creature 3

[fas723]

See next post for explanation.

Ranked creatures:

	Name	True Cost	Calc Cost	Diff Cost	Index
1	Blue Gremling	7	10,01	3,01	1,43
2	Dwarf Panzergarde	11	13,82	2,82	1,26
3	Tataree	6	8,77	2,77	1,46
4	Gargoyle Sentry	11	13,61	2,61	1,24
5	Ichthellid	9	11,52	2,52	1,28
6	Guardian Angel	12	14,39	2,39	1,20
7	Kralathor, The Devourer	16	18,18	2,18	1,14
8	Ludwig Boltsorm	13	15,16	2,16	1,17
9	Earth Elemental	20	22,15	2,15	1,11
10	Psylok	8	10,08	2,08	1,26
11	Ravenous Ghoul	13	15,05	2,05	1,16
12	Asyrian Cleric	5	7,05	2,05	1,41
13	Whirling Spirit	12	14,02	2,02	1,17
14	Thunderift Falcon	6	7,91	1,91	1,32
15	Flaming Hellion	13	14,78	1,78	1,14
16	Darkfenne Hydra	16	17,74	1,74	1,11
17	Deathfang	8	9,68	1,68	1,21
18	Necropian Vampiress	16	17,63	1,63	1,10
19	Emerald Tegu	9	10,60	1,60	1,18
20	Screech Harpy	11	12,54	1,54	1,14
21	Dire Wolf	12	13,52	1,52	1,13
22	Gray Angel	12	13,43	1,43	1,12
23	Mana Leech	8	9,42	1,42	1,18
24	Firebrand Imp	5	6,36	1,36	1,27
25	Darkfenne Bat	5	6,35	1,35	1,27
26	Bitterwood Fox	5	6,33	1,33	1,27
27	Shaggoth-Zora	8	9,22	1,22	1,15
28	Zombie Crawler	4	5,00	1,00	1,25
29	Dark Pack Slayer	13	13,92	0,92	1,07
30	Thornlasher	7	7,89	0,89	1,13
31	Dwarf Kriegsbiel	11	11,78	0,78	1,07
32	Spitting Raptor	11	11,77	0,77	1,07
33	Orc Butcher	8	8,69	0,69	1,09
34	Timber Wolf	9	9,64	0,64	1,07
35	Skeletal Minion	5	5,58	0,58	1,12
36	Goblin Builder	5	5,55	0,55	1,11
37	Tarok, the Skyhunter	13	13,54	0,54	1,04
38	Goblin Bomber	8	8,46	0,46	1,06
39	Iron Golem	13	13,43	0,43	1,03
40	Venomous Zombie	7	7,40	0,40	1,06
41	Vine Snapper	7	7,23	0,23	1,03
42	Zombie Minion	7	7,20	0,20	1,03
43	Skeletal Sentry	8	8,18	0,18	1,02
44	Acolyte of the Bog Queen	5	5,16	0,16	1,03
45	Makunda	17	17,12	0,12	1,01
46	Sosruko, Ferret Companion	7	7,09	0,09	1,01
47	Grey Wraith	10	10,05	0,05	1,00
48	Steelclaw Grizzly	17	17,04	0,04	1,00
49	Brogan Bloodstone	15	14,95	-0,05	1,00
50	Galador, Protector of Straywood	16	15,95	-0,05	1,00
51	Skeletal Knight	13	12,85	-0,15	0,99
52	Togarah, Forest Sentinel	21	20,73	-0,27	0,99
53	Mort	16	15,69	-0,31	0,98
54	Feral Bobcat	5	4,57	-0,43	0,91
55	Devouring Jelly	13	12,55	-0,45	0,97
56	Goran, Werewolf Pet	15	14,51	-0,49	0,97
57	Goblin Grunt	4	3,40	-0,60	0,85
58	Royal archer	12	11,29	-0,71	0,94
59	Zombie Brute	11	10,13	-0,87	0,92
60	Samandriel, Angel of Light	21	20,08	-0,92	0,96
61	Highland Unicorn	13	12,03	-0,97	0,93
62	Cervere, The forest Shadow	15	13,96	-1,04	0,93
63	Plague Zombie	9	7,75	-1,25	0,86
64	Unstable Zombie	9	7,72	-1,28	0,86
65	Bridge Troll	13	11,69	-1,31	0,90
66	Sir Corazin, Bladmaster	16	14,63	-1,37	0,91
67	Mountain Gorilla	16	14,57	-1,43	0,91
68	Valshalla, Lightning Angel	21	19,54	-1,46	0,93
69	Knight of Westlock	13	11,46	-1,54	0,88
70	Moonglow Fearie	8	6,42	-1,58	0,80
71	Gorgon Archer	16	14,26	-1,74	0,89
72	Goblin Slinger	7	5,12	-1,88	0,73
73	Giant Wolf Spider	15	13,01	-1,99	0,87
74	Adramelech, Lord of Fire	24	22,00	-2,00	0,92
75	Redclaw, Alpha Male	16	13,97	-2,03	0,87
76	Malacoda	16	13,81	-2,19	0,86
77	Thorg, Chief Bodyguard	17	14,74	-2,26	0,87
78	Stonegaze Basilisk	12	9,67	-2,33	0,81
79	Skeletal Archer	11	8,48	-2,52	0,77
80	Fella, Pixie Familiar	12	9,03	-2,97	0,75
81	Invisable stalker	15	11,79	-3,21	0,79
82	Hugin, Raven Familiar	11	7,74	-3,26	0,70
83	Thoughtspore	8	4,56	-3,44	0,57
84	Grimson Deadeye, sniper	15	11,04	-3,96	0,74
85	Selesius, the East Wind	21	16,56	-4,44	0,79

[fas723]

Here it is again, the third version.

This time I won’t go into details how the actual calculation is done. I will only publish the table. For you who like to know you can start read the first two “Most efficient creature” topics.

http://forum.arcanewonders.com/index.php?topic=7979.msg8010#msg8010

You can also take a look at the excel which you find here:

https://drive.google.com/file/d/0Bz0fnLKKUKlxM2oyMzlNVWZWQ28/edit?usp=sharing
It is the 3.3 version.

However I can tell you what is new in this version. This time I have focused a lot more at the traits. The equation is the same as before, but with one more exponent at the trait product. In earlier calculations I have used traits evaluated by myself to rank which ones were better than the others. This resulted in some feedback for you guys. So this time I have done a loop for the traits as well.

Basically each trait has been tested several times in order to generate the most accurate curve. Initially I just let the loop run until it found the optimum point, but that didn’t turned well at all. Depending on all the other attributes each creature had it more than often resulted in strictly worse traits to be evaluated higher than better traits, like Double strike being higher rated than Triple strike and so on. To prevent this I made a few rules that traits with fewer occurrences than 5 were bundled together in its group and had a pre-set difference before the loop started. This was a success and now all the traits are also mathematical evaluated.

Again I will point out that this rank does not consider any synergies or combinations of cards or abilities. A None living creature will for sure be better with the Idol in play. This calculation does not take these aspects into account. However the trait None living have been evaluated according to all the other traits and properties, and finally optimized towards each creature cost.

[Aylin]

First, how did you arrive at the "values" for attack type, traits, or even the weights given in your formula? Best I can tell is the numbers were arbitrarily chosen.

Second, the value of any spell is going to be determined by the mage's traits/abilities, other cards in the spellbook, and the current meta. Any chart that cannot take that into account (ie all charts unfortunately) will never be accurate.

[Zuberi]

I am not sure what information I am gleaning from this chart, and therefore don't know how it could be helpful. It doesn't match my idea of efficiency in creatures, which measure's how much I can expect to get out of a creature before it dies and compares that to its mana cost. In that regards, Tataree is most definitely not a very efficient creature. I am paying 6 mana to get maybe 2 or 3 healing/channeling out of her before she dies to a single unavoidable attack. Yet you list her as the 3rd most efficient creature? I call shenanigans.

I personally like Piousflea's suggestion to one of your previous charts. He suggested that you forget about the traits, and calculate the expected mana cost of each creature based purely on their Life, Armor, and Attacks. You can then compare this to their actual cost and determine how much of a premium you are paying for the abilities and traits that they actually have. I don't believe this would allow you to rank them in any significant manner either, but at least it would give some useful data.

[The Dude]

I am not sure what information I am gleaning from this chart, and therefore don't know how it could be helpful. It doesn't match my idea of efficiency in creatures, which measure's how much I can expect to get out of a creature before it dies and compares that to its mana cost. In that regards, Tataree is most definitely not a very efficient creature. I am paying 6 mana to get maybe 2 or 3 healing/channeling out of her before she dies to a single unavoidable attack. Yet you list her as the 3rd most efficient creature? I call shenanigans.

I personally like Piousflea's suggestion to one of your previous charts. He suggested that you forget about the traits, and calculate the expected mana cost of each creature based purely on their Life, Armor, and Attacks. You can then compare this to their actual cost and determine how much of a premium you are paying for the abilities and traits that they actually have. I don't believe this would allow you to rank them in any significant manner either, but at least it would give some useful data.

But then, we would all know that Steelclaw grizzly is the greatest thing since Buffy stopped airing!

[MrSaucy]

I am inclined to agree with Aylin on this one.

The problem I have with these calculations is that traits are too difficult to evaluate absolutely and numerically. Is having the "fast" trait better than having the "regenerate 2" trait? Is having the "vampiric" trait better than having the "flying" trait? You can't know for sure because the helpfulness of traits vary from game to game. Sure, some creatures can seem overall better than others, but summoning the right creature for the job is more tactically valuable than summoning the creature that is "usually" better. It would be a shame if we discredited creatures simply because they weren't always useful.

In a card game like Magic The Gathering, calculating creature efficiency was important because you never knew which creature you were going to draw (if any), so you had to hand pick the creatures that were "overall" better. The beauty of a game like Mage Wars is that you have the power to select any creature you want at any time, thus allowing you to include creatures in your spellbook that are useful under certain (and perhaps narrow) circumstances.

Just my two cents.

[MrSaucy]

I personally like Piousflea's suggestion to one of your previous charts. He suggested that you forget about the traits, and calculate the expected mana cost of each creature based purely on their Life, Armor, and Attacks. You can then compare this to their actual cost and determine how much of a premium you are paying for the abilities and traits that they actually have. I don't believe this would allow you to rank them in any significant manner either, but at least it would give some useful data.

This is a great suggestion. I honestly feel like most of the creatures in MW are balanced enough such that any weaknesses with respect to damage output, armor rating, life total, and/or mana cost are compensated with helpful traits. For example, the Invisible Stalker, ignoring his invisibility, is not all that great of a creature. But with the invisibility trait he can be devastating.

Perhaps ranking creature usefulness with respect to mages is more helpful than ranking creature usefulness overall. For example, for a mage like the Warlord, creatures like Goblin Grunts are fairly useless since the Warlord is not built to swarm.

I am assuming the calculations given didn't take into account how many spellpoints it costs to include each creature, another important variable to consider.

[Aylin]

Perhaps ranking creature usefulness with respect to mages is more helpful than ranking creature usefulness overall. For example, for a mage like the Warlord, creatures like Goblin Grunts are fairly useless since the Warlord is not built to swarm.

Instead of mage, it should be made for mage archetypes. Solo or Thoughtspore Forcemaster and Grizzly Forcemaster are both the same mage, but they play very differently. Cards that would be useful to one might be useless to another.

[MrSaucy]

Instead of mage, it should be made for mage archetypes. Solo or Thoughtspore Forcemaster and Grizzly Forcemaster are both the same mage, but they play very differently. Cards that would be useful to one might be useless to another.

Another good idea.

[jacksmack]

nr 1 and number 85 are correct.

everything else is incorrect.

(Some would argue that the last spot is shared between selisius, mountain gorilla, goblin bomber and Makunda  - all creatures i have never seen summoned)

[Aylin]

nr 1 and number 85 are correct.

everything else is incorrect.

(Some would argue that the last spot is shared between selisius, mountain gorilla, goblin bomber and Makunda  - all creatures i have never seen summoned)

I've used Selesius a few times. A ranged sweeping attack is nothing to scoff at.

Makunda will get better when some level 2 and level 3 cats are introduced.

[ChimpZilla]

As always, thanks for the analysis fas. Despite everyone's objections, we need more of this, not less.

BUT

The moneyball approach doesn't work well in Mage Wars versus Magic the Gathering:

1. MtG has progressive mana mechanics. You start at 0 and build up to cast more expensive spells as the game goes on. MW is flat/regressive and gives you a large reserve to start. You can reasonbly play the equivalent of an Emerakul in the first two turns. In MtG, you cheat it in like a degenerate or you ramp your b**** off.  It's why going swarm (i.e. goblins, fish, merfolk, WW) is more viable in that game. Creature cost determines how quickly they see play, and playing more efficient threats than your opponent is actually relevant.

Which brings us to:

2. MW is wayyyyyyy more match-up dependent. Earth Elemental is probably the most efficient of the 20 drop fatties, but he's currently usable in maybe only two builds, and Darkfenne Hydra is an awful draw. The boogeyman of the meta, Iron Golem, bricks a lot of creatures above him. Sir Corazin, rated well below, eats him alive one-on-one (Yes, I've tested this in many game situations, including ones that favor IG, and he comes out on top). Overrated by the community? IMO, sure. But the forums have been in pro/anti-Piousflea mode since Bashcon and IG is the poster child of Piousflea hate.

[fas723]

Ones again strong opinions about my tables. I like. 

First I will punch you all down in the "trait discussion".  I have not set any values for anything. Anything! Not for traits or coefficient.  This is simple math. The only thing I have done is to tell the algorithm to consider obvious rankings like Double strike is worse than Triple strike. The rest is set by the math and the convergation towards the best possible curve for all creatures. So, if "Regeneration 3" is higher evaluated than "Fast" it will only tell that the creatures with "Fast" have higher values in Armor, Attack and Health compared to creatures with "Regeneration 3". This is how AW has set the parameters and how the numbers falls out based on the creature pool we are currently looking at. Simple, but NOT arbitrarily chosen.

I also thought about Piousflea's suggestion and remove the traits totally. Problem is you can not compare it towards the true cost in that case (the traits are naturally included there), and that means you can not do the comparison at all.

I never had MtG in mind when I did this. I didn't even know this type of evaluation existed for MtG. The little I know about MtG is what I learned playing as a teenager 15-20 years ago.

I will point this out again (as in the first post): This table doesn’t tell you which creatures are the best one in any given situation. It will tell you how much Attack, Armor, Health and abilities/Traits you will get out of your invested mana.

[Zuberi]

It will tell you how much Attack, Armor, Health and abilities/Traits you will get out of your invested mana.

Unfortunately I believe your chart has failed then. There are quite a few highly ranked creatures that provide very few substantial benefits.

I personally gain no useful information from your chart. I don't mean that to be offensive, and I am really sorry because I can tell you have put a lot of thought and effort into creating this...but I can't think of a single thing I can gain from it.

I also thought about Piousflea's suggestion and remove the traits totally. Problem is you can not compare it towards the true cost in that case (the traits are naturally included there), and that means you can not do the comparison at all.

I think there might be a slight bit of miscommunication here. We're not suggesting that you ignore their traits when comparing their statistics to their current mana cost. We're suggesting you try to figure out how much they should cost based on their statistics if they didn't have any traits. Then we could compare their expected cost to their actual cost to determine how much extra you are paying for the extra traits.

I'm not completely sure how you would go about reverse engineering this information. But I think it could probably be done.

[Aylin]

Ones again strong opinions about my tables. I like. 

First I will punch you all down in the "trait discussion".  I have not set any values for anything. Anything! Not for traits or coefficient.  This is simple math. The only thing I have done is to tell the algorithm to consider obvious rankings like Double strike is worse than Triple strike. The rest is set by the math and the convergation towards the best possible curve for all creatures. So, if "Regeneration 3" is higher evaluated than "Fast" it will only tell that the creatures with "Fast" have higher values in Armor, Attack and Health compared to creatures with "Regeneration 3". This is how AW has set the parameters and how the numbers falls out based on the creature pool we are currently looking at. Simple, but NOT arbitrarily chosen.

And I'll ask again, how did you arrive at the numbers for any of the variables you use? How did you arrive at your equation? How did you test it?

You're asking people to just accept your graph without giving any of your methodology. It doesn't work that way.

Not to mention the fact that your chart has several entries that are obviously wrong? Psylok being the 10th "most efficient" creature? Adramelech being the 74th? There are some serious problems with your work.