Save gold for stronger rounds
Treat your gold as a budget, not as a scoreboard. You start with 6 gold, and the round income is 6, 6, 7, 7, then 8 from round 5 on. Every card costs 3 gold, a refresh costs 1 gold, and selling a card gives back 1 gold.
Unspent gold carries over, so a quiet round can be a savings account that pays for a stronger shop later. Holding gold instead of refreshing is like keeping cash in a high-yield savings account: small now, bigger later.
The Interest talent turns that savings account into compound interest. At round start you gain +1 gold per 2 gold you hold, up to +5, and holding 10 gold pays the full +5.
Treasury works from round 7: your total gets +1 per 2 gold you hold, up to +12, as in France + Spain + Portugal with 10 gold held in round 7 scores 13 + 5 = 18. That is return on investment measured before the fight begins.
Refreshing is a fee, not an investment. A refresh costs 1 gold, and Free Dispatch makes the first 2 refreshes each round free. Street Market makes cards cost 2 instead of 3 in rounds 1-4, which stretches gold into more purchases.
War Chest starts with 10 gold but every round gives 1 less gold, like a loan that raises your opening balance while lowering your income. Blood Tax loses 8 HP immediately, then gives +4 gold every round, closer to carrying a credit card balance: the gold is real, but the cost is paid in survival.
The most common gold mistake is refreshing for a card you cannot afford to keep. Selling a card returns only 1 gold while every card costs 3 gold, so a churn cycle is an early withdrawal fee on your savings account. Lock keeps your current shop for next round, like holding a quote.
When you do spend, spend for a concrete return on investment: completing a chain, enabling a homeland bonus, or filling a slot that stands on its own. A purchase that needs several more purchases to make sense is a loan against a board you may never finish.
Spend when the next fight is close and the upgrade has a clear payoff. Early rounds have smaller damage and smaller shops, so saving toward a stronger round can be the better budget. From round 5 the shop shows 5 cards and stronger cards appear; from round 9 it shows 6.
That is when a reserve has done its job and should become board strength. The board has 5 card slots and 2 effect card slots, so every purchase competes for limited space. A budget that never pays out is just a pile of gold with no return.
Interest and Treasury reward a held balance, but they do not reward hoarding through a losing board. If holding 10 gold pays +5 through Interest, that bonus is worth more when your board already has a chain to amplify.
If Treasury gives +12 at the cap from round 7, the gold must still support a real board. The budget question is not "how much gold can I keep" but "which purchase gives the larger return this round."
Chain neighbours for bigger scores
Border chains are the map version of a Schengen visa, where one permit lets you cross many neighbouring borders without stopping. Two neighbouring countries on your board form a border link, and links join into a chain. The chain bonus is 3 cards +5, 4 cards +10, 5 cards +18.
A connected route is easier to justify than a scattered set because every new border can strengthen the existing chain. The same logic applies to continent bonuses: 2 give +2, 3 give +3, 4 give +5, 5 or more give +8.
The simplest chain is France + Spain = 5. Extending it gives France + Spain + Portugal = 13, then France + Spain + Portugal + Italy = 22.
A full set shows the payoff: Germany + France + Spain + Portugal + Italy = 35, compared with 5 for the same five flags scored alone. A dense cluster can be even better because shared borders add links: France + Spain + Italy + Germany + Switzerland = 37, with France and Switzerland each touching several of the others.
Person cards reward the same connectivity when they sit with their home country. A person card placed with their home country adds their own score again, so France + Napoleon = 11 and China + Qin Shi Huang = 13. Napoleon alone = 5, which shows how much the homeland placement changes the result.
Think of a student visa: it is issued for study in one country, so a visa for Germany or France pays off only when you actually study there. A person works the same way and counts twice only next to their own home country. The same double count is why Portugal + Spain + Magellan + Vasco da Gama = 27, with two people sharing one homeland, Portugal.
Effect cards turn a good chain into a better return on investment. Edict gives a chain of 3 or more +6, so France + Spain + Portugal + Edict = 19. Tide gives each border link +2, so France + Spain + Portugal + Tide = 17.
Bloodline makes the homeland bonus count twice, so France + Napoleon + Bloodline = 16. Dynasty gives each person with their homeland +4, so France + Napoleon + Dynasty = 15.
Let the shop support the chain you are building. France + Spain + Portugal + Napoleon = 28 shows a chain plus a homeland person in a package. China + Mongolia + Qin Shi Huang + Genghis Khan = 39 shows a long connected group with two people in their home countries.
The practical rule is to count links before buying a shiny card, because a card that breaks a chain can cost more score than it adds. A chain is a route, and every stop should make the next stop worth more.
Build a floor to protect HP
Everyone starts with 32 HP, and the last player standing wins, so HP is an insurance policy with a coverage limit. The lower total loses HP: 1 base damage in rounds 1-4, 2 from round 5, 5 from round 9, 7 from round 12, plus 1 for every 4 points you are behind.
The damage cap is the deductible you can afford: at most 3 in rounds 1-4, 6 in rounds 5-8, 14 in rounds 9-11, and 20 from round 12. When only 3 players are left there is no cap, so the worst case becomes much more expensive.
Early damage limits make a weak round survivable. Losing by 4 in round 1 costs 2 HP, while losing by 40 in round 1 still costs only 3 HP. That is coverage doing its job.
In round 9 the same safety net is thinner: losing by 4 in round 9 costs 6 HP, losing by 20 costs 10 HP, and losing by 60 costs 14 HP, the cap. With 3 players left, the same 60-point loss costs 20 HP, so a floor line-up matters most when the caps come off.
Defensive talents are the deductibles and payouts that keep a bad round from becoming a knockout. Pavise reduces damage taken by 2, but at least 1, so a 10-damage hit becomes 8.
Phoenix heals 8 HP when picked and the first time you drop to 0 HP you come back with 5 HP, which is a comeback payout. Last Stand gives +12 total at 10 HP or less, so France + Spain + Portugal scores 25 instead of 13.
A reliable floor line-up does not need to win every fight. Effect card Militia gives each card with base score 1-2 +2, and talent Levy gives each card with base 1 +2.
Those bonuses are coverage for cheap cards: the floor holds while your better pieces come online. A floor line-up should still respect chains and homeland bonuses rather than filling space with random flags.
Protect HP when the cost of an upgrade is only marginal. A 3-gold card that breaks a chain can be a poor trade even if it looks stronger, because the lower total loses HP and the cap may not save you later.
The worst case is not a 60-point loss in round 9; it is repeated medium losses after the cap rises and after only 3 players remain. Insurance is cheapest before the accident, and a reliable floor is cheapest before the first bad shop.