Big Two Strategy: Control, Counting & the Endgame
Nobody wins Big Two by holding the best cards. You win by spending them in the right order — and by noticing when the cheapest card that wins is the one to play.
Big Two is a race with interference. Both players are trying to empty a hand of thirteen cards, and the only way to slow the other one down is to take the lead away from them. That single fact explains most good play: the strongest hand in the world loses if it never gets to choose what the exchange is about. Nothing below is a solved line — the deal is random and half the deck is hidden, so there is no move that always wins. These are the principles that raise the quality of your decisions, in the order they tend to matter. They follow the Paofuqi two-player rules; if you need the rules themselves first, start with the complete Big Two guide.
Step 1: Understand what you are actually racing for
Paofuqi does not score Big Two. There is no penalty for the cards left in your hand and no reward for finishing second — the first player to run out wins, and nothing else is measured. That simplifies your goal enormously: you are not trying to shed cards, you are trying to shed them faster than your opponent.
The practical version of that is a counting question you should ask on the very first turn: how many plays do I need? Thirteen singles is thirteen turns. Thirteen cards arranged as a five-card straight, two pairs, a triple and a single is five turns. The player who needs fewer plays is ahead in the race before a single card is on the table.
So the first thing to do with a new hand is not to look for your biggest card. It is to look for the five-card combinations, because they are the only plays that move five cards at once — and because they are the easiest to destroy by accident when you spend one of their cards to win an unimportant exchange.
Step 2: The lead is the prize, not the trick
Winning an exchange in Big Two gives you something far more valuable than the cards you just played: the right to choose what the next exchange is about. When you lead you decide the size — one card, two, three or five — and your opponent has to answer in that shape or pass.
This is the whole reason weak-looking cards are playable. Leading a low pair is not a waste; it is a probe that costs you two cards and either takes another turn off your plan or forces the opponent to spend a real pair to stop you.
It also means the two questions you ask on your own turn are different. When you are following, the question is what is the cheapest thing that beats this? When you are leading, the question is what shape is my opponent least able to answer?
In a two-player game the lead swings faster than people expect, because a single pass ends the trick. If your opponent passes, you did not just win those cards — you get the table back immediately.
Step 3: Spend the cheapest card that wins
The most common way to lose a hand you should have won is to overpay. If a queen is on the table and you hold both a king and a 2, the king wins the exchange just as completely as the 2 does — and it leaves you holding the single most useful card in the deck.
There are four 2s and each of them is an unanswerable single. A 2 in your hand is not really a card; it is a guaranteed recovery of the lead, at the exact moment of your choosing. Spending one to win a trick you could have won with a jack trades a certainty for a card.
The same principle applies to structure, not just to rank. Beating a single with a card pulled out of your straight costs you the straight — you win five cards' worth of tempo and give up a five-card play. Look at what a card is doing in your hand before you look at what it beats.
The exception is the endgame, where the value of a held 2 collapses to zero the moment the hand ends. A 2 you never spend is a card you lost with.
Step 4: Count what has not appeared — and remember half the deck is missing
Counting in the two-player game works differently from every four-player version, and it is the single biggest adjustment. Only 26 of the 52 cards are dealt. The other 26 never enter the game and nobody sees them.
So when a card you fear has not appeared, there are two explanations and not one: your opponent is holding it, or it was never in play. If you hold the K♠ and two of the other kings have already been played, the last one is as likely to be sitting in the undealt half as in your opponent's hand. In a four-player game that same count would be certainty; here it is a probability, and playing as if it were certainty is a mistake in both directions — too timid with good cards, too bold with medium ones.
What is certain is the count of your opponent's hand, which is on the table in front of you at all times. Watch it more closely than you watch their cards. A player who has gone from 13 to 4 in four turns is playing five-card hands and is winning the race whatever their cards look like.
Two things are worth tracking properly rather than vaguely: how many 2s have gone, because each one that appears makes your own high singles safer, and whether the opponent has passed on a size. A pass on a pair is real information — for the rest of the hand, that shape is a weakness you can attack.
Step 5: Pass on purpose
Passing is not surrender, and the beginner mistake is treating it as one. If beating the table costs you a card that is holding a plan together, passing keeps your structure intact and costs you only this one exchange.
Pass when the exchange is cheap and the answer is expensive. Somebody leads a 7 and your only answer is to break your straight or spend a 2 on it — let it go. Pass when you are already winning the race: if you need three more plays and your opponent needs six, you do not have to contest anything except the exchanges that actually threaten to run them out.
Do not pass when the count says the hand is nearly over. Once your opponent is down to two or three cards, every trick they win might be their last — at that point, spending your best card to take the lead away is not extravagant, it is the whole game.
And remember the one moment you cannot pass: leading an empty table. That is why holding at least one comfortable lead — a low pair, a spare single you do not mind losing — is worth planning for. Being forced to open a fresh trick with a card you wanted to keep is a small disaster that repeats.
Step 6: Use the straight ladder — it is not intuitive on purpose
Two runs sit outside the obvious order, and both are worth building deliberately. 2-3-4-5-6 is the highest straight in the game — it beats 10-J-Q-K-A, which is exactly the run your opponent will assume is safe. A-2-3-4-5 is the lowest, because the ace is demoted to a one.
That has a concrete consequence for your low cards. Threes, fours and fives are the weakest singles in the deck and the hardest cards to unload one at a time, which makes them a liability late — but 2-3-4-5-6 turns five of them into the strongest run there is. If you are holding a 2 and a pile of small cards, check for that run before you decide your hand is bad.
The other half of the ladder is knowing that a stronger type beats a weaker one regardless of rank. A modest full house beats the best straight; four of a kind beats the best full house. So when you lead a straight, you are not really asking your opponent to find a better straight — you are asking them to find anything five cards long that outranks it, which is a much easier question than beginners assume. Lead your straights early, while they still do work; hold the four of a kind for when the hand is close.
Step 7: Close it out — the endgame is about sizes, not ranks
When your opponent is down to their last few cards, the rules hand you a very concrete weapon: a play must be answered with the same number of cards. If they hold one card, they cannot answer a pair at all. Not with a better pair, not with anything — their only legal action is to pass.
So when the opponent is on one card, lead two or more. They pass, the trick ends, and you lead again — and you can keep doing that for as long as your multi-card plays last. That is often the difference between winning and handing them the single they were waiting for.
The mirror image applies to you. If you are down to your last card and it is not a certain winner, you need the lead, not a good card. Plan two turns earlier: keep a lead-worthy play in hand so that your last card goes down on your own terms rather than as a desperate answer.
Finally, watch the shape of your own last four or five cards. Two pairs is two turns; a pair and two singles is three; four singles is four. Late in the hand, breaking a pair to win one exchange can quietly add a turn to your own race — and that extra turn is usually the game.
Step 8: Play the computer — and know what it is doing
The Paofuqi Big Two opponent plays with the same information you have: its own thirteen cards and the count of yours. It cannot see your hand, and the engine that decides what is legal is the same one that judges your plays.
When it follows, it plays the cheapest combination that beats the table, and it deliberately holds back a lone 2 unless that 2 is the only thing that works. When it leads, it prefers the play that sheds the most cards. Those two habits are worth copying — they are the same two principles from the sections above.
On Novice it also plays loosely on purpose: it will sometimes decline a trick it could win, and sometimes leads something other than its best option. That is a real if weak tactic, and it makes the game less relentless while you are learning. On Normal it stops doing both.
The most useful practice routine is not playing more hands; it is pausing on the first turn of each hand to count how many plays your thirteen cards need, then checking at the end whether you actually got there. That number, not the cards, is what you are competing on.
Play Banqi free