This article won't tell you how to build tournament-winning Shadowfist decks. It will, hopefully, give you a general idea of how to build a deck that won't get blown away by experienced players. It also includes some hints on how to play the decks you create effectively, although it is by no means a guide to playing the game. (That's a different article.)
The ideas presented here are only those that work for me, and I know other people will have other approaches. By all means, send me comments and even entire alternative theories; I'll be very happy to provide them a home on the Web.
Much credit is due to Dave von Domelen's article on Shadowfist deck construction in Scrye #11, which you should buy if you can find it. While my strategy is different than his, it's certainly influenced by his recommendations. Last but not least, of course, the original deck construction guidelines in the rulebook are the seed of it all.
The first thing I think about is deck size. Smaller decks allow you to increase the probability that a given card will come up, but they also limit your options simply because they limit the number of cards you can have. Personally, I build sixty card decks, allowing myself up to ten cards of slippage, but rarely more than seventy cards. This is a pretty typical CCG deck size, but I've seen Shadowfist decks as small as fifty cards and as large as a couple of hundred.
I choose my cards in units of four, which is probably a holdover from my Magic days, since you can have up to five of a given card in Shadowfist. However, sixty cards divides into five groups of twelve cards, and twelve divides much more nicely into fours than into fives. Who cares? Well, there are five basic types of cards for the purpose of deck-building, so there are your five groups of twelve, and four cards is a fifteenth of your deck -- so in a sixty card deck, you can usually count on seeing at least one of any card you've got four of.
And planning on getting a specific card at some point in the game -- or at least a card to fulfill a specific purpose -- is what deck-building is all about. At the most basic level, you have to plan on getting a Feng Shui site and a character in your hand before you can win the game. You can't make plans much more specific than that unless you think in multiples, and you'll need to in order to build a good deck, so it's pretty much essential to start with that in mind.
Deck size dealt with, let's consider the different kinds of cards you can put into a deck.
Sites are the backbone of your deck. You need sites to win, you need them to generate power so you can win, and you need them for the useful effects they can provide. There are four ways sites can help you. They can count towards victory (Feng Shui sites), they can provide resources, they can provide power, and they can have a useful effect. No site does all four things; some do only one.
A fifth of your deck should be Feng Shui sites. These sites are going to be most of your power structure, so you need to choose them with an eye towards defense. For a rough rule of thumb, try to make sure at least half of them have a Body of 7 or higher; most of them should have a Body of 6 or higher. The stronger ones will be the ones you're trying to play in the front row. It's OK to mix in a few weak sites for the sake of the effect they generate, but if you have too many weak ones you'll wind up getting a lot of sites taken. Don't forget that your other sites may be weak as well, since most of the pure resource-providing sites have a Body of 5.
After you've chosen your Feng Shui sites, you can think about other sites you might want. I try to have about four other power providing sites in my deck. (If I have any non-power providing sites, I'll add an extra power providing card for each one. Mostly that's a site, but it can be something else.) Some of these will most likely be resource providers, whether that's an aligned site like the House on the Hill or a site that provides independent resources, like the Alchemist's Lair; the others will be aligned sites that require resources (a lot of these in Netherworld) or pure power providers like the Trade Center.
Characters are the second most important card type; you (generally) can't take sites away from other people -- which you need to do to win -- without characters. Characters can do three things. They can provide resources without requiring any (these are usually called foundation characters), they can provide a useful effect, and they can be thugs. A very few do all three; some do two.
First off, let's look at foundation characters. My basic rule with regard to foundation characters is that for every three cards you have in your deck requiring a given resource, you should have one foundation character of that type. Thus, on average, you'll end up with about one fifth of your deck made up of foundation characters. This ratio ensures that you'll almost always get a foundation character within the first turn if you're careful about which ones you pick and if you discard liberally. This is important because you don't want to get stuck the first few turns without a character to protect your site or sites; there are a number of ways to take an unprotected site on the first turn. Besides, the earlier you get out the resources you need to play your nasty cards, the better.
Once you've resigned yourself to putting all those low Fighting cards in your deck, there are a few more things you may want to consider. First of all, weight your foundation characters towards the cheap end. If you have a lot of foundation characters that cost four Power, you will be hosed early in the game. I almost always have five foundation characters costing one Power, and scatter the rest around to fit. I rarely use the four Fighting foundation characters, because they're difficult to get out early in the game and late in the game you'll likely be able to bring out characters who cost three for four Fighting, or cost four for six Fighting.
Second, in mixed faction decks, try to make sure at least a tenth of the deck provides resources for faction X if you have cards requiring three faction X resources to play. (In other words, about half the foundation characters should be of faction X.) If you have cards requiring four or five faction X resources, that faction should be most of the deck with any other factions being there solely for the sake of a few cards. Otherwise, you will be short of resources. Cards like Shi Ho Quai or Orbital Laser Strike that take a variable number of a given resource count as cards requiring four faction X resources to play.
Finally, yes, you can replace a foundation character with a resource generating site. Maybe even two. But probably not more than that. Remember: it's a lot easier to take sites than characters, and you don't want to lose those precious resources; also, you can't play a Feng Shui site and a resource providing site on the first turn.
Now, on to the other character types. Choosing these is a bit more tricky. In general, I try to have about as many thugs and effect generators as I have foundation characters, plus a few extra. (Those few extra, though, are almost completely an element of my play style, and many people play successful decks without them.) I make about two-thirds to three-quarters of these thugs, and the rest effect generators. However, if I'm playing a deck with a lot of Fighting boosters -- vehicles, weapons, and other States -- then I allow for that in my calculations, since one or two of those turns any character into a thug.
In general, any character with a Fighting of four or higher counts in my eyes as a thug. A character is a better thug if it has a special ability related to combat -- so the Masked Avenger, who ignores damage from certain characters, is about as good as a character with five Fighting and no special abilities, but the Brain Sucker, who turns off Masterminds, is just a thug who happens to have a useful ability.
This is a somewhat fuzzy concept. Cheap stuff is any card not already discussed with a Cost of zero or one. These are vital: they have effects that you can almost always use, and they allow you to cycle through your hand quickly. Quick cycling is one of the marks of a good deck, because the more cards you play in a game the more use you're getting out of your deck. Cheap stuff can be Events, States, or even (in some particularly annoying cases) Edges.
They should be at least a fifth of your deck, and probably more if you can manage it. Often, if you've mixed another faction in, that faction's cards will be mostly cheap stuff. This is especially true when you consider that devastating cards like Operation Killdeer, Final Brawl, Iron and Silk, Inexorable Corruption, Brain Fire, Death-O-Rama, and Helix Chewer are all zero Cost. I'll expand on this in the section on combining factions.
Beyond this, there's little I can recommend but experimentation. Again, the common cheap cards are the backbone of your deck; they are your surprise defense and your surprise offense. Pick them with a mind towards getting attacks to their target, because that's what it's all about in the end.
By now, you've accounted for around eighty percent of your deck. A fifth of it is Feng Shui sites, and a fifth of it is cheap States and Events. Unless you're building a deck with a lot of unaligned States and Events, you've got about two fifths of your deck in characters. That leaves, in a sixty card deck, twelve cards left to add. What do you do here?
Well, first off, you see if there are any expensive States, Events, and Edges you want to add. Edges that hose other factions are recommended, because the chances are the factions you can hose are exactly the ones that can hose you. For the rest, this is a good place to experiment; you can't hurt yourself too badly with things you put in here. Try out expensive States like the Combat Aircar, or Shadowy Mentor. Play around with the big ticket Events like Neutron Bomb. Almost all Shadowfist cards are balanced in play cost, so it's hard to screw this up.
It's also OK to add in cards that would normally count as one of the other card types. For example, I'll often add in a few more fighters here, because that fits my play style. You might want to load up on more cheap cards, too. If you're building a deck that relies heavily on getting as many of a given resource as possible, you may want to put in more inexpensive resource providing characters. This is probably the area that's the most fun to fill out.
When you're done, take a look at your deck. Make sure you didn't put anything in which will hose you -- if you're playing an Ascended/Lotus deck, you probably don't want Chains of Bone in there. Think a little about whether you've got something in there that will let you get past defenders, since you are going to have to take at least one site through heavy resistance to win. This can be tough fighters, or special abilities, or whatever, but it has to be there. Then shuffle the deck, and deal yourself out a few hands. Make sure you're getting resources out the first turn that can be used to play the cards you're getting. If you keep finding you aren't getting enough resources, retune the deck. And finally, go out and play it.
After play, of course, you'll want to tune it again. One friend of mine trashes everything he discarded during play and adds new cards to taste, which seems to me a little extreme. The thought is good, though; see what you're not using, and if there's a good reason why you aren't using it, cycle in something else. You probably also want to tune a bit to deal with the habits of your play group. You'll also want to try new cards, and new combinations. Again, the goal of the game is to have fun.
Finally, let's take a quick look at the issues surrounding multi-faction decks. Shadowfist is not Magic; single-faction decks are not, IMHO, as weak as single-color Magic decks. For one thing, most of the ways to screw up a given faction are symmetrical -- the Ascended can make the Lotus very unhappy, but the Lotus can do the same to the Ascended. Still, there are plenty of reasons to combine factions; you reduce your vulnerability, you can make up for weaknesses in the primary faction, and there's something really fun about playing a Demon Within on a $10,000 Man to turn him into a Demonic Jury-Rigged Cyborg.
The first of the classic ways to combine factions is to choose about ten cards from another faction that provide abilities the first faction doesn't have; then add five resource providing characters and either a few more expensive States and Events from that faction or a few fighters. For example, I often cover the weaknesses in my Dragon decks by adding five Students of the Bear, five Realpolitiks, five Operation Killdeers, a couple of Family Estates, and three Military Commanders. The Realpolitiks allow me to remove States, which the Dragons otherwise can't do. The Operation Killdeers help block damage, and the Military Commanders boost damage; these complement the basic Dragon philosophy. Finally, the Family Estates make it cheaper to get the Students and the Commanders out and provide me with more Power and more Ascended resources.
The second classic faction combination style is the theme deck. A theme deck is a deck built around a theme or a title, rather than sheer winning ability. My Ascended/Architect deck isn't really designed to be a killer deck, although it does well, but rather to tell the story of the Ascended scientists who captured a horde of Test Subjects and used their findings to construct advanced weapons from the future. There's not a whole lot to say about these decks, other than that you should keep the basic guidelines in mind while you're being creative. (For an excellent example of this sort of deck, check out the Sumo Surprise.)
Finally, one can build a multi-faction deck around combinations. This is pretty advanced, as you're not only working out strange combinations, but also trying to tune your deck such that you've got a good chance of getting them in play. I don't build these much, but a good example is the Architect/Dragon deck containing Vivisectors, Proving Grounds, Golden Comebacks, and the like. The general idea is that you use the Vivisectors in combination with the Golden Comebacks to generate power, which can be used on Neutron Bombs, which leave the Vivisectors enough time to smoke the characters all over again. In theory, this is a wonderful little loop, but the problem is making it work.
Videodrome is a deck I recently designed, recording the thinking that went into it. I'm also keeping track of how it does in play. After all, it's no fair presenting a deck design strategy without testing it, and hopefully seeing the thoughts I had as I worked will illuminate anything obscure I've written.
In theory, you should now be able to design a decent deck. In practice, well -- you'll have to find out in play. If you have any questions, or if you'd like opinions on your deck, I encourage you to post them to rec.games.trading-cards.misc where not only I but dozens of people who are better at this than me will happily tell you what they'd do in your situation.
And good luck!