Bluetracker

Tracks Blizzard employees across various accounts.

What kind of total dork would write a program to simulate the effect of a horrible Snap card, to try to optimize its (horrible) performance? https://t.co/4CeVt2SAUH

What kind of total dork would write a program to simulate the effect of a horrible Snap card, to try to optimize its (horrible) performance? https://t.co/4CeVt2SAUH

Chadd Nervigg

Posted 2 years, 11 months ago (Source)

What kind of total dork would write a program to simulate the effect of a horrible Snap card, to try to optimize its (horrible) performance? https://t.co/4CeVt2SAUH

Chadd Nervigg

Posted 2 years, 11 months ago (Source)

So anyway, I think this is the theoretical best Agatha deck, if your goal is for Agatha to play the whole game, and you have my collection (CL 1000). https://t.co/XEqjTyHZe4

StormWindBoy

Posted 2 years, 11 months ago (Source)

@Celestalon This is awesome! I'd love to see a writeup of the program - were locations factored in? What cards White Queen might give you? And I assume we ruled out taking control of the game if Agatha is discarded/played early (out of principle)

Chadd Nervigg

Posted 2 years, 11 months ago (Source)

@StormWindBoy This is purely based on total stats thrown at the board, simulating the hand/draw/plays each turn. No locations or other random effects. White Queen given card is expected to be irrelevant, since Agatha always picks herself. I suppose it could give you a 5drop, but that's ignored

Chadd Nervigg

Posted 2 years, 11 months ago (Source)

@StormWindBoy (And yes, excluding cards that could potentially discard Agatha early, because then she's not playing.)

SirLanik

Posted 2 years, 11 months ago (Source)

@Celestalon Huh. I would have expected Quicksilver and Domino to be part of it. I look forward to testing Agatha when I finally get her.

Chadd Nervigg

Posted 2 years, 11 months ago (Source)

@SirLanik Saw them in lots of others' Agatha lists. It's not that they're necessarily bad, just that those decks tend to try to take advantage of the guaranteed 1&2 by filling the rest of the deck with 3/4/5s. That just isn't the max power route. Wastes energy, and lower power/energy.

Chadd Nervigg

Posted 2 years, 11 months ago (Source)

@SirLanik Going with a whole bunch of 1s and 2s results in very little wasted energy, and you have lots of the high power per energy 2/4s and 1/2s. It's just better to swap a 3/4 for a 2/3+1/2, until you start running out of cards.

RickMKick

Posted 2 years, 11 months ago (Source)

@Celestalon Neato. Tell us (programming nerds) what the program looked like?

Chadd Nervigg

Posted 2 years, 11 months ago (Source)

@RickMKick Just a C# console app that I fed a ton of potential decks to, iterating manually to find the highest output.

<blockquote class="quoted">
 
</blockquote>
What kind of total dork would write a program to simulate the effect of a horrible Snap card, to try to optimize its (horrible) performance? https://t.co/4CeVt2SAUH (<a href="https://twitter.com/i/web/status/1605848658586398722">Chadd Nervigg</a>)
<hr/>

<blockquote class="quoted">
 
</blockquote>
So anyway, I think this is the theoretical best Agatha deck, if your goal is for Agatha to play the whole game, and you have my collection (CL 1000). https://t.co/XEqjTyHZe4 (<a href="https://twitter.com/i/web/status/1605849042927333376">Chadd Nervigg</a>)
<hr/>

<blockquote class="quoted">
 
</blockquote>
@Celestalon This is awesome! I'd love to see a writeup of the program - were locations factored in? What cards White Queen might give you? And I assume we ruled out taking control of the game if Agatha is discarded/played early (out of principle) (<a href="https://twitter.com/i/web/status/1605853335449239552">None</a>)
<hr/>

<blockquote class="quoted">
 
</blockquote>
@StormWindBoy This is purely based on total stats thrown at the board, simulating the hand/draw/plays each turn. No locations or other random effects. White Queen given card is expected to be irrelevant, since Agatha always picks herself. I suppose it could give you a 5drop, but that's ignored (<a href="https://twitter.com/i/web/status/1605855270881480704">Chadd Nervigg</a>)
<hr/>

<blockquote class="quoted">
 
</blockquote>
@StormWindBoy (And yes, excluding cards that could potentially discard Agatha early, because then she's not playing.) (<a href="https://twitter.com/i/web/status/1605855417921257472">Chadd Nervigg</a>)
<hr/>

<blockquote class="quoted">
 
</blockquote>
@Celestalon Huh. I would have expected Quicksilver and Domino to be part of it. I look forward to testing Agatha when I finally get her. (<a href="https://twitter.com/i/web/status/1605855430273400833">None</a>)
<hr/>

<blockquote class="quoted">
 
</blockquote>
@SirLanik Saw them in lots of others' Agatha lists. It's not that they're necessarily bad, just that those decks tend to try to take advantage of the guaranteed 1&2 by filling the rest of the deck with 3/4/5s. That just isn't the max power route. Wastes energy, and lower power/energy. (<a href="https://twitter.com/i/web/status/1605858446649065472">Chadd Nervigg</a>)
<hr/>

<blockquote class="quoted">
 
</blockquote>
@SirLanik Going with a whole bunch of 1s and 2s results in very little wasted energy, and you have lots of the high power per energy 2/4s and 1/2s. It's just better to swap a 3/4 for a 2/3+1/2, until you start running out of cards. (<a href="https://twitter.com/i/web/status/1605858969888530432">Chadd Nervigg</a>)
<hr/>

<blockquote class="quoted">
 
</blockquote>
@Celestalon Neato. Tell us (programming nerds) what the program looked like? (<a href="https://twitter.com/i/web/status/1605928306930380800">None</a>)
<hr/>

<blockquote class="quoted">
 
</blockquote>
@RickMKick Just a C# console app that I fed a ton of potential decks to, iterating manually to find the highest output. (<a href="https://twitter.com/i/web/status/1605971804643659781">Chadd Nervigg</a>)
<hr/>

Tweet

What kind of total dork would write a program to simulate the effect of a horrible Snap card, to try to optimize its (horrible) performance? https://t.co/4CeVt2SAUH (<a href="https://twitter.com/i/web/status/1605848658586398722">Source</a>)

So anyway, I think this is the theoretical best Agatha deck, if your goal is for Agatha to play the whole game, and you have my collection (CL 1000). https://t.co/XEqjTyHZe4 (<a href="https://twitter.com/i/web/status/1605849042927333376">Source</a>)

@Celestalon This is awesome! I'd love to see a writeup of the program - were locations factored in? What cards White Queen might give you? And I assume we ruled out taking control of the game if Agatha is discarded/played early (out of principle)

@StormWindBoy This is purely based on total stats thrown at the board, simulating the hand/draw/plays each turn. No locations or other random effects. White Queen given card is expected to be irrelevant, since Agatha always picks herself. I suppose it could give you a 5drop, but that's ignored (<a href="https://twitter.com/i/web/status/1605855270881480704">Source</a>)

@StormWindBoy (And yes, excluding cards that could potentially discard Agatha early, because then she's not playing.) (<a href="https://twitter.com/i/web/status/1605855417921257472">Source</a>)
 
 <hr/>

@Celestalon Huh. I would have expected Quicksilver and Domino to be part of it. I look forward to testing Agatha when I finally get her.

@SirLanik Saw them in lots of others' Agatha lists. It's not that they're necessarily bad, just that those decks tend to try to take advantage of the guaranteed 1&amp;2 by filling the rest of the deck with 3/4/5s. That just isn't the max power route. Wastes energy, and lower power/energy. (<a href="https://twitter.com/i/web/status/1605858446649065472">Source</a>)

@SirLanik Going with a whole bunch of 1s and 2s results in very little wasted energy, and you have lots of the high power per energy 2/4s and 1/2s. It's just better to swap a 3/4 for a 2/3+1/2, until you start running out of cards. (<a href="https://twitter.com/i/web/status/1605858969888530432">Source</a>)
 
 <hr/>

@Celestalon Neato. Tell us (programming nerds) what the program looked like?

@RickMKick Just a C# console app that I fed a ton of potential decks to, iterating manually to find the highest output. (<a href="https://twitter.com/i/web/status/1605971804643659781">Source</a>)
 
 <hr/>