Animate Dead > Recurring Nightmare

Status: OPEN, Castle AI Version: 1.0.2

Midrange creature based reanimator deck took Animate Dead over Recurring Nightmare. I'm guessing it's because Animate Dead is shown as a combo with Fauna Shaman, but for some reason Recurring Nightmare isn't. Maybe this is subjective but I don't see it being particularly close. Recurring Nightmare is a significantly more powerful card and has high synergy with this version of the reanimator deck.

2015-07-04 10:35
Recurring Nightmare is such a unique card & hopefully the tags to make it drafted appropriately can be added. Not sure how much work it is but it's very high in combo value with many Cube cards. At first I thought you were proposing that Animate Dead was better & I had my downvote button ready =)
2015-07-04 15:25
Hah! I couldn't figure out how to convey what I wanted in 35 characters or less, and it got a little muddy.
2015-07-05 06:19
At the very least Recurring Nightmare should have a 6 rating for quality, I think. Also, doesn't Recurring Nightmare have the same relevant tags that Animate Dead does? What's different about the tags that caused it to be rated lower for synergy and strategy?
2015-07-05 23:18
It has similar tags, I would imagine costing 2 CMC vs 3CMC was the deciding factor. However, I agree the quality for Recurring Nightmare probably should be higher and maybe introduce a tag for "combo enters the battlefield trigger (creature)". Cards like Blade Splicer and Shriekmaw should definitely push the AI towards Nightmare over Animate.
2015-07-08 17:01
In this case we can see that the AI has rated Recurring Nightmare as a better card from a Quality point of view (as has been discussed). There is also some consideration to Animate Dead being lower on the curve. If you check out the Strategy Guide you'll see that this strategy much prefers two drops over three. However the main influencing factor here is the fact that Animate Dead has been marked as a Combo and Recurring Nightmare has not. This in incorrect. Essentially the AI detects combos using a threshold on the "Two Way Synergy" between two cards. Synergy itself is a measure of the pick ratio (pick count / pass count) for any two given cards, but it is directional i.e. "If I have Voltaic Key and I then pick Time Vault" is a different stat to "If I have Time Vault then I pick Voltaic Key". Two Way Synergy is a weighted average of these two statistics. The reason that the AI uses this particular threshold is that a flat out list of two card combos would simply be two cumbersome to maintain. By inferring whether something is a "combo" based on how real Cube Tutor drafters have picked, we've got something that is constantly evolving and doesn't require maintenance. That does come with a cost. It isn't 100% accurate at detecting combos and this is one of those cases. I'd be happy to open up this method of detecting combos for discussion. What would you do differently?
2015-08-15 03:04
If you use a confidence interval rather than a simple pick ratio, the threshold approach should be more effective with smaller sample sizes. Had mentioned this here:
Please Login to post a reply.
Thumb Up 12
Thumb Down 0
Login to vote on this report.

Strategies Before Pick:

Primary: Golgari Reanimator    Secondary: Orzhov Aggro


#Card NameReasoning
1 Animate Dead Quality[3]AVERAGE  Strategy[10]STRONGEST  ComboSTRONGEST  Synergy[3]STRONGEST 
2 Go for the Throat Quality[3]STRONG  Strategy[10]STRONGEST  Synergy[3]STRONG 
3 Recurring Nightmare Quality[3]VERY_STRONG  Strategy[10]VERY_STRONG  Synergy[3]VERY_STRONG 
4 Herald of Torment Strategy[10]VERY_STRONG  Quality[3]VERY_WEAK  Synergy[3]AVERAGE 
5 Gray Merchant of Asphodel Strategy[10]WEAK  Quality[3]AVERAGE  Synergy[3]AVERAGE 
6 Wheel of Fortune Quality[3]AVERAGE  Strategy[10]NONE  Synergy[3]WEAK 
7 Qasali Pridemage Strategy[10]NONE  Quality[3]WEAKEST  Synergy[3]AVERAGE 

Previous Picks: