(Turns out I only got the motivation to write this when the server stopped working. Of course.)
This article should be considered a supplement to
http://dominionstrategy.com/2011/06/24/guest-article-deck-control/---
What do you track?The correct response for a new player should be, "What?".
The response from an okay player should be, "The amount of VP in both players decks", to at least think about PPR.
The response from an absolutely perfect player should be "Everything derivable from a log of all cards played and bought." Does anyone actually do this? I'd wager, no. However, a strong player should get pretty close, and utilize that information to make better tactical plays to get small advantages that snowball up. Personally, I consider myself to be strong in understanding and getting information, and pretty terrible in actually using it properly.
Things worth tracking:The number of each Victory card in the game in both players decks. This is better than just tracking the amount of VP, because it accounts for Silk Roads and gives you good intuition about whose deck is more clogged, which can influence whether you build more or go for points.
The number of each action card in the game in both players decks. Sometimes, knowing the exact counts is overkill, but knowing how many +Buy, +Cards, and +Actions there are can be a big help.
In games with Cursers/Ruins, how many of the 10 junk cards you have. Again, often not immediately applicable, but useful as a heuristic for whose deck is more clogged.
Alone, the above don't help very much. However, in combination with tracking how many are still left in the current shuffle, you can make vastly better plays. Knowing that you're due for $8 because you've seen mostly junk can help in PPR situations. Knowing that you have seen most of your good treasures can influence you in the other direction. This applies to tracking your opponent as well; if they've blown through their Gold and Silvers, you can be much more bold with PPR than you normally would.
However, this deck tracking helps the most with playing engines. Just for your own deck, you should be thinking about
- Number of terminals your deck can support. This should affect your buy decisions and overall strategy.
- Maximum money your deck can produce, if you draw your deck. In the late game, you should be thinking about how you get this to $16 or $22, or whatever important cost point you want.
- Number of Buys/cards you can gain in one turn. Important to keep in mind for 3-pile considerations.
- In games where you can't draw your deck, a rough estimate of how strong the cycling in your deck is. If your cycling is strong, you can trigger reshuffles with much less fear. In general, it also lets you make better reshuffle decisions: if you know you've played most actions, you can decide based on number of non-action cards in your discard. If you still have actions, you can trigger the reshuffle with a good chance of continuing to go off.
Again, tracking this for your opponent helps you immensely more. For example, if your opponent has played all +Buy cards this shuffle, and aren't going to reshuffle soon, you can empty victory piles much lower than you normally would.
In addition to this, you should be keeping information on the piles. Which ones are running low and are potential 3 pile targets, which piles are important to win the split on. Usually +Actions, more rarely +Cards because you don't usually have the +Actions or time to empty those piles.
This, in addition to actually making the right buys, and managing your reshuffles correctly, is what makes engines so hard to play correctly. There are many more decisions to make than it may seem, and continually making the right ones can make top players get wins for seemingly weird reasons.
Sample Gameshttp://gokologs.drunkensailor.org/static/logprettifier.html?20140523/log.53784910e4b0557728cb855c.1400903624556.txt - on turn 12, I buy an Alchemist instead of a Familiar. At that point, there are 4 Curses left, and we each have 2 Familiars, so if I had bought a Familiar, it would be a very slim chance I could play all 3 before my opponent only plays 1.
http://gokologs.drunkensailor.org/static/logprettifier.html?20140523/log.505c6645a2e6c78ad2ed5ad3.1400900241914.txt - on turn 14, I buy 2 Envoys, both because I knew my deck could support more, and because I hadn't played many of my Festivals or Highways that shuffle yet, so I knew I was due for a megaturn and wanted to set up a 3-pile.
http://gokologs.drunkensailor.org/static/logprettifier.html?20140518/log.511bd320e4b01008292d004b.1400403187517.txt - on turn 14, I draw my deck and hit $21. I buy Province-Province-Mystic-Copper, giving me a total of $24 in my deck and setting up a triple Province. After this turn, there are 5 Provinces left, meaning if I hit the triple Province, I'll have 5 out of the 8. This pays off when I manage to draw my entire deck the next turn for the win.
http://gokologs.drunkensailor.org/static/logprettifier.html?20140518/log.5101a6c4e4b02b7235c3860f.1400471202741.txt - on turn 13, I know my Forge is in one of the next 5 cards. Although I have the option of Throne Room-Envoy, I instead Throne Room Mining Village, to maximize my chances of drawing the Forge instead of losing it to Envoy. I still play this turn a bit incorrectly: if I had known I had a possible TR, I should have played Mining Village first instead of Treasury, because Mining Village is much worse to Throne Room at this point.
(Edit: There was another game where I left one Province in a game with a small VP lead but lots of +Buy. I knew my opponent had played all sources of +Buy and didn't have enough cycling to trigger a reshuffle on his turn, so I could go for the win over the next 2 turns. Unfortunately, I can't find this game anymore, so you'll have to live with the summary.)
Concluding RemarksThe unfortunate truth is that a lot of this isn't generalizable. Of course, everything depends on the kingdom, but when you start talking about game state, any hope of generalization flies out the window. The best advice I can give is to get in a habit of considering the information at your disposal whenever you have to make a decision.
- Obtain as much information as possible, on both your deck and your opponent's
- Synthesize that information into useful insights
- Use those insights to make better plays
Above all else, think about your turn. Autopiloting is easy; it's also one of the worst ways to play.
(Edit: Did some revisions, fixed some typos.)