NOT RECOMMENDED
The fundamental problem is that the developers appear to have tested the game with advance knowledge of its systems. The player is not privy to any of it.
You arrive at a boss with a deck that isn’t optimized for that particular fight, equipment that doesn’t help, wounded characters, and a hand selected by RNGesus who decided today is your turn to be nailed to the cross. You get your ass kicked, try again, and keep losing until you respec, pare your deck down to precisely what the encounter demands, and hope the cards arrive in the correct order.
That's not good encounter design. It's an answer key disguised as character building.
The Blacksmith encounter is a perfect example. The game allows you to attack the Fire Priest, which suggests that you might be able to defeat him and save the Blacksmith intact. The Priest has a comically inflated health pool, however, and the real objective is to decapitate the Blacksmith. Fighting the Priest is technically presented as an option, but the encounter makes that option functionally meaningless with the resources available at that point. You were never really meant to choose. You were meant to discover the action the developers had already selected for you.
Later, you are expected to save a were-creature by applying a mixture to it three times during combat. What happens if the necessary card doesn’t appear when you need it? You can end up killing the creature because you didn’t know in advance how the encounter was scripted and hadn’t optimized your deck around repeatedly drawing that one card.
The game is full of encounters like this. It frequently presents what looks like a tactical or moral choice when there is actually one intended solution enforced by card availability, inflated enemy statistics, or hidden encounter rules.
Health persistence creates another problem. You often finish one fight with the party looking as though it went through a food processor, only to be thrown into another battle without meaningful healing or rest. The sensible move is sometimes to enter the next encounter, lose, respawn at full health, and then make a serious attempt.
The game has turned death into the ultimate healing device.
Attrition can produce excellent tension when players can manage supplies, choose when to rest, retreat, or decide whether to risk pressing onward. Here, victory can leave you in a worse strategic position than dying. Your punishment for surviving one battle is starting the next one half dead; your reward for losing is a fully restored party.
Equipment balance is equally inconsistent.
The Scholar obtains the Sun Attire early in the game. It restores his shields to maximum every time he plays a Scholar card. The armor initially provides only twelve shields, but it can be upgraded, and the Scholar possesses numerous zero-cost cards. With modest planning, he can repeatedly restore himself to full shields between enemy attacks while performing actions he already wanted to take.
Why would I replace that?
Later armor might offer higher numbers or conditional effects, but very little competes with automatically restoring your entire shield whenever you play one of the character’s abundant class cards. The game hands you what is effectively an early best-in-slot defensive engine and then continues dropping armor that is far less reliable.
The Memory of the Materialist has the opposite balance problem. It damages every enemy according to the number of cards discarded during a redraw. That means discarding five or six cards can inflict five or six guaranteed damage on every enemy on the board, every round.
Against the hordes the game throws at you, redrawing your entire hand can be considerably more efficient than using actions to attack individual enemies. The relic bypasses bad draws, card costs, targeting restrictions, and much of the ordinary combat system. It isn’t merely strong. It is an escape hatch from the game’s combat design. Sometimes the most effective way to play this card game is to refuse to play your cards.
Then there is the Memory of a Trader, which may be the most useless trinket I found. It reduces the time cost of the wearer’s cards by one but only during the first round.
You control three characters, and each draw randomly produces cards belonging to all three. Even if you load one character’s deck with one-cost cards, there is no guarantee that you will draw enough of them during the opening round to justify occupying an equipment slot. Perhaps you get lucky and receive some free attacks or buffs. Perhaps you draw one card whose discount barely matters. Either way, after that first round the trinket does absolutely nothing for the remainder of the battle.
It is a high-variance relic occupying a permanent equipment slot in a game that already punishes variance.
These are not isolated balance quirks. Weak equipment depends upon random conditions, while the strongest equipment works by insulating you from the randomness of the card system altogether.
This is not much of an RPG. It is a railroad with a card game attached.
The developers have now added the ability to refund skill points. That is better than the original system, which allowed players to become trapped by earlier choices and accumulate upgraded cards that were actively harmful to later decks. However, respecs address the symptom rather than the cause.
The encounters still do not adequately support a broad range of reasonable builds. Now, when your deck cannot pass the next test, the game simply allows you to change your answers after seeing the question. Reach a boss, discover which cards it demands, refund your choices, reconstruct your deck, and try again.
Experience and progression are tightly controlled. Areas remain locked until the game permits you to enter them, encounters are finite, and enemies do not randomly respawn for additional experience. You cannot grind a few levels to compensate for an unconventional build. The developers largely know what level you will be, what equipment you can possess, and which cards can be available when you reach a particular encounter. Despite controlling those variables, the game still needed a respec system and even a major area rebalance after release.
That suggests the developers proved that their preferred builds could finish the game. They did not adequately test whether the range of builds encouraged by their own cards, skill trees, relics, and armor remained viable.
The music also deserves a warning. Dzivia’s magnificent work helped lure me into the game, but his released album material is heavily front-loaded and reused. “Dead Deities” and “Vataha” appear under their proper names, while the eleven-minute “Imperivm” is reused under multiple incorrect cue names. “Ascension,” which appears in the trailer, isn’t even included in the in-game soundtrack. Much of the later music is comparatively generic. It isn’t bad, but it probably won’t stay with you after turning off the game the way Dzivia’s work does.
Not everything here is terrible. The artwork is excellent, the world is dark and atmospheric, and there are some wonderfully grim jokes. The story remains interesting enough that I continue playing despite my frustration. The underlying card mechanics can also be satisfying when the game permits you to use them normally rather than demanding an encounter-specific solution.
But that isn’t enough for me to recommend it.
If you can tolerate dying because you brought the “wrong” reasonable deck, arriving at battles already mauled, rebuilding your character to counter individual bosses, and following a rigid sequence of objectives through areas that open only when the game permits, you may enjoy Shards of Order. It has style, atmosphere, and some genuinely compelling ideas.
For everyone else, wait for a STEAM sale. You can always refund the game if you hate it—which, appropriately enough, may be its most functional respec system.
