Splendor Complete Guide 2026: Strategy, Expansions, Duel, and the Playmat Verdict

Buy standard Splendor when you need one clean game for two, three, or four. Buy Splendor Duel when two is the permanent player count and you want sharper interference. The two expansion boxes require standard Splendor; they do not expand Duel.

For an established standard-Splendor table, The Silk Road is the friendlier first expansion because Cities changes the finish line and Trading Posts rewards focused engines. The Sun Never Sets is the more interactive box: Orient expands the card market, while Strongholds lets players occupy and contest cards. None is required. Base Splendor remains the best teach and the easiest travel purchase.

Build three face-up rows of four development cards, one row from each level, and reveal nobles equal to the player count plus one. Use four gem tokens of each color at two players, five at three, and seven at four; all five gold jokers stay in every game. Each player begins with an empty hand and tableau.

On your turn take exactly one action: take three different gem colors; take two of one color only when that stack held at least four before you touched it; reserve one face-up card or the top card of a deck and take a gold if one remains; or buy one card from the display or your reserve. Purchased cards stay in your tableau and discount every later purchase of their color. You may hold at most ten tokens and three reserved cards. Nobles arrive automatically at the end of your turn when your permanent card bonuses meet their requirement. Reaching 15 prestige triggers the final equal-turn round; highest prestige wins, with fewer purchased development cards breaking a tie.

Do not buy the cheapest card merely because you can. First scan level two and level three for efficient prestige, then buy level-one discounts that shorten those exact routes. Early nobles are route signals, not mandatory quests: chase one only when its colors overlap cards you already want.

At an experienced table, estimate turns to 15, not current score. A ten-point opponent with a mature engine and an affordable five-point card may be one turn from ending the game; a twelve-point opponent with the wrong colors may be three away. Reserve when it protects a hinge card, denies a nearly guaranteed score, or converts a spare action into gold. Reserving speculative cards until all three slots are clogged is a quiet way to lose.

Duel is not standard Splendor with fewer tokens. Players draft gems and pearls from a shared 5-by-5 board, and taking a line changes what the opponent can reach. Privilege scrolls create a tactical response when a refill or greedy take gives the rival compensation. Crowns create a separate race, and there are three victory routes: reach 20 total prestige, reach 10 prestige in one color, or collect 10 crowns.

That changes the buying decision. Standard Splendor is smoother and scales to four. Duel is tighter, meaner, and better when the same pair wants repeated study. Neither replaces the other for every household.

The Silk Road contains two independently playable modules. Cities removes nobles and reveals three City tiles. A city asks for both prestige and specific permanent card colors; meeting one City’s full requirement triggers the final equal-turn round. If multiple players qualify, prestige breaks the tie, then the player with fewer development cards. It is a cleaner alternate finish for groups that find nobles too passive.

Trading Posts gives each player five matching post tiles. At the end of a turn, after resolving a noble, you may place at most one post whose card-color requirement you meet. The five powers are concrete: gain a non-gold gem after a purchase; take an extra different gem after taking two identical; make spent gold count as two gems of one color; score one prestige per post; or, when reserving blind, draw two cards and keep one while placing the other beneath its deck. Trading Posts accelerates specialized engines and is the module most likely to reward planning several purchases ahead.

Orient adds three small decks, one for each development level. Reveal two Orient cards beside each normal row. Their powers can reserve nobles, grant double permanent bonuses, attach to another card, or create more explosive purchases. The market becomes broader and stronger, so everyone must watch replacement cards instead of playing from a private plan.

Strongholds gives each player three fortress pieces. After buying a development card, place or move one of your strongholds onto a face-up card, or remove one opponent stronghold. A card occupied by an opponent cannot be bought. Put all three of your strongholds on one card and you may perform a special Conquest purchase after your normal action. This is Splendor’s most direct blocking module, and the easiest one to dislike if your group prefers low-conflict efficiency.

Cities of Splendor is the older 2017 expansion containing four modules: Cities, Trading Posts, Orient, and Strongholds. Its rulebook generally treated those modules as separate experiences. The newer Silk Road and Sun Never Sets boxes split those concepts into two releases, revise their rules and presentation, and explicitly include compatibility guidance. Do not mix old and new copies of the same module.

“Splendor Super Combo” is a retailer bundle name seen in Brazil and some import listings. It is not a fifth rules system or a deluxe edition designed by Space Cowboys. The bundle combines a base game with the older Cities of Splendor expansion, which is why its listing can promise four modules. If the contents are genuine and the language works for your table, it is a convenient old-line bundle; it is not automatically better than buying the current base plus one current expansion.

The current rulebooks allow modules to be played separately, together, and with other current Splendor expansions, but combinations need their printed interaction rules. Cities removes nobles, so any effect that refers to a noble must follow the Cities clarification. Orient adds cards to all three levels, while Strongholds can occupy eligible face-up cards; use the box’s exact exceptions.

For actual play, add one module for two games before combining. A sensible ladder is base, Cities, Trading Posts, then Cities plus Trading Posts. For the other box, try Orient before Strongholds. Only after every player understands both should you combine them. A legal four-module night is not automatically the best Splendor night; it increases market reading, setup, and exception load.

A playmat is nice to have, never required. Neoprene makes cards easier to lift, keeps the three development rows aligned, quiets the heavy chips, and protects glass or finished wood. It does not improve storage or strategy. A generic mat is more versatile than an edition-specific one and still gives nearly every handling benefit.

Sleeves are similarly conditional. Sleeve the development decks when the game travels, sees drinks, or gets played weekly. Do not automatically sleeve every expansion before checking insert tolerance. The base box’s best component is already included: the weighted chips.

Show one cheap card, pay its printed cost, and leave it in front of the learner. Then show a second card and physically subtract that permanent bonus. Explain the four actions, the ten-token cap, three-card reserve limit, nobles, and the 15-point trigger. Start.

Do not lecture about engine building. On the first two turns, point to one attractive level-two card and ask, “Which cheap card makes this arrive sooner?” That question teaches the entire game. Mention that gold comes only from reserving and that the player who triggers the end does not necessarily win. Leave expansion boxes closed.

Before every move, ask: How many turns does this save? What does it reveal? What does it deny? Count opponents’ chips and permanent bonuses, especially when they can buy a high-value card next turn. Denial is strongest when the same move advances you: reserve a card you can plausibly purchase, or take a scarce color you already need.

The decisive transition is knowing when to stop buying zero-point infrastructure. Once an engine can reach several scoring cards, calculate the shortest sequence to 15 and convert. A beautiful six-color tableau that delays scoring can lose to an ugly three-color route that finishes two actions earlier.