Hello, tabletop friends in the community, I'm linlin.
If you compare the table of Splendor to a Renaissance-era commercial city-state, then the thing most likely to bankrupt a gem merchant may be exactly that row of increasingly cheap mines right in front of them.
Imagine a table like this: the five colors are neatly arranged in front of you, and every time you buy a new card, you pay several fewer tokens. The zero-point card just flipped can be taken for free, and as your fingertip brushes the corner of the card, the abacus of the next chain purchase is already clicking in your mind. You look up, and your opponent pushes a development card with points toward you, casually welcomes a noble, and announces they have passed fifteen points.
Your commercial empire is finally built, and the game is almost over.
Why do the people who look richer and richer come closer and closer to losing? Because in Splendor, the hardest asset to value is the actions you still have time to take. Tokens can buy cards, discounts can save money, but turns already spent will not be refunded just because the engine is pretty.
In today's article, we continue with another ledger behind tokens and discounts: when to keep expanding, and when to hit the brakes and cash the whole engine into points.
1. A free mine still costs a turn: do a payback calculation for discounts
Seeing a zero-point card that costs no tokens easily creates an illusion: it's free anyway, so why not take it.
But the card's price tag does not list the full cost. On the turn you buy it, you cannot also take tokens, reserve a target others are also eyeing, or directly buy another scoring card. Whether it is worth it depends on whether the resources saved can truly pay off in the remaining game.
Push each current purchase route two steps forward, and the problem becomes clear.
Suppose you have already identified a target, and after subtracting your existing discounts and the tokens you can pay from hand, you are only two red short. The public red tokens are enough to take two, and the target has not been snatched away for now: take two red directly, then buy, and you can cash out in two of your own turns.
If you detour to buy a zero-point card that provides a red discount, and after buying you are still one red short with no gold in hand, then you still need to take tokens and then buy. Two turns become three; you did pay one fewer red, but you did not get the points faster.
However, if this red card can also save another target one token-taking step, or simultaneously complete a noble requirement, the calculation must be redone. The value of a discount comes from stacking with later routes; you cannot stare only at the card in front of you, nor can you turn "never buy zero-point cards" into a new dogma.
What really needs practice is: before reaching out, first find a future in which this card can pay off.
2. The tide of the token pool: taking the most does not necessarily move the fastest
If discounts determine your long-term purchasing power, then the tokens in hand are the waterway for your next transaction.
Three different colors gets you three tokens; two of the same color gets you two. The former gives more tokens, the latter concentrates colors. Which is more effective depends on what the target still lacks.
If the same card still lacks one red, one blue, and one green, and all three colors can be taken, then three different colors can fill the gap at once. Force-taking two red leaves the extra red temporarily useless while the gaps in blue and green remain.
Conversely, if the target is only two red short and the public red pile has at least four before you take, then two of the same color is more direct. Do not take two colors you cannot use right now just to "earn one extra token," and then spend another turn repairing the original gap.
There is another easily overlooked variable here: the public piles rise again after you pay.
When you buy a card and return scarce white tokens, you may exactly let the next player complete their high-scoring card. You cannot refuse to buy on that basis alone, but among several cards you can afford, what resources will be released after payment is worth another look.
In two-player games especially, watch this exchange of first and second position. You gain a development card, and your opponent may thereby gain the last batch of tokens leading to the endgame. Calculating only your own income will miss the other half of the table.
3. Gold is not a prize, and reserving is not free: leave an exit for the safe
The moment in Splendor that most resembles commercial warfare is when you have spent ages accumulating tokens, and your opponent calmly tucks the target into their hand.
This is certainly aggressive, but a beautiful snipe does not equal a profitable snipe.
Reserving consumes your action and also occupies a limited hand slot. The base game allows at most three unpurchased development cards to be kept, and they cannot be discarded casually. When the gold pool has tokens remaining, reserving can take one gold; when the pool has no gold, reserving can still be performed, but without that wild token to compensate you.
Therefore, before deciding to reserve a card, first look for an exit along your opponent's route.
Is this the only card with which they can widen the score gap on their next move? If another card beside it is also affordable to them, then taking one away only makes the choice for them. If you could originally have bought points directly, but instead turned twice in a row to block someone, and their alternative route remains open, you may be buying time for the whole table while moving yourself farther from victory.
A more attractive reservation is when defense and redirection happen to fall on the same card: the opponent wants to buy it, your discount can also support it; a color you cannot currently obtain can be filled in with the gold you gain; and after buying this card, it can also connect to your next score.
But you cannot secretly count one step ahead here. Reserving and buying are two actions; even if you can already afford it after taking the gold, you must wait until a later turn of your own to buy. Mistaking "I have locked it down" for "I have scored" will misalign the entire endgame timetable.
4. The noble's invitation: three points along the way, and three points by detour
Nobles are where people most easily develop route obsession.
Even when several color cards are still missing, your eyes keep getting drawn to those three points: one more white, one more green, wait for the next refresh, and perhaps it will soon be complete.
The problem is that opponents are not necessarily willing to wait for your collection to finish.
Nobles should enter your purchase ledger, but should not monopolize the entire ledger. If a scoring card you were going to buy anyway happens to complete the last bonus requirement of a noble, those three points are tightly connected to the main route; if chasing a noble requires filling in several cards in a row that help other targets very little, then those extra turns must be counted into the cost.
For example, if you are one blue bonus away from a noble, and an affordable blue card with points simultaneously satisfies the requirement, that is a compound return worth comparing closely. Conversely, ignoring cost and the opponent's endgame threat just because the card face has blue may mean spending a very long setup to exchange for an invitation that cannot be cashed in time.
Also hold the rule boundary: nobles check the bonuses of purchased cards, and blue tokens in hand cannot count toward them; even if multiple nobles are satisfied in the same turn, only one can be chosen. Strategy can be bold, but scoring cannot invite future guests in early.
5. Fifteen points is only the endgame alarm: check seating first, then press the switch
What truly determines the quality of the finish is often whether you saw, before crossing the line, who else at this table has not yet acted.
In the base game, after someone reaches fifteen points, the current round must be completed so that everyone has the same number of actions. Therefore, the first person to reach fifteen points is not necessarily the final winner.
Suppose in a two-player game you are the first to act this round, currently at twelve points, and buying a three-point card would take you to fifteen; your opponent is at fourteen and can already buy a two-point card. If you directly sound the alarm, your opponent still has a response opportunity this round, and can then exceed you at sixteen.
At this point, another card that incidentally triggers a noble, a reservation of a key scoring card, or a route that does not immediately trigger the end are all worth comparing. But every route must be rechecked: after you block this one, does the opponent have another they can buy? If you wait one more round, will your extra gain pay off in time? Have you left a noble to someone else?
Do not understand this example as "never end while ahead." What it reminds you is: endgame calculation must include the actions the opponent still has, and cannot stop just because your own score crosses fifteen.
If the final scores are equal, the number of purchased development cards is also compared, and the one with fewer is favored. Those zero-point expansions that no longer have time to pay back can even become a burden in the tiebreak comparison.
Conclusion: a truly mature engine knows when to stop growing
What is fascinating about Splendor is that the same cheap mine can be either the key that opens a trade route or a beautiful fork that makes you one turn late.
From discounts to tokens, from gold to nobles, every visible asset in front of you must pass the same question: can it, before the game ends, deliver me to a better result?
Next time you sit down, perhaps leave two lines in your purchase plan: one for how to cash out when things go most smoothly, and another for how to reroute after a key card is taken. First estimate how many of your own turns you still need, then look up at the opponent's public score, discounts, and tokens. You may not be able to read all their reserved cards, but at least you can see whether you are still building a distant future for a game that is about to end.
The gem merchant's last craft is to be willing to stop expanding when the machine is humming most pleasantly, and turn it into points that truly belong to you.
Applicable version and rules basis: Space Cowboys' Splendor 2014 base game, excluding expansions and Splendor Duel. Official English rules PDF page 2 "Taking Tokens, Reserving, Purchasing," page 3 "Bonuses, Nobles, End of the Game." All situations in the text are strategic deductions; when later actions are involved, the conditions are that the target has not been taken and resources can still be obtained.
#Splendor# #EngineBuilding# #TurnEfficiency# #ReservationMindGames# #EndgameCalculation#