Practical Advice: Playing From Behind

Game Log

 

...with Shelters. Witch into Rebuild seems pretty clear here.

My opponent gets a 5/2, opens Witch/Estate.

I have 3/4 and get Silver/Feast. Tuns 3 and 4, my Feast becomes a Witch, and I get another feast and a Silver. Going into turn 6, my Feast and Witch are colliding in my hand, and the decks look like this:

At this point, I feel very far behind. I hit $4, and I make what I feel is a pretty bold choice: (with hovel in hand) I buy a Silk Road. What?! Well, tactically, I'm not sure the time is right. But the idea, the strategic thought, is something I am very much on board with.

The concept here is that, though I am very far behind in any case, I'm especially down in the long game. So I make the decision to try to turn the game as short as possible, such that there is more randomness and luck (or less time for the overall quality of his deck to shine through). This should be really losing, but on the other hand, I think everything is really losing, this gives me better random chances, and most importantly, it gives my opponent the chance to make mistakes.

The first of these comes on turn 13, when he buys a Duchy. This is a classic trap in 2 ways: "I'm playing Rebuild, better get Duchies pretty soon" (which is often not true in non-mirrors, Colony games, and various odd scenarios), and "my opponent is playing to end the game quickly, better get points soon in case that happens" (you're often better served by not helping them end things). At that point in the game, he had 3 Rebuild, 1 Estate, 1 Province, and 1 Duchy. If he just focuses on getting those cards up to colony, he's going to have a near-insurmountable lead. He can then finish off by grabbing 1 more green card LATE. As is, the Duchy both slows him down, and it helps me get the piles empty to end the game sooner. This should have been a Rebuild. A similar situation occurs on turn 15, except at that point, he's also locking in one of his own Estates by emptying the LAST Duchy.

He follows this up by grabbing Silvers, in a game where he has very little chance of actually making $8 (though to be fair, actually buying Province would essentially seal the game). By his turn 19, through some luck as well as my mad greening and his missteps, he is actually down by one point, with $6 in hand. He makes a final mistake by buying the last Silk Road, which is only worth 1 point to him. I have one of my Rebuilds, Duchy->Province, and buy the final curse to win.


Let's dig a little as to why this Silk Road is a mistake. There is first of all the general principle of him wanting the game to go long, and this helping me close it out. But specifically here, if I have Rebuild, I can upgrade Duchy into Province or Province into Colony and just win; also, if I have Witch at all, I win immediately - and I am about even money to have at least one of these things happen for me (I also have some very small chance of being able to buy Province). Beyond that, I will probably buy an Estate. If I do that, my one point lead would let him win if: he Rebuilds a Duchy or Province (which requires him both drawing Rebuild and nabbing the correct hit off of its play) OR he has his only Witch AND (in either case) can buy Estate (or, of course, having enough money to buy Province - really unlikely considering he's just seen 3 Silvers). Note that he can't really buy an Estate that turn himself, since even if I were to whiff AGAIN the next turn, one more Estate from me would end the game. But even not buying that Estate, an extra one from me the following turn is worth 6 for me, really putting the squeeze on him (though he would indeed have some chance).

Alternatively, if he doesn't get the Silk Road (instead getting... I'll get to this in a second - for now, let's assume nothing, though that's obviously not strictly best), then what? Ok, well, I can win my next turn if I get Rebuild AND Witch. If I only get Rebuild, turning Estate into the last Silk Road and buying Curse would let me win by a point, but more likely I would upgrade an expensive green card, and then buy the best green card I could afford. If I get only Witch, I win if I can also hit $4. Failing this, I just buy the best green card I can, which gives him a turn to score some points (but he most likely NEEDS to do so right then). Alright, so what should he buy? Estate. It gives me no new wins, scores as many points as Silk Road, doesn't hurt his deck (as he'll never see it again), and there are some scenarios where I can't buy the last Curse, or he can GIVE me the last curse with Witch on his next turn, where it actually turns a loss into a win. Would it have changed the outcome of the game? As it turns out, almost certainly no (I would have been able to buy the last SR instead of curse on my turn 20, which gives him another turn, but he needs to Shanty Town into Witch into a good Rebuild, or ST into two good Rebuilds, which is very very unlikely). And in general, it's very unlikely to. But playing the slim percentages is something you have to do. More to the point, without the earlier too-soon-duchies, I probably would have simply had not enough time to get all the gains I need, he would have been able to get his Rebuilds going to the point where he has a few more points, and then I'd be locked out. These kinds of situations don't come up often, because I put myself into a position where I'm likely to lose, and quickly, which is something people don't like to do. But the volatility turns out to be my best chance, and as the player in the lead, you really need to be vigilant and know which way to turn, because if you aren't precise, you really can open up the door.

A Tale of Traps

Game Log

Setup:

First of all, I want to say I don't actually think this game was played particularly well, by any stretch, by either player. Indeed, the reason I'm posting this is that there were a number of errors being made which I see as fairly common even for players who are pretty strong. As such, I'm going to refrain from posting any images here which would show my opponent's name, as the point really isn't to shame him, but rather to try to point out these pitfalls so that you, the reader, will be able to avoid them. Obviously, you can figure out who it is if you really want to.

Trap 1: University
The University trap is very easy to fall into. There are lots of good and shiny 5-cost cards, and you want lots of them, so you go for it. The problem is, this is incredibly slow. You have to buy a Potion, buy a University, and only then can you start getting those 5s to pour in. A player who just went straightforwardly for buying those 5s the old-fashioned way is just going to be 5s ahead for several shuffles, and particularly if those 5s have an effect of improving your deck's ability to build itself (as is often true), perhaps just indefinitely. On top of this, as a village, University is a Necropolis - and that's card-negative, generally not where you want to be. Given all this, in order to make University worth it, you usually want some plan to make the game go late (often extra points) and/or ways to see those reshuffles quickly (i.e. deck-thinning). You do have some of that here (Tunnel and Island give more points, Spice Merchant thins a bit, Embassy and Margrave can help you draw). But the draw is pretty weak - only Margrave can combine with University to actually increase your hand-size, and then you are not doing it in a very potent way - the draw power is equivalent to Village-Moat. Yeah. So I really don't believe in University or the Engine here.

Trap 2: Opening Potion
This trap is a bit tough, and perhaps I shouldn't call it a trap, because it's very often correct. The thought is 'if I want cards that cost 2p or 3p, I should open potion'. And because you typically want those cards as soon as possible, this is indeed most of the time correct. But there are lots of exceptions. And I think this is likely one of those exceptions - if you want to go for the engine, I think you want to open Spice Merchant and defer those Universities for a while, as getting your deck thinned and spinning a bit is more important than loading up on those 5-costs super-fast (and indeed, you aren't so much slower anyway, as you are more likely to be able to just buy some 5s). A good example of delayed potion happened in a game this morning on Adam Horton's stream which can be found here.

Trap 3: Tunnel
People often like to rush for Tunnel+Enabler, thinking they'll gain lots of golds and get way ahead. This can sometimes be good, but generally it's quite mediocre, even with fairly decent enablers. More importantly, in doing this, you're playing a deck that is more like a slog than anything else - it takes an awful long time for you to get a decent number of golds out of this, and in the meantime, you have added a nothing card to your deck (and this is even assuming you do hit your collisions pretty well). In the long run, you do have a gold-heavy, fairly robust deck, and some extra points to boot. So it is generally like a slog, and not necessarily a great one. Of course, there are other ways to play the card. What doesn't make sense, though, is trying to get this to work in an engine. You're adding lots of non-draw cards to your deck, which makes it exceptionally difficult to draw your deck. Of course, the enablers usually sift you, so you may well still be able to have some ability to draw your important cards, but the point still remains that as you increase the number of cards you have in your deck, you decrease the chances of getting everything you need to come together. In this game, your draw isn't great, and especially given your few villages and mostly needing Embassy over Margrave to enable, you're eventually turning yourself into something near Province-per-turn mode. And keep in mind, you were pretty slow to get there. This really lowers your ability to reap the benefits of an engine.

Trap 4: One Copper Trasher
I have to give credit here for pointing this one out to Stef, the longtime top player in the game. Here's the deal: usually, you only want one copy of a copper trasher (Moneylender or especially Spice Merchant, as here). This is because you're going to run out of fuel for them fairly quickly, and there's diminishing returns on these cards, as you get to trash a lot fewer cards per trasher with each successive one. However, in games where there is some trash-for-benefit card, getting the second one gets much, much better, and is very often the thing you want to do. Here, I was the player who overlooked this: I slam lots of silvers and only one Spice Merchant. This is fine, but given my Butcher strategy, getting the second SM just seems a good bit better.



As for how the actual board goes? Well, in the game, my SM+2 Butchers was fine enought o overcome the wonky kind-of-engine from my opponent. In general, though, I am not sure which way to turn between Margrave-based BM, Embassy-based BM (picking up tunnels at the right moment), or 2SM+1-3Butcher BM.