The safest early unlock in Shards of Order is not the card with the largest damage number. It is the card that completes a line your shared deck can already start and finish. Lev, Ayla, and Thade pull from one pool, so an exciting card for one hero can delay defense, arrow recovery, or mana setup for everyone. Build one compact engine per hero, test what the hand fails to produce, and stop adding cards when the problem is already solved.
This guide was checked on August 11, 2026 for the released Steam game. The latest official developer note in the Steam feed is Hotfix v1.0f21, which reduced skill-tree tier costs, improved respec handling and intent hover, corrected a displayed movement cost with Momentum, and fixed targeting and combat issues. Those changes make early experimentation cheaper, but they also mean pre-launch demo counts and old fixed deck recipes are not a safe basis for the current client.
Start from the hand that failed
Do not open the skill tree with a permanent tier list in mind. Finish a normal encounter, name the exact moment the deck stopped working, and spend the next point against that failure. “Damage felt low” is too vague. “Lev reached high Strength but drew no payoff,” “Ayla emptied the Quiver and could not reload,” or “Thade drew a mana spender before the party generated its colors” describes an edit you can test.
| Failure visible in combat | Check before unlocking anything | Useful type of next edit |
|---|---|---|
| The hand has setup but no finish | Was the payoff absent, or merely too expensive before the next attack? | One payoff that consumes the resource already being made |
| The hand has several finishers | How many could legally resolve before the nearest countdown? | Remove a redundant finisher; add no new one yet |
| A hero’s resource stays empty | Is the generator missing, or being drawn after its spender? | A cheap generator or a card that is useful in both states |
| Redraw repeatedly finds the same dead role | How many cards perform that role, including equipment injections? | Prune duplicates before buying a different answer |
| One knockout makes every later hand worse | Was the original problem damage, positioning, or recovery timing? | Fix prevention first; do not call Revive clutter a damage problem |
This is also the cleanest way to use the cheaper tiers in v1.0f21. A respec is an experiment, not a reason to activate every affordable node. Change one branch or one pair of cards, repeat a comparable fight, and keep the change only if the failed hand becomes playable without creating a new dead draw.
Price the whole sequence in Time
A combo is only real if it fits before the action that matters. Hover an enemy’s intent, identify the threatened position, and read the current Redraw cost before queueing cards. Include movement in the bill. Hotfix v1.0f21 fixed both the Momentum movement display and cases where enemies attacked a slot they were not showing, so update first if the hover line and the resolved target disagree.
The released-game references agree that Redraw begins at four Time and becomes cheaper as actual Time is spent. Some older demo-oriented pages still describe a three-card threshold. Do not build around that stale shortcut. Two expensive plays and four zero-cost plays are not interchangeable: watch the live Redraw number and the enemy countdowns, then stop the queue before either reaches the wrong state.
Use a three-part price check for every proposed line:
- Setup: count the Time needed to create Rage, Strength, arrows, Sharp, Rot, or colored mana.
- Payoff: add the finisher’s current printed cost, including any encounter effect that has changed it.
- Exit: reserve enough room to move, defend, remove a priority target, or accept the Redraw that follows.
If the line spends the whole safe window on setup, the payoff is not “almost online.” It belongs to a later cycle. If it resolves but leaves no exit against the next clock, remove one setup card or keep a cheaper payoff in the deck. This arithmetic matters more than a screenshot’s damage value because card balance and equipment can change while enemy intent remains the immediate truth.
Give Lev one Rage loop and one conversion
Lev’s early cards divide into resource builders, Strength scaling, and cards that reward being the intended target. Begin with the pair that makes his next action better rather than collecting every card from all three groups.
Commanding Shout is a flexible first engine piece because it adds Strength to a chosen character while building Lev’s Rage. It is useful when Lev will finish the sequence, but it can also prepare Ayla or Thade instead of becoming dead support after Lev has enough resources. Reckless Strike is the more aggressive partner: its current early version is a low-Time attack that adds two Rage. Together they answer “how do I start?” without requiring the finisher to be in the opening hand.
Choose the conversion by the fight you are actually losing:
- Crushing Blow applies Strength twice to its damage. Take it when Strength reliably appears before the payoff; skip it if the deck usually draws the attack before its buffs.
- Duel pays defensively when Lev is the enemy’s current target, turning Rage into Shield. It is strongest when intent hover and positioning already make that condition deliberate. If Lev is not targeted, do not pretend the defensive half happened.
- Irritation trades Lev’s health for a zero-Time hit and a cheaper Redraw. Use it to repair a cycle that is one step short, not as a free card in a party already losing health to Fracture.
- The Judge begins a staged chain by putting The Jury on top of the draw pile, which in turn leads to The Execution. That predictability is valuable, but the later cards still consume Time. Add the chain only when the deck can pay for its planned follow-up instead of abandoning defense to chase it.
A practical Lev test hand is one builder, one conversion, and one defensive option. If both Commanding Shout and Reckless Strike appear but neither chosen payoff does, the deck needs pruning before it needs a second finisher. If the payoff appears too early, retain the more generally useful builder and remove whichever narrow card cannot help another party member.
Make Ayla reload in the same way she spends
Ayla already has Barrage when she joins. That gives the deck an arrow payoff before it needs another one. Her first unlock problem is therefore not “find burst”; it is “choose how the Quiver becomes ready again.” Mixing every reload style produces hands full of preparation after the Quiver is already full.
Lock and Load is the cleanest manual starter. It loads an arrow and deals damage that improves with the arrows already present, so it is not useless at either end of the refill. Build around it if you prefer to inspect the Quiver and decide exactly when to prepare Barrage.
The discard route uses Quickstrike and Salvo as a pair. Quickstrike asks for another card to be discarded; Salvo can be played as damage or sacrificed to load arrows. The pair is flexible only when both modes remain relevant. If Salvo is always discarded, ask whether a direct reload card would require fewer moving parts. If Quickstrike repeatedly consumes the defensive card you needed, the deck does not have “versatility”—it has an unsafe discard tax.
Powerload and Fully Automatic support a different rhythm. Powerload rewards a Redraw by adding an arrow, which fits a deck already cycling on schedule. Fully Automatic refills the Quiver immediately and is best after Barrage has emptied it. Do not add both merely because both mention arrows. Powerload wants repeated planned redraws; Fully Automatic wants a clear empty-to-full burst cycle.
Keep Caltrops as a matchup answer rather than part of every damage loop. Its party-wide Weakness is valuable against enemies whose danger comes from several hits. A single huge attack is better answered by Shield, movement, countdown control, or removing the attacker. The failure signal is visible: if Weakness barely changes the threatened hit, stop spending a slot on the wrong kind of mitigation.
Choose Thade’s destination before his mana cards
Thade gains colored mana as the party plays cards associated with the three heroes. His expensive cards can therefore be dead for two separate reasons: the deck did not generate the required colors, or it generated them only after the safe Time window closed. Pick the destination first, then keep only the generators and spenders that reach it.
The shortest early route begins with Rot. It does not require mana and applies a persistent effect for a low current Time cost, so it remains usable when Thade joins a hand before the party has built colors. From there, choose one continuation:
- Contagion asks for red mana and becomes most valuable when Rot itself finishes the target and transfers onward. Apply it to a target that can actually die from the effect; using it on a healthy boss removes the condition that makes the card special.
- Decay is a later commitment behind Freeze and Rot. It applies Rot based on the target’s current Time counter, so countdown manipulation and target selection matter more than adding another generic damage card.
- Arcane Path consumes all three mana colors and can improve its own damage when it scores a kill. Treat it as an engine destination. If the deck cannot present all colors before a safe kill, it is a draw tax rather than a scaling plan.
- Unstable Currents exhausts after use and adds damage when mana is consumed. It belongs in a mana-spending sequence, not in a Rot list that rarely pays mana after setup.
Equipment can create a different Arcane route. The current combat reference gives Crystal Staff plus Shifting Fabric as a concrete example: one interacts with Redraw and Arcane Bolt, while the other can add more Arcane Bolt copies. That is a synergy, but it is also a bloat test. Equip the pair, count how often added Bolts replace the defense or colored setup the hand needed, and remove the armor if the engine makes Thade stronger while making the party less playable.
Use a staged unlock plan, not a universal order
The three heroes gain and spend points on their own trees, so there is no honest single list from card one to card fifteen. Use stages instead. At each stage, advance only the character whose current line lacks a required job.
| Stage | Lev | Ayla | Thade | Stop condition |
|---|---|---|---|---|
| Establish | Commanding Shout or Reckless Strike | Lock and Load | Rot | Each hero can contribute before a specialized resource is ready |
| Convert | Crushing Blow or conditional Duel | Quickstrike plus Salvo, or a direct refill route | Contagion for Rot, or the setup leading to a chosen mana spender | Every generated resource has one reliable use |
| Repair | Irritation only if the cycle needs Redraw relief | Caltrops for repeated-hit pressure; Powerload for redraw rhythm | A defensive or support card if the engine arrives too slowly | The new card fixes a named failure in two comparable encounters |
| Commit | The Judge chain when the deck can afford its follow-ups | Fully Automatic when the plan repeatedly empties the Quiver | Decay, Arcane Path, or Unstable Currents after prerequisites work | Adding another payoff would compete with the one already online |
The important column is the last one. Once a stage works, continue the campaign with it. Do not buy a second answer because the point is available. Hotfix v1.0f21 lowered tier costs, but a cheaper unlock still occupies the shared pool when selected for combat.
Audit equipment as cards, not decoration
Weapons and armor can define a hero’s mechanic and add cards to the shared deck. That means a higher stat line may still be a downgrade to the draw. Before equipping a new piece, record the current number of cards serving each job: setup, payoff, defense, movement or control, and recovery. Equip the item, then repeat the count with its injections included.
Test equipment in an ordinary encounter before carrying it into a boss. Draw until you have seen every injected card at least once and ask three questions:
- Can the card be played when its owner is low on Rage, arrows, or mana?
- Does it replace an existing card’s job, or create a new job that needs more support?
- When it is the wrong answer, can it be discarded, sacrificed, exhausted, or used cheaply enough to leave the hand?
Keep a piece when its worst draw is still manageable. Reject it when the best-case combo is spectacular but every ordinary hand must carry its setup. This is especially important with duplicate injections: two items that separately improve a hero may together crowd the party’s only deck.
Separate a bad deck from a damaged combat state
When a hero falls, that hero’s cards become one-Time Revive actions. Using ordinary Revive returns the hero but adds Fracture pressure, shown by the red value around Redraw. From that point, a hand can fail because the original list was weak, because useful cards have become Revives, or because cycling now damages the party. Do not change the skill tree until you know which state caused the loss.
Use the visible Fracture number rather than memorizing a universal damage promise. Before Redraw, compare that cost with every hero’s health and the nearest enemy clock. If the redraw itself will create another knockout, play a useful card, stabilize, or restart the attempt; adding a high-damage unlock after the fight does not repair that arithmetic.
Ethereal Power can restore fallen heroes and supplies an Ethereal card, but it is limited to once per battle and is linked by current walkthrough evidence to later campaign consequences. Treat it as a run decision, not a deck component you assume will cover every glass-cannon test. If a build becomes functional only after that emergency button, the test has shown where the build fails.
Swap answers for encounter rules without rebuilding the article’s walkthrough
A compact deck still needs a few situational edits. Make them from the encounter’s visible rule, then remove them when that rule is gone:
- An exact-Time objective: keep several playable costs and watch their current values. The Frozen Marina Whale checks the total Time spent before Redraw, so a deck made only of one large finisher and zero-cost setup can be awkward even if its damage is high.
- A shielded one-health target: Rot is a direct mechanical answer because it can bypass the need to chew through the entire Shield. This justifies keeping Rot; it does not justify adding every Rot payoff.
- Hidden or altered card costs: stop queueing from memory. The Wine King changes what the cards cost, so build decisions must follow the visible current value rather than the price recorded during a previous fight.
- Several fast attacks: Caltrops or another wide Weakness source becomes more valuable. Against one heavy hit, reposition Lev, build Shield, delay the threat, or remove it instead.
- An invulnerable boss with an active object: prioritize cards that can hit the actual objective without waiting for a long resource engine. A perfect finisher aimed at an invalid target is still a dead card; v1.0f21 specifically fixed target-requiring cards when no legal target exists.
For the region routes, puzzle inputs, boss phases, and ending checkpoint behind these tests, use the existing Shards of Order walkthrough. This article deliberately stops at deck selection and diagnosis so the two guides solve different player tasks.
Respec when the engine’s premise is wrong
Pruning one card fixes a noisy list. A respec is appropriate when the engine assumes a state the party rarely reaches. Remove the branch, not just its final card, when any of these remains true across comparable fights:
- Lev’s Strength payoff repeatedly arrives before Strength and cannot serve another job.
- Ayla reaches Barrage with an empty Quiver even though reload cards were drawn and playable.
- Thade holds a three-color spender while one color consistently fails to appear before the safe window closes.
- An equipment injection creates more dead draws than the item’s mechanic rescues.
- The deck needs Ethereal Power in ordinary encounters rather than during an exceptional collapse.
After the respec, test with one engine per hero: a Lev builder plus conversion, an Ayla reload plus spend, and a Thade destination plus the mana or Rot path that reaches it. If those six jobs appear in playable order, the shared deck is doing its work. The next upgrade should answer a new encounter rule, not reward you for opening the skill tree.
Sources and current boundary
Identity and release facts come from the Steam listing for Shards of Order, its API record, and Fardust’s news posts. Together they establish the released PC game, Fardust and Awaken Realms, the shared-deck and equipment premise, official screenshots, and the v1.0f21 boundary. Named early cards and current-client interactions were cross-checked against readable community guides dated around launch. The unofficial wiki’s demo-only card totals, fixed deck percentages, pre-launch review language, and conflicting three-card draw rule were not used as released-game facts. Published visuals are official Steam JPGs; this article embeds no community screenshot or video.