Dominion Strategy Forum

Please login or register.

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

Author Topic: Fabian's Game Analysis Series, Game #1  (Read 6295 times)

0 Members and 1 Guest are viewing this topic.

Fabian

  • 2012 Swedish Champion
  • *
  • Offline Offline
  • Posts: 666
  • Respect: +542
    • View Profile
Fabian's Game Analysis Series, Game #1
« on: July 22, 2012, 09:29:32 pm »
+3

(this is a series where I take a real game I've played against a real opponent, and analyze the board in as much detail as I find necessary, trying to figure out how it should be played. I've not selected specifically interesting boards, I will go with whatever game happens whenever I decide to do a game like this. It is mostly an exercise for me to become a stronger player, but hopefully we can all learn something through discussion and shared analysis.)

Game #1
Feast, Jester, Margrave, Nobles, Remodel, Thief, Trading Post, Tribute, Walled Village, Witch, !Colony

4/3 opening, going first.


The key card early is Witch. You want to play it early, and often.

Option #1 is to open Feast/Silver, looking to get two early Witches and play them often with Walled Villages. This would transition into more of an engine, with Silver at 3, Walled Village at 4, Nobles at 6, and at 5 you'd likely want to pick up a Trading Post quickly, and possibly a Margrave in the midgame (although the +buy isn't that likely to be important on this board; as the cards you want are not cheap, it's not very likely you'll be able to start picking up stuff like Walled Village+Nobles with $10, and at that point you might prefer Province anyway if you can draw most of your deck reliably enough to get there). It's possible Tribute could be a nice card to pick up at $5 too in the midgame, and Remodel could be reasonable in the mid/endgame to turn Nobles into Provinces to accelerate the game ending.

Option #2 is to go for a deck focusing less on enginey things and go more for a money approach. Silver/Silver or Feast/Silver opening, looking to buy two early Witch, Silver on 3/4, Gold over Nobles at 6, and $5 cards to pick up would probably be.. Margrave? In a deck like this with less reliable card draw, Trading Post shouldn't be very impressive in the midgame since you can't reasonably rely on it to get rid of the cards you really want to get rid of (Curses primarily). Even buying Margrave might be bad as that would add a third/fourth terminal draw action; then again that might not be the worst as you're probably adding around 5 Curses to your deck, so clashing is less of an issue than normal.

After writing this out, I played a test game between the two strategies, and as I predicted option #1 crushed pretty hard. I would expect it to win very comfortably over the more money-centric approach if the two kept playing for a long time:

http://dominion.isotropic.org/gamelog/201207/22/game-20120722-181313-818d312e.html

Trading Post seemed key here, to the point where I think the Money deck would want to add one early too. As mentioned above, it would run into problems of not being able to draw it consistently, or drawing it dead, too often, and at that point I believe the deck would want to morph into something like deck #1 anyway.



Some questions to get discussion started:

What do you think the optimal strategy on this board looks like? Does it involve Nobles over Gold? Is it close?

Knowing your opponent would go for this optimal strategy, what adjustments, if any, would you make to your own strategy to gain an edge? Does Jester come into play here at all?

Feast/Silver, Silver/Silver, or other for the opening? Why? Is it close?

What parts of my analysis do you disagree with? Why? What key things have I missed in my analysis?


General discussion is also very welcome!
« Last Edit: July 22, 2012, 09:32:10 pm by Fabian »
Logged

cayvie

  • Explorer
  • *****
  • Offline Offline
  • Posts: 317
  • old
  • Respect: +236
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #1 on: July 23, 2012, 04:04:00 am »
0

I would think, in a matchup between these strategies, the money player would want Tributes.
Logged
18:28 MEASURE YOUR LIFE IN LOVE: you shouldve done the decent thing and resign rather than go on being that lucky all the time

she/her

dondon151

  • 2012 US Champion
  • *
  • Offline Offline
  • Posts: 2522
  • Respect: +1856
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #2 on: July 23, 2012, 04:15:24 am »
0

I generally find that it's helpful to try to recover from cursing attacks if there is a card in the kingdom that deals with Curses well, and Trading Post does it exceptionally.
Logged

DG

  • Governor
  • *****
  • Offline Offline
  • Posts: 4074
  • Respect: +2624
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #3 on: July 23, 2012, 09:31:01 am »
+1

I think there should be some caution against the trading post here as it matches up badly against a margrave. It would really need to be used in a strong drawing engine. There's also an opportunity to use the walled village for its benefit instead of as an expensive village. This would suit a deck with one village and a couple of terminals that could be drawing cards, perhaps a couple of witches.
Logged

RisingJaguar

  • Minion
  • *****
  • Offline Offline
  • Posts: 527
  • Respect: +184
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #4 on: July 23, 2012, 11:51:40 am »
0

The analysis is really good here.  I just want to caution using this game as a guideline of how this match up would go.  Mainly, the curse war was 7-3.  BM player misses his early witch play on T6/T7, then after trying to recoup from this, his witches collide on T11 (see T10 hand). 

To be honest, I was pretty surprised to see the engine actually succeed because the engine parts are so expensive and the flood of silvers cannot help.  Although I guess you don't need much when you have that many silvers.

The BM player (or non-noble player) should be able to react with tributes and can daringly use them as villages against the engine player. 

Jester looks like a fun card here, but probably incredibly ineffective.  The early part of the game is filled with trashing/cursing, both price points jester collides with.  Then after that is settled, trying to steal engine pieces will be clogged with curses that won't give out anything. 

If the engine player wanted to ensure their engine was incredibly effective, I would suggest a remodel the same time you would purchase a margrave (so after the curses are given away and moving into greening).  It would be to remodel silvers into engine pieces.  This would be only if the deck had a decent engine and looking for more consistency really.  Although at this point, the engine player may be already ahead and the marginal benefit wouldn't even be that large here.
Logged

WanderingWinder

  • Adventurer
  • ******
  • Offline Offline
  • Posts: 5275
  • ...doesn't really matter to me
  • Respect: +4386
    • View Profile
    • WanderingWinder YouTube Page
Re: Fabian's Game Analysis Series, Game #1
« Reply #5 on: July 23, 2012, 12:10:54 pm »
+1

The game you link is extremely good luck for the engine player. I also don't know why Mr. BM has so darned many terminal draw cards.
I still like BM-ish type stuff here - I just don't see the engine getting up to snuff in time.

I am really not so sold on tribute here though - drawing cards is not so great, with the curses; getting actions doesn't help a BM player much if at all; and +cash is good,but not SO reliable. Yeah, there's nobles, but the actions aren't so good, and there's also decent chances of hitting curse. I just don't like it as a cross-matchup counter here much at all.

BM player should also be conscious of racing out duchies and three-piling, though that's not so so great an option against an engine, most of the time.

methods of rationality

  • Bishop
  • ****
  • Offline Offline
  • Posts: 102
  • Respect: +13
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #6 on: July 23, 2012, 01:12:22 pm »
0

First of all I don't like feast for the money player. Since he's not trying to play 2 witches at once he really only needs 1 witch and double silver could get you the witch easily. Secondly, I like trading post. It will help him get rid of curses and estates and if he buys it over margrave he will have less of an issue about drawing it dead. Besides the attack of margrave is not as power full with the curses you give to the other player. I would open silver silver, buy a witch quickly and then a trading post (probably preferring it to gold). Then play mostly money and maybe green a bit earlier than usual if I manage to flood my deck with enough silver. I probably wouldn't buy any more actions except nobles of course when I'm greening, unless I get 5 again which I could spend on another witch if there are at least 5 or so curses left in the supply or, if not, a margrave. But then again I'm probably too much of a big money guy.
« Last Edit: July 23, 2012, 01:14:36 pm by methods of rationality »
Logged

HiveMindEmulator

  • Mountebank
  • *****
  • Offline Offline
  • Posts: 2222
  • Respect: +2118
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #7 on: July 23, 2012, 05:01:03 pm »
+1

I think engine should almost certainly be best here. I tend to always like engine when there is Nobles+village+attack+buy. The attacks make BM slow enough that you can grab all your Nobles and then end up drawing most of your deck every turn and reliably buying (multiple) VP cards before the BM player can end the game. Attacks don't hurt Nobles decks that much since you're in no rush to get to $8, and you eventually end up drawing tons of cards. I'd say my strategy is buy 1 Witch, then TP, then Margrave, taking WVs at 4 and Silvers at 3. Then start buying Nobles with $6+ and Walled Villages with less, getting 1 more Margrave the next time I $5 exactly. Then just start on the other green cards when the Nobles are gone.
Logged

dondon151

  • 2012 US Champion
  • *
  • Offline Offline
  • Posts: 2522
  • Respect: +1856
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #8 on: July 23, 2012, 06:06:52 pm »
0

I still like BM-ish type stuff here - I just don't see the engine getting up to snuff in time.

Engine should win with Nobles in the kingdom.
Logged

Eevee

  • Saboteur
  • *****
  • Offline Offline
  • Posts: 1010
  • Shuffle iT Username: Eevee
  • A wild Eevee appears!
  • Respect: +867
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #9 on: July 23, 2012, 06:28:01 pm »
0

Someone interested in playing this one out a couple of times? I think the engine should be better too but sadly I dont build them too well so I might not be the guy for the job.
Logged

Fabian

  • 2012 Swedish Champion
  • *
  • Offline Offline
  • Posts: 666
  • Respect: +542
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #10 on: July 23, 2012, 06:35:48 pm »
0

I would think, in a matchup between these strategies, the money player would want Tributes.

I would think the opposite. Those tributes are going to hit an awful lot of Walled Villages and Curses. It would get better in the mid/endgame when the engine deck starts adding more Nobles and remove Curses, but I'm not sure that's better than the alternatives (though I'm not sure what that alternative is exactly; WW seems to think the Margraves are a mistake for the BM deck?). RJ seems to agree with you (then again WW doesn't), it really seems like more of an mid/endgame pickup to me though.

First of all I don't like feast for the money player. Since he's not trying to play 2 witches at once he really only needs 1 witch and double silver could get you the witch easily. Secondly, I like trading post. It will help him get rid of curses and estates and if he buys it over margrave he will have less of an issue about drawing it dead. Besides the attack of margrave is not as power full with the curses you give to the other player. I would open silver silver, buy a witch quickly and then a trading post (probably preferring it to gold). Then play mostly money and maybe green a bit earlier than usual if I manage to flood my deck with enough silver. I probably wouldn't buy any more actions except nobles of course when I'm greening, unless I get 5 again which I could spend on another witch if there are at least 5 or so curses left in the supply or, if not, a margrave. But then again I'm probably too much of a big money guy.

Both you and HiveMind are saying only pick up 1 Witch (you for the BM deck, HiveMind for the engine deck), do others agree with this? Is it correct no matter how many Witches the opponent picks up? As for Feast vs Silver, I'm curious what others think. If we take the BM deck and consider what happens on turns 3/4, it seems likely that a Silver/Feast opening will allow a gaining of Witch and buying Silver on turns 3 and 4, which would mean the same end result as opening Silver/Silver and buying Witch/Silver on turns 3 and 4 (deck of 7 Copper 3 Estate 3 Silver 1 Witch). I'm sure some analysis has been done on Feast and what happens in various circumstances (Feast being card 11 or 12, the Silver you would have bought instead being card 11 or 12, etc), but I'm not smart enough to figure it out off the top of my head. Intuitively it would seem opening Feast/Silver makes it more likely you'll be able to buy 2 Witches before the second reshuffle, but also more likely you might miss buying Witch entirely (though that chance has to be smaller, no?)

Someone interested in playing this one out a couple of times? I think the engine should be better too but sadly I dont build them too well so I might not be the guy for the job.

Meet me in the Secret Chamber! (Eevee or anyone else)
« Last Edit: July 23, 2012, 06:39:40 pm by Fabian »
Logged

Fabian

  • 2012 Swedish Champion
  • *
  • Offline Offline
  • Posts: 666
  • Respect: +542
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #11 on: July 23, 2012, 07:17:20 pm »
0

I'll see if I can find someone to play some test games with me like Eevee suggested, here's what I'm thinking as far as buy rules go:

Engine deck:

Open Feast/Silver
Buy Walled Villages for $4 or less
Get a Witch
Get a Witch
Get a Trading Post
Get Nobles at $6
Get a Margrave at $5 (bad?)
From here it's probably just go with the flow, seems silly to set up super rigorous buy rules. Let the human play it out.

The big questions for me are if we want Witch #2, and where Trading Post falls on the priority list. Do we prefer it to Nobles at $6 in the early/midgame? Do we want to get Witch #1, Trading Post, Witch #2? No Trading Post?


----

philosophyguy turned up and we got some games in:

http://dominion.isotropic.org/gamelog/201207/23/game-20120723-155531-0123698b.html
http://dominion.isotropic.org/gamelog/201207/23/game-20120723-160454-8af99780.html
http://dominion.isotropic.org/gamelog/201207/23/game-20120723-161312-ccb40666.html

Overall it seemed like the engine deck handled itself very well, and none of the games felt very close in the end. We didn't branch out with possible Tribute/Jester/Remodel shenanigans, preferring to keep the decks more streamlined. Turn 13 in game 1 I felt like Tribute/Walled Village/Jester were all sort of reasonable, don't know what I should be picking there.

Edit: And here's the original game log from yesterday, which started this thread:

http://councilroom.com/game?game_id=game-20120722-175312-89ceb29d.html
Logged

HiveMindEmulator

  • Mountebank
  • *****
  • Offline Offline
  • Posts: 2222
  • Respect: +2118
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #12 on: July 23, 2012, 07:45:44 pm »
0

The big questions for me are
1. if we want Witch #2, and
2. where Trading Post falls on the priority list.
1. I'm pretty sure we don't want Witch #2. The way I see Witches working in general is that the first Witch causes a major slowdown, and the second one has a much more minor slowing effect, and seems primarily useful for scoring purposes. Here we don't care about the scoring, since we're going to trash the Curses anyway, and the slowdown can be achieved more effectively by the Margrave, which provides more all-around benefits (better draw, +buy). Margrave just has to be better than Witch #2.
2. I think TP should be had on the second $5+ hand. You want to get that going early. It's the only thing building your economy. You want to be buying Walled Villages and not Silvers.

The question I want to add is:
3. When do you want to get Remodel?

You probably want one at some point, and I'm inclined to believe you want it as an opening. You run the risk of delaying your Witch or hitting the ultimate disaster of getting all Coppers with Remodel and Estates with Silver, but I think more often than not you end up ahead. It's probably okay to get a later Witch if you start removing Estates and stocking up on $4 cards (Remodel, WVs) sooner.
Logged

dondon151

  • 2012 US Champion
  • *
  • Offline Offline
  • Posts: 2522
  • Respect: +1856
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #13 on: July 23, 2012, 08:06:56 pm »
0

1. I'm pretty sure we don't want Witch #2. The way I see Witches working in general is that the first Witch causes a major slowdown, and the second one has a much more minor slowing effect, and seems primarily useful for scoring purposes.

I don't feel the same way about this. If you get 1 Witch to your opponent's 2, then you're almost certainly going to lose the Curse split. You'll have trashed some of the Curses with your Trading Post, but your opponent has fewer Curses as well. That said, I don't think that one choice is particularly stronger than the other.

I also don't agree about opening with Remodel. If you fail to hit $5 on turns 3-4 (about a 50% chance, technically lower if you count a Remodel-Silver collision as "hitting $5"), then you're prone to just not hitting $5 very often for a long time, and that's really bad. Your Estates are going to get TP'd away for the most part anyway, so I don't really see the benefit of getting a couple of extra WVs out of it. I'm guessing that you'd rather pick it up later so that you can do Witch -> Gold/Nobles or Silver -> Margrave or something.
Logged

HiveMindEmulator

  • Mountebank
  • *****
  • Offline Offline
  • Posts: 2222
  • Respect: +2118
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #14 on: July 23, 2012, 09:20:34 pm »
0

1. I'm pretty sure we don't want Witch #2. The way I see Witches working in general is that the first Witch causes a major slowdown, and the second one has a much more minor slowing effect, and seems primarily useful for scoring purposes.

I don't feel the same way about this. If you get 1 Witch to your opponent's 2, then you're almost certainly going to lose the Curse split. You'll have trashed some of the Curses with your Trading Post, but your opponent has fewer Curses as well. That said, I don't think that one choice is particularly stronger than the other.
With a second Witch, you're talking about adding 2 Curses, in the 4th and 5th shuffles or something. This does do something, but is it that much stronger than a Margrave attack that it's worth delaying/omitting Margrave? I doubt it. 6-4 Curse split and 5-5 Curse split are not all that different in engines with trashing. Curse split is important in BM games, but not really important here, imo.

Quote
I also don't agree about opening with Remodel. If you fail to hit $5 on turns 3-4 (about a 50% chance, technically lower if you count a Remodel-Silver collision as "hitting $5"), then you're prone to just not hitting $5 very often for a long time, and that's really bad. Your Estates are going to get TP'd away for the most part anyway, so I don't really see the benefit of getting a couple of extra WVs out of it. I'm guessing that you'd rather pick it up later so that you can do Witch -> Gold/Nobles or Silver -> Margrave or something.
You're not actually that prone to not hitting $5 for a long time, since you're removing an Estate, which is the primary obstacle to hitting $5. The chances of not getting a Witch before the third shuffle should be pretty small. You are definitely risking taking a tempo hit though, so it might not be worth it. It's hard to say. I'm still pretty sure you want one eventually, so if you don't open with it, you still have to find a timing to get it. And when is that? Before any Walled Villages? After 1? 2? Once you're drawing most of your deck?
Logged

verikt

  • Apprentice
  • *****
  • Offline Offline
  • Posts: 272
  • Respect: +65
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #15 on: July 23, 2012, 11:19:21 pm »
0

I'm not all that sure that witch is key. Seems to me that margrave would slow down your deck, more than a curse would the other player. Worst case buy another trading post and ignore witch altogether. Not to mention that with trading post to trash two cards, a jester to give you coppers is not much worse than witch either. I'd rather go silver silver trading post margrave, and then either jester or tribute to counter your nobles. And I mean as an engine deck.
Logged

Fabian

  • 2012 Swedish Champion
  • *
  • Offline Offline
  • Posts: 666
  • Respect: +542
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #16 on: July 25, 2012, 03:07:02 am »
+1

I created a bot, attempting to roughly emulate the main strategy we've been discussing. I then made some small changes to it, running the new version(s) against the baseline bot, to see how the result would change. It's not an attempt at a perfect optimized bot, partly because I'm not good enough with the simulator to figure out the necessary buy rules, and partly because it would be more work than I'm willing to put in. The baseline bot is just the very basics; it has no Duchy buying rules and no Estate buying rules. Here's what it does:

Opens Feast/Silver
Gets 2 Witch
Gets 1 Trading Post
Gets Nobles
Gets 1 Margrave at 5
Gets Walled Village at 4
Gets Silver at 3
Starts buying Province when 5 Nobles are left (in a non-mirror match, as in "engine" vs "money", I suspect this number would be a lot lower. In the mirror match, 5 seemed to beat 6 slightly and be equal against 4, so I made this the baseline)

Here's the baseline bot:

Code: [Select]
<player name="Fabian Game #1 Baseline"
 author="Fabian"
 description="A very basic bot attempting to emulate the key parts of the strategies discussed in this thread: http://forum.dominionstrategy.com/index.php?topic=3560.0">
 <type name="Province"/>
 <type name="UserCreated"/>
 <type name="TwoPlayer"/>
 <type name="Bot"/>
   <buy name="Province">
      <condition>
         <left type="countCardsInSupply" attribute="Nobles"/>
         <operator type="smallerThan" />
         <right type="constant" attribute="6.0"/>
      </condition>
   </buy>
   <buy name="Witch">
      <condition>
         <left type="countCardsInDeck" attribute="Witch"/>
         <operator type="smallerThan" />
         <right type="constant" attribute="2.0"/>
      </condition>
   </buy>
   <buy name="Trading_Post">
      <condition>
         <left type="countCardsInDeck" attribute="Trading_Post"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Nobles"/>
   <buy name="Margrave">
      <condition>
         <left type="countCardsInDeck" attribute="Margrave"/>
         <operator type="equalTo" />
         <right type="constant" attribute="0.0"/>
      </condition>
   </buy>
   <buy name="Feast">
      <condition>
         <left type="countTurns"/>
         <operator type="smallerThan" />
         <right type="constant" attribute="3.0"/>
      </condition>
   </buy>
   <buy name="Walled_Village"/>
   <buy name="Silver"/>
</player>

Here are some tests I did. I wasn't too concerned with exact winning %, so these are Accurate Simulations unless otherwise noted. The more interesting thing to me was in which direction the resulting matchup went, not by how much exactly:

I bought only 1 Witch instead of 2. Baseline lost ~44.5 to ~53.5
I skipped buying Margrave. Baseline won ~56 to ~42.5
I opened Silver/Silver. Baseline won ~58 to ~40.5 (!)
I skipped buying Trading Post. Baseline won ~65.5 to ~32
I bought a Gold before the first Nobles. Baseline won ~51.5 to ~46.5
I bought a Gold before Trading Post. Baseline won ~53.5 to ~44.5
I bought 1 Witch, then 1 TP, then 1 Witch instead of Witch, Witch, TP. Baseline lost ~46 to ~52 (Ultimate Simulation)
I bought TP, then 2 Witch. Baseline lost ~45.5 to ~52.5 (Ultimate Simulation)
I bought Margrave ahead of the Witches/TP. Baseline won ~63.5 to ~35
I bought Margrave ahead of TP. Baseline won ~56 to ~42.5
I bought 0 Witch. Baseline won ~65 to ~33
I bought Margrave over Nobles. Baseline won 50.1 to 48.24 (Ultimate Simulation)
After the deck contains x Silver, I stopped buying Silver. For low x, this made the baseline bot win. For higher x, it seemed to make no difference.

The results of these simulations led me to believe the following things are likely to be true:

1. You only want 1 Witch.
2. You don't want to buy Gold in this deck.
3. Trading Post is hugely important.
4. Margrave is hugely important, but not a priority like TP and Witch. It's a good pickup when you already have them and miss $6 for Nobles.
5. You most definitely want to open Feast/Silver over Silver/Silver.

After this, I made a new baseline bot. This bot is identical to the first baseline bot, except it buys 1 Witch instead of 2. The previous baseline mirror lasted 21.7 turns (Ultimate Simulation), the baseline2.0 mirror lasted a full 2½ turns less, 19.2 turns (Ultimate Simulation) , which I found noteworthy. Then some more simulations:

I bought TP before Witch. Baseline lost ~48 to ~50.5
I skipped buying Witch. Baseline won ~63 to ~35.5
I bought Margrave over TP and Nobles. Baseline won ~55 to ~43.5
I bought Margrave over Witch, TP and Nobles. Baseline won ~57.5 to ~41
I bought 1 Gold over Witch, TP, Nobles, Margrave. Baseline won ~59 to ~39
I bought 1 Gold over TP, Nobles, Margrave. Baseline won ~59 to ~39.5
I bought 1 Gold over Nobles, Margrave. Baseline won ~57 to ~41.5
I opened Silver/Silver. Baseline won ~58.5 to ~40

These simulations led me to believe it's very possible you want Trading Post before you want Witch, but the difference should be small. It also made me believe you definitely don't want to skip Witch.

I'd be curious to hear thoughts on the results of these simulations, and on whether or not the conclusions I draw can be trusted despite the bots' incomplete play rules (most notably not bothering with end game conditions with Duchy and Estate). My feeling is they still can be trusted.

I didn't bother trying to create an Engine vs Money match-up, as properly evaluating a non-mirror match would probably require complete bots with good end-game play, which I didn't want to bother with. I'd be very interested if someone wanted to give it a shot though.


Edit to add: I tried baseline2.0 against baseline2.0 + one Tribute, both after Margrave in the buying order, and in place of Margrave. I did the same with 1 Jester. In all 4 cases, the new bot lost somewhere between convincingly and horribly. I don't believe those cards would improve the mirror match.
« Last Edit: July 25, 2012, 03:19:19 am by Fabian »
Logged

HiveMindEmulator

  • Mountebank
  • *****
  • Offline Offline
  • Posts: 2222
  • Respect: +2118
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #17 on: July 25, 2012, 03:36:48 am »
0

One thing you didn't try was adding more than one Margrave. That should improve things.

I also think Remodel wants to be involved, but simulator can't handle that.

And I think you're right about TP first, though I didn't think of it at first Witch adds one junk card to their deck, TP removes 2 from yours AND gives you a Silver. I think generally when you want one of each, TP should be first. I think this has been discussed before, but none of us remember to apply it...
Logged

Fabian

  • 2012 Swedish Champion
  • *
  • Offline Offline
  • Posts: 666
  • Respect: +542
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #18 on: July 25, 2012, 03:48:48 am »
0

That's true HiveMind. Baseline3.0 (TP over Witch compared to 2.0) against 3.0 with an extra Margrave goes 49.181% vs 49.285% on an Ultimate Simulation. With two extra Margrave (3 total), it's 50.58% to 47.854% in favor of the baseline deck. From this I would conclude that 2 Margrave makes no difference (which is certainly better than Jester or Tribute, which makes the deck lose terribly), and 3 Margrave is slightly worse.

Edit: When you buy 2 Margrave, it seems you want to buy the first Margrave over the first Nobles actually. Beats the baseline bot ~51.5 to ~47. Buying 2 Margrave over the first Nobles is essentially the same. Overall it seems you just want to buy Margrave over Nobles, period, which I guess I missed in my initital simulations? Well, I tried it (and lost) when buying 2 Witch > TP, but not with the 1 Witch bot. Oh well, interesting.
« Last Edit: July 25, 2012, 04:07:35 am by Fabian »
Logged

DG

  • Governor
  • *****
  • Offline Offline
  • Posts: 4074
  • Respect: +2624
    • View Profile
Re: Fabian's Game Analysis Series, Game #1
« Reply #19 on: July 27, 2012, 10:30:18 am »
+1

Upon reflection, this is a very good example kingdom for nobles

- by putting vp into the kingdom the nobles give more time for deck development
- nobles are added into an existing drawing engine
- either +2 actions or +3 cards might be useful each turn
- silver from the trading post provides coins from treasures, so nobles actions are not needed to support +coin actions
- gold is not required since the trading post provides silver
Logged
Pages: [1]
 

Page created in 0.063 seconds with 20 queries.