Dominion Strategy Forum

Please login or register.

Login with username, password and session length
Pages: [1] 2  All

Author Topic: Seprix's Awful Simulations  (Read 10406 times)

0 Members and 1 Guest are viewing this topic.

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Seprix's Awful Simulations
« on: June 10, 2016, 03:49:06 pm »
+2

I've never really figured out Geronimoo's simulator, but it can't be all that tricky, right? Tips and tricks on how to make better stuff would be awesome.

My pet project so far is to get Castles to beat Big Money, which is a humble little project for sure. It might not even be possible, who knows. I'm trying to get the hang of all of the buy rules and stuff.
« Last Edit: June 10, 2016, 03:52:26 pm by Seprix »
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #1 on: June 12, 2016, 03:51:22 pm »
0

After I realized that my Castles strategy was buying Duchies and Estates and Provinces for no reason (I borrowed some lingo from basic Big Money and forgot to clear those), I wiped those and I noticed a noticeable improvement in my Castle Bot. It still is awful and cannot even beat Basic Big Money, but at least I've gotten the basic hang of Geronimoo's Simulator, right?

Conclusion 1: Pure Castles Big Money is freaking awful, don't ever do it. Basic Big Money beats it 100% of the time.

Real Conclusion 1: Castles BM is better than Basic BM, but it's still almost a coin flip

Now that the super obvious is out of the way, I will begin work on something else: Castles Engines, or Castle + BM + something else, like Wharf.

The real question is if Castles BM beats Scout Big Money.

Another edit: I increased the win percentage of Castles Big Money against Basic Big Money by a whopping 52.7%, because I simply neglected to have Castles BM pick up Provinces or Duchies after the Castles pile was empty. Dumb me. :)

the (bad) code: Buys Province, Duchy, Estate if the Castles runs out. Otherwise, buys Castles > Gold > Silver. I'll probably try some continued optimization with Castles BM instead so that it buys things that aren't Castles in the opening and beyond.

edit 2: Buys Province <= 4 in supply, buys Province, Duchy, Estates if Castles are out. Otherwise, buys Castles > Gold > Silver. It's a little better in improved odds, but still basically a coin flip.

edit 3: Continual improved odds! I simply programmed a buy Castle over Province if <= 3 CastlesInSupply and it jumped all the way to 67% winnings. Excellent, excellent. Continuing with optimization and screwing around with some numbers.

Big Money Ultimate (by WW) flattens my poor little Castles code 56% to 40%, so there's still work to do. :(
« Last Edit: June 12, 2016, 04:19:14 pm by Seprix »
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #2 on: June 12, 2016, 04:31:41 pm »
0

Another big breakthrough. It is bad to instantly go for Castles, no duh. However, It is strictly better to get at least 3 Golds before going hard for Castles. Adding that makes Castles BM beat Ultimate Big Money (WW) about 48-51% of the time, which is basically a coin flip, but that probably means there's more of a breakthrough to be done.

I also added a Buy Province rule if Opponent's VP < Castles Vp -6.
« Last Edit: June 12, 2016, 04:34:42 pm by Seprix »
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

liopoil

  • Margrave
  • *****
  • Offline Offline
  • Posts: 2587
  • Respect: +2479
    • View Profile
Re: Seprix's Awful Simulations
« Reply #3 on: June 12, 2016, 04:33:52 pm »
0

If you have all 8 castles you have:

8 + 2 + 2 + 2 + 3 + 7 + 5 + 16 = 45 VP, just short of 8 provinces. So I think it should skip provinces until castles are out. Not counting extra VP from grand castle.

It should probably also not open double castle. Try humble/silver instead, and buy gold over small/crumbling after that.

Playing small castle loses 3 VP even if you trash the crumbling castle. So do it only when it would make you gain the grand or king castle.

I think you can probably beat BMU upwards of 80% of the time. Castles are generally cheaper than provinces and only small castle and the last three really slow you down.
Logged

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #4 on: June 12, 2016, 04:36:00 pm »
0

I don't know how to program the bot to not play Small Castle, but I'll look into it.
« Last Edit: June 12, 2016, 04:41:47 pm by Seprix »
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

liopoil

  • Margrave
  • *****
  • Offline Offline
  • Posts: 2587
  • Respect: +2479
    • View Profile
Re: Seprix's Awful Simulations
« Reply #5 on: June 12, 2016, 04:41:44 pm »
0

Another big breakthrough. It is bad to instantly go for Castles, no duh. However, It is strictly better to get at least 3 Golds before going hard for Castles. Adding that makes Castles BM beat Ultimate Big Money (WW) about 48-51% of the time, which is basically a coin flip, but that probably means there's more of a breakthrough to be done.

I also added a Buy Province rule if the amount of points the opponent has is less than the Castles player by at least 7.
Your last rule sounds bad to me, especially because you could end up with an early lead. You will almost always win if you get all 8 castles, so focus on that.

My buy rules:

gold if castles in supply > 5
silver if total money < 16
castles
province if castles in supply = 0
duchy if castles in supply = 0
duchy if provinces in supply < 4
gold
estate if castles in supply = 0
estate if provinces in supply < 3
silver

I don't know how to program the bot to not play Crumbling Castle, but I'll look into it.
You should check some logs to see how it's playing Small castle. I'm not sure it's possible to program play rules, but a good one would be "play only if you have 6 or 7 castles"
« Last Edit: June 12, 2016, 06:44:49 pm by liopoil »
Logged

DG

  • Governor
  • *****
  • Offline Offline
  • Posts: 4074
  • Respect: +2624
    • View Profile
Re: Seprix's Awful Simulations
« Reply #6 on: June 12, 2016, 04:46:17 pm »
+1

You can post up simulator code by using the copy button in the simulator and using the # (insert code) function on these message boards.

Code: [Select]
<player name="Castles dg"
 author="DG">
 <type name="Bot"/>
 <type name="BigMoney"/>
 <type name="UserCreated"/>
 <type name="SingleCard"/>
 <type name="Generated"/>
 <type name="TwoPlayer"/>
 <type name="Province"/>
   <buy name="Gold">
      <condition>
         <left type="getTotalMoney"/>
         <operator type="smallerThan" />
         <right type="constant" attribute="19.0"/>
      </condition>
   </buy>
   <buy name="Silver">
      <condition>
         <left type="getTotalMoney"/>
         <operator type="smallerThan" />
         <right type="constant" attribute="16.0"/>
      </condition>
   </buy>
   <buy name="Castles"/>
   <buy name="Province"/>
   <buy name="Duchy">
      <condition>
         <left type="gainsNeededToEndGame"/>
         <operator type="smallerOrEqualThan" />
         <right type="constant" attribute="4.0"/>
      </condition>
   </buy>
   <buy name="Estate">
      <condition>
         <left type="gainsNeededToEndGame"/>
         <operator type="smallerOrEqualThan" />
         <right type="constant" attribute="2.0"/>
      </condition>
   </buy>
   <buy name="Gold"/>
   <buy name="Silver"/>
</player>
Logged

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #7 on: June 12, 2016, 04:50:39 pm »
0

Another big breakthrough. It is bad to instantly go for Castles, no duh. However, It is strictly better to get at least 3 Golds before going hard for Castles. Adding that makes Castles BM beat Ultimate Big Money (WW) about 48-51% of the time, which is basically a coin flip, but that probably means there's more of a breakthrough to be done.

I also added a Buy Province rule if the amount of points the opponent has is less than the Castles player by at least 7.
Your last rule sounds bad to me, especially because you could end up with an early lead. You will almost always win if you get all 8 castles, so focus on that.

My buy rules would be:

silver if you don't have any
castle if you don't have any
province if no more castles
duchy if no more castles
gold if you have fewer than 3 castles
castles
gold
estate if no more castles
silver

I might have done something wrong, but this code gives a 0% winning chance for Castles BM.

Code: [Select]
<player name="Castles"
 author="Seprix"
 description="A bot that goes for Castles.">
 <type name="TwoPlayer"/>
 <type name="Province"/>
 <type name="UserCreated"/>
 <type name="Engine"/>
 <type name="Combo"/>
 <type name="Bot"/>
 <type name="Fun"/>
   <buy name="Castles">
      <condition>
         <left type="countCardsInDeck" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Silver">
      <condition>
         <left type="countCardsInDeck" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Province">
      <condition>
         <left type="countCardsInSupply" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Duchy">
      <condition>
         <left type="countCardsInSupply" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Gold">
      <condition>
         <left type="countCardsInDeck" attribute="Castles"/>
         <operator type="smallerThan" />
         <right type="constant" attribute="3.0"/>
      </condition>
   </buy>
   <buy name="Castles"/>
   <buy name="Gold"/>
   <buy name="Estate">
      <condition>
         <left type="countCardsInSupply" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Silver"/>
</player>
« Last Edit: June 12, 2016, 04:55:36 pm by Seprix »
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

liopoil

  • Margrave
  • *****
  • Offline Offline
  • Posts: 2587
  • Respect: +2479
    • View Profile
Re: Seprix's Awful Simulations
« Reply #8 on: June 12, 2016, 04:58:34 pm »
0

That's very strange. Can you post a log from a sample game?

Perhaps it is having a very hard time hitting $10 for the king castle

Anyway I revised my post a bit. Maybe fiddle with how many silvers to get first. I can see anywhere from 0 to 5 being plausible.
« Last Edit: June 12, 2016, 05:00:28 pm by liopoil »
Logged

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #9 on: June 12, 2016, 05:00:42 pm »
0

That's very strange. Can you post a log from a sample game?

Perhaps it is having a very hard time hitting $10 for the king castle

I'll try to figure it out. It's buying Provinces (!?) so I probably messed something up. And now it's buying *no* Provinces. Hmm.

Any way to paste code into the simulator for use? That way, I could get some simulators other people made. I suspect I have to have a textfile with the code in it and load it. edit: Never mind. You simply click copy/paste on an empty slot and it takes whatever is in the clipboard.

DG's simulation is not much stronger than the one I wrote from before, so I think I can potentially write a better one given enough time. Hmm.

Also, nice. There is a suicide clause in the simulator, so it will not end the game while behind.

edit: I think I fixed your suggested sim, lio. It doesn't win that often though, so I'll keep looking into it.

edit 2: I found the problem. It continues to buy Castles because playing a card is not "in the deck" apparently?
« Last Edit: June 12, 2016, 05:16:55 pm by Seprix »
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #10 on: June 12, 2016, 05:21:22 pm »
0

Code: [Select]
<player name="Seprix's Castles Sim 1.0"
 author="Seprix"
 description="A bot that goes for Castles.">
 <type name="TwoPlayer"/>
 <type name="Optimized"/>
 <type name="Province"/>
 <type name="BigMoney"/>
 <type name="UserCreated"/>
 <type name="Engine"/>
 <type name="Combo"/>
 <type name="Bot"/>
 <type name="SingleCard"/>
 <type name="AppliesPPR"/>
 <type name="Competitive"/>
   <buy name="Silver">
      <condition>
         <left type="countCardsInDeck" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
      <condition>
         <left type="countCardsInDeck" attribute="Silver"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Castles">
      <condition>
         <left type="countCardsInDeck" attribute="Castles"/>
         <operator type="greaterThan" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Gold">
      <condition>
         <left type="countCardsInDeck" attribute="Gold"/>
         <operator type="smallerThan" />
         <right type="constant" attribute="3.0"/>
      </condition>
   </buy>
   <buy name="Castles">
      <condition>
         <left type="getTotalMoney"/>
         <operator type="greaterThan" />
         <right type="constant" attribute="19.0"/>
      </condition>
   </buy>
   <buy name="Province">
      <condition>
         <left type="countCardsInSupply" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Duchy">
      <condition>
         <left type="countCardsInSupply" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Gold"/>
   <buy name="Estate">
      <condition>
         <left type="countCardsInSupply" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Silver"/>
</player>

edit: I fixed the problem. I shouldn't have had the items = 0, but greater than 0 instead. Stupid me. :p
« Last Edit: June 12, 2016, 05:34:21 pm by Seprix »
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #11 on: June 12, 2016, 06:08:49 pm »
0

I wrote a simulator for Castles BM that beats DGs in fighting percentage against WW's Ultimate BM Strategy. I am still tweaking it, but it should also beat DG's sim in a Castles mirror (it's very close, but my sim always has the edge), if that's even relevant, which it isn't. :)

It is strictly better to not have the Humble Castle T1/2 when not playing the mirror, but better when playing the mirror.
« Last Edit: June 12, 2016, 06:12:58 pm by Seprix »
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

liopoil

  • Margrave
  • *****
  • Offline Offline
  • Posts: 2587
  • Respect: +2479
    • View Profile
Re: Seprix's Awful Simulations
« Reply #12 on: June 12, 2016, 06:13:58 pm »
0

The castles mirror is probably completely different. Can you try my revised version in my previous post?
Logged

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #13 on: June 12, 2016, 06:29:28 pm »
0

The castles mirror is probably completely different. Can you try my revised version in my previous post?


Tried it, fixed some mistakes I made.

edit: ....aaaaand it's terrible with a 37% win chance, but at least it buys Castles now.
« Last Edit: June 12, 2016, 06:33:17 pm by Seprix »
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

liopoil

  • Margrave
  • *****
  • Offline Offline
  • Posts: 2587
  • Respect: +2479
    • View Profile
Re: Seprix's Awful Simulations
« Reply #14 on: June 12, 2016, 06:31:45 pm »
0

The castles mirror is probably completely different. Can you try my revised version in my previous post?

Your strategy doesn't work currently, due to the fact that your strategy will never buy Castles until the Golds and Silvers run out. :)

You said Golds if Castles in Supply > 5, I think you meant < 5, so I'll try that.
It should buy castles over silvers after getting three silvers and castles over gold after getting three castles. Possibly you are using the wrong total money function.
Logged

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #15 on: June 12, 2016, 06:34:31 pm »
0

The castles mirror is probably completely different. Can you try my revised version in my previous post?

Your strategy doesn't work currently, due to the fact that your strategy will never buy Castles until the Golds and Silvers run out. :)

You said Golds if Castles in Supply > 5, I think you meant < 5, so I'll try that.

edit: ....aaaaand it's terrible, but at least it buys Castles.
Nope, I meant > 5. I think you implemented the silver wrong.

Here's what happened:

I implemented it right, then forgot to have it buy Castles. Then I switched the Golds symbol, and then I didn't get Castles buys, then I fixed it, and got it. Your sim has a win rate of 37% about, which is bad. Mine is currently about 68%.
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

DG

  • Governor
  • *****
  • Offline Offline
  • Posts: 4074
  • Respect: +2624
    • View Profile
Re: Seprix's Awful Simulations
« Reply #16 on: June 12, 2016, 06:37:21 pm »
+1

Since the castles are not individually listed in the simulator there is going to be a significant difference between real play and the simulation result. Although the simulator can give good results it is always worth knowing where the limitations kick in.
Logged

liopoil

  • Margrave
  • *****
  • Offline Offline
  • Posts: 2587
  • Respect: +2479
    • View Profile
Re: Seprix's Awful Simulations
« Reply #17 on: June 12, 2016, 06:43:40 pm »
0

The castles mirror is probably completely different. Can you try my revised version in my previous post?

Your strategy doesn't work currently, due to the fact that your strategy will never buy Castles until the Golds and Silvers run out. :)

You said Golds if Castles in Supply > 5, I think you meant < 5, so I'll try that.

edit: ....aaaaand it's terrible, but at least it buys Castles.
Nope, I meant > 5. I think you implemented the silver wrong.

Here's what happened:

I implemented it right, then forgot to have it buy Castles. Then I switched the Golds symbol, and then I didn't get Castles buys, then I fixed it, and got it. Your sim has a win rate of 37% about, which is bad. Mine is currently about 68%.
Okay, I guess I am still buying castles too soon. I adjusted the silver cutoff to 16, which moght make a big difference. When does your sim buy humble castle usually?


Hmm my sim is becoming more and more like DG's.
« Last Edit: June 12, 2016, 06:52:35 pm by liopoil »
Logged

liopoil

  • Margrave
  • *****
  • Offline Offline
  • Posts: 2587
  • Respect: +2479
    • View Profile
Re: Seprix's Awful Simulations
« Reply #18 on: June 12, 2016, 06:49:51 pm »
0

Since the castles are not individually listed in the simulator there is going to be a significant difference between real play and the simulation result. Although the simulator can give good results it is always worth knowing where the limitations kick in.
Unless you mean different play rules, you can give different buy rules depending on how many castles are in the supply to accomodate with the simulator.
Logged

DG

  • Governor
  • *****
  • Offline Offline
  • Posts: 4074
  • Respect: +2624
    • View Profile
Re: Seprix's Awful Simulations
« Reply #19 on: June 12, 2016, 06:55:09 pm »
0

Okay, I guess I am still buying castles too soon. I adjusted the silver cutoff to 16, which moght make a big difference. When does your sim buy humble castle usually?

Maybe you can't buy the humble castle too soon.
Logged

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #20 on: June 12, 2016, 07:07:44 pm »
0

Okay, I guess I am still buying castles too soon. I adjusted the silver cutoff to 16, which moght make a big difference. When does your sim buy humble castle usually?


Hmm my sim is becoming more and more like DG's.

Your sim is worse now. The problem with your Sim is that it buys Golds only after it greens most of the time and greens way too soon.

My sim starts greening when it has more than $19. I haven't adjusted numbers much yet. Here's the code:

Code: [Select]
<player name="Castles BM [Seprix]"
 author="Seprix"
 description="A bot that goes for Castles.">
 <type name="TwoPlayer"/>
 <type name="Optimized"/>
 <type name="Province"/>
 <type name="BigMoney"/>
 <type name="UserCreated"/>
 <type name="Engine"/>
 <type name="Combo"/>
 <type name="Bot"/>
 <type name="SingleCard"/>
 <type name="AppliesPPR"/>
 <type name="Competitive"/>
   <buy name="Gold">
      <condition>
         <left type="getTotalMoney"/>
         <operator type="smallerThan" />
         <right type="constant" attribute="19.0"/>
      </condition>
   </buy>
   <buy name="Silver">
      <condition>
         <left type="countCardsInDeck" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
      <condition>
         <left type="countCardsInDeck" attribute="Silver"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Silver">
      <condition>
         <left type="getTotalMoney"/>
         <operator type="smallerThan" />
         <right type="constant" attribute="16.0"/>
      </condition>
   </buy>
   <buy name="Castles"/>
   <buy name="Province">
      <condition>
         <left type="countCardsInSupply" attribute="Castles"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Province">
      <condition>
         <left type="gainsNeededToEndGame"/>
         <operator type="smallerOrEqualThan" />
         <right type="constant" attribute="10.0"/>
      </condition>
   </buy>
   <buy name="Duchy">
      <condition>
         <left type="gainsNeededToEndGame"/>
         <operator type="smallerOrEqualThan" />
         <right type="constant" attribute="4.0"/>
      </condition>
   </buy>
   <buy name="Estate">
      <condition>
         <left type="gainsNeededToEndGame"/>
         <operator type="smallerOrEqualThan" />
         <right type="constant" attribute="2.0"/>
      </condition>
   </buy>
   <buy name="Gold"/>
   <buy name="Silver"/>
</player>

I'm trying to figure out why my win percentage is so high (73%!!!) when I have one little buy condition in, even when I inflate the number incredibly high, but not when I remove it.
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #21 on: June 12, 2016, 07:14:46 pm »
0

Okay, I think I understand it now. My bot is dumb, so it won't buy Provinces until all of the Castles are gone, even if it can hit above 8 when I remove that particular buy condition. I programed it before so that it did buy Provinces after it couldn't buy Castles, and this makes it win against UBM a lot more often in the particular case of this bot, but I think it's a losing play against other bots.
« Last Edit: June 12, 2016, 07:16:08 pm by Seprix »
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #22 on: June 12, 2016, 07:36:02 pm »
0

Another major breakthrough. I'm going to be the premier Castles player at this point! :)

I pitted my former Castles BM against BM Smithy, where it barely won the coin flip. Now I can get Castles BM to beat Smithy BM 64% of the time. In comparison, DG's sim can beat Smithy BM 34% of the time.
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)

liopoil

  • Margrave
  • *****
  • Offline Offline
  • Posts: 2587
  • Respect: +2479
    • View Profile
Re: Seprix's Awful Simulations
« Reply #23 on: June 12, 2016, 07:39:45 pm »
0

How could it be worse? I just made it wait longer to green. Oh well.
Logged

Seprix

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5607
  • Respect: +3676
    • View Profile
Re: Seprix's Awful Simulations
« Reply #24 on: June 12, 2016, 07:52:05 pm »
0

I just discovered something crucial about Castles BM.

I can get DG's sim to win 57% against Smithy BM.

When I apply this crucial discovery to my simulator against Smithy BM, I get about 64%, but more consistent. I think my Bot is a little 'smarter' when it comes to purchasing things, so I'm pretty sure that is the main 'difference' between my sim and his. I think the winning % goes higher because the bot doesn't play optimally, and that my bot plays closer to optimally on average than DGs, but that is not my crucial discovery. It's quite obvious, actually. I think anybody who has played with Castles could have told you what I discovered.
Logged
DM me for ideas on a new article, either here or on Discord (I check Discord way more often)
Pages: [1] 2  All
 

Page created in 1.371 seconds with 21 queries.