Poker Hand Reading Exercises: 10 Drills and How to Score Them

Playing more poker is a slow way to get better at reading hands. In a normal session you commit to a real read on a handful of hands, you see the cards on fewer, and the ones you do see are a skewed sample: hands that went to showdown, which leaves out every bluff that took the pot. You can play for months without finding out which of your reads were wrong.

The drills below do three things a normal session doesn't. They make you commit to a range before the answer arrives, they get you the answer as often as possible, and they give you a way to score the result. Scoring comes first, because without it the ten drills are just more practice with no measure of progress.

How to score a read

A lot of hand reading material comes without an answer key, on the grounds that there's no single right range. There isn't, but a read can still be scored. It takes two numbers.

• Hit: was the hand they showed inside the range you committed to? Yes or no.
• Width: how many combinations your range held. Count it, or estimate it from the share of hands.

Either number on its own can be gamed. Say "any two cards" every time and you hit 100% of the time with a 1,326-combo range that tells you nothing. Name one exact hand every time and your ranges are as narrow as they get, but you miss most of them. A good read is as narrow as you can make it while still containing the hand most of the time.

The target I use is a hit rate between 70% and 85%, with the average width coming down from month to month. Above 85% you're almost certainly padding ranges with long shots. Below 70% you're cutting hands that belong. The exact band matters less than the trend: hit rate steady, width falling.

If you want one number, score each read in bits: log2(1,326 ÷ combos) when the hand is in your range, zero when it isn't. Each bit is one halving of the deck.

Range width	Share of all hands	Bits when it hits
663 combos	50%	1.00
265	20%	2.32
133	10%	3.32
66	5%	4.33
33	2.5%	5.33
6 (one pocket pair)	0.5%	7.79

The averages show why both numbers count. A player who always names a 663-combo range and always hits averages 1.00 bit a read. One who names 66 combos and hits three times in four averages 3.25. One who names a single pocket pair and hits one time in four averages 1.95, well behind the 66-combo reader despite much tighter ranges.

For quick widths, convert share to combos: 5% of hands is about 66 combos, 10% about 133, 20% about 265, 30% about 398, 50% about 663.

Watch which hands you're scoring. If you grade only hands that reached showdown, you grade the value bets and the calls, never the bluffs that won uncontested. Your showdown hit rate flatters you on the lines that cost the most: big bets that got a fold. Drills 9 and 10 get around this because they show you the cards in every hand.

A practice log needs one line per read:

Date	Type	Spot	Range committed	Combos	Hit?	Actual
11/03	TAG	CO open, bets K♦9♣4♠, checks turn; I hold A♠K♣	KQ, KJ, KTs, QQ–TT, A9s	38	Yes	K♠J♦

Fifty lines is a small sample. It pins your hit rate down to about six points either way, which is enough to tell 60% from 80% but not 75% from 80%.

At the table

One rule covers everything here: say nothing and type nothing during a hand. Most rooms, and the TDA rules many tournaments use, restrict phone use while you're in a hand, and talking about a hand in progress breaks the one-player-to-a-hand rule. These drills happen in your head and get written down after the hand is over.

1. Read one player in each hand you fold

A folded hand is free study time. Pick one player who's still in and run their range as if you were facing them: what the preflop action left, what the flop bet removed, what the turn check means. Commit to a range at each street, silently. If the hand reaches showdown, check it. Stick to one player per hand. Trying to read three at once leaves you with a vague read on all of them.

2. Name the hand before the cards turn over

When a river bet is called, there are a couple of seconds before the hands are tabled. Name to yourself the exact hand each player will show. Then score on two levels: the hand class (top pair, a missed draw, a set) and the exact holding. Most players find they get the class right far more often than the kicker, which tells them where the work is.

3. Commit one range an orbit, and write it down after the hand

Once an orbit, on a hand you're playing, fix your read on the turn in shorthand ("TT–QQ, AQ, KQ, JTs") and hold it. When the hand is over, write it down with the result, between hands if your room allows it, or batch the orbit's reads at the next break. Writing matters because memory edits itself: the reads you remember are better than the reads you made. Score the log once a week.

4. Type every player, and hold the label loosely

You can't read a range until you know whose range it is, so give everyone a provisional type early: nit, TAG, reg, LAG, fish, calling station or aggressive fish. Then treat the label as a guess, because short samples mislead.

Here is by how much. In one six-handed orbit, a TAG who plays 22% of hands enters no pots at all 22.5% of the time, and a nit who plays 13% sits out the whole orbit 43% of the time. A fish who plays half their hands will enter one pot or none about 11% of the time, looking every bit as tight as the TAG. Hands-played counts settle slowly: for a player who enters a quarter of pots, the standard error on that number is nearly 10 points after 20 hands, 4.3 after 100 and 2.5 after 300.

So weight early evidence by how much it says. An open-limp, a 3-bet with a weak hand, or a showdown no solid player would reach that way tells you more than an orbit of folds. Pick the label from those, then let the counts confirm or overturn it as the sample grows. When a player's action contradicts the label, re-type them rather than explaining the action away.

5. Count one big pot a session

Pick the biggest decision you face and count it properly, afterward if you can't do it in time: how many combinations beat you, how many you beat, and what price you were getting. At the table the shortcuts in drill 8 are enough. Doing the full count once a session is how the shortcuts get fast.

Away from the table

6. Replay hands with the result hidden

Take a hand history, yours or a posted one, and read it street by street with the showdown covered. Write a range at each street before you look at the next action, then uncover the result and score it. Forum threads where the poster reveals the hand later work well, with one bias to know about: people post unusual hands, so the answers skew toward unusual holdings.

Your own online histories show only hands that reached showdown. A few sites have put every player's hole cards into downloadable histories after a delay. Ignition, an offshore site, has done this. It removes the showdown bias completely. Check the current policy, and whether the site is legal where you live, before you rely on it.

7. The narrator drill

You need a partner. One of you holds a hand history with the hole cards and reads the action aloud, street by street. The other names the range after each action and says which hands came out and why. The narrator scores the final range against the real hand, then you swap. The reasons are what you're practicing. "He probably has a pair" doesn't count. "He'd have raised the flop with a set, so it's top pair or a draw" does.

8. Combination drills

Set yourself a board, a hand and an opponent's range, and count against the clock. Three facts carry most of it. An unpaired hand has 16 combos and a pocket pair has 6. Every card you can see removes some: an unpaired hand with one of its cards showing drops to 12, and with one of each rank showing to 9. A pocket pair drops 6, 3, 1, 0 as its cards appear. Three graded sets with answers are below.

9. Watch streams with the hole cards covered

TV and streamed poker show the hole cards on screen. Cash-game streams read them with RFID cards, which is why they broadcast on a delay: so nobody at the table can be fed the cards. The graphic makes the footage useless for practice until you cover it. Put a sticky note over the hole-card graphics, or size the window so they're cut off, and mute the commentary. Pause before each big action, commit to a range, and uncover the cards when the hand ends.

This is the best free source of feedback there is, because you see the cards in every hand, including the bluffs that never reach showdown.

10. Use a trainer that grades every read

Soul Read was built around this drill. You paint a range on a 13×13 grid, the app shows the cards every hand, and it grades the read A+ to F on the same two inputs as above: whether the range contained the hand, and how tight it was. The free browser demo runs the real app: a short tutorial, then three hands.

Combo drills, with answers

Work each set before you open the answers. Pass mark for each set: every answer right, in under ten seconds a question.

Set 1: card removal

1. You hold A♠. How many combos of ace-king can your opponent have?
2. You hold A♠ and the board shows K♦. How many now?
3. The board shows K♦. How many combos of pocket kings are left? And if you also hold K♣?
4. The flop is 8♠5♦2♣. If your opponent plays every pocket pair, how many set combos can they have?
5. The flop is J♥8♥3♣. How many two-heart combos are possible if you hold no heart? If you hold A♥?
6. A player opens 22% of hands. About how many combos is that?

Answers to set 1

1. 12: three aces times four kings.
2. 9: three times three.
3. 3, then 1.
4. 9: three each of 88, 55 and 22.
5. 55 (eleven hearts left, choose two), then 45 (ten left).
6. About 292: 1,326 × 0.22.

Set 2: counting inside a range

A TAG opens from the cutoff with the range below, the one Soul Read's TAG players use: 270 combos. You hold A♠K♣ and the flop is K♦9♣4♠.

AAAKsAQsAJsATsA9sA8sA7sA6sA5sA4sA3sA2sAKoKKKQsKJsKTsK9sK8sK7sK6sK5sK4sK3sK2sAQoKQoQQQJsQTsQ9sQ8sQ7sQ6sQ5sQ4sQ3sQ2sAJoKJoQJoJJJTsJ9sJ8sJ7sJ6sJ5sJ4sJ3sJ2sAToKToQToJToTTT9sT8sT7sT6sT5sT4sT3sT2sA9oK9oQ9oJ9oT9o9998s97s96s95s94s93s92sA8oK8oQ8oJ8oT8o98o8887s86s85s84s83s82sA7oK7oQ7oJ7oT7o97o87o7776s75s74s73s72sA6oK6oQ6oJ6oT6o96o86o76o6665s64s63s62sA5oK5oQ5oJ5oT5o95o85o75o65o5554s53s52sA4oK4oQ4oJ4oT4o94o84o74o64o54o4443s42sA3oK3oQ3oJ3oT3o93o83o73o63o53o43o3332sA2oK2oQ2oJ2oT2o92o82o72o62o52o42o32o22

raise (270 combos)

1. How many of their combos are still possible?
2. How many beat you right now?
3. How many tie?
4. How many are top pair with a worse kicker?
5. How many have no pair at all?

Answers to set 2

1. 200. Your two cards and the flop remove 70.
2. 12: AA (3), KK (1), 99 (3), 44 (3), K9s (2).
3. 6: AK, one suited and five offsuit.
4. 20: KQ (8), KJ (8), KTs (2), K8s (2). KTo isn't in the range.
5. 88. The other 74 are smaller pairs: QQ–TT, 88–55, 33, 22, and the suited nines and fours.

Before anyone bets, you're ahead of 182 of the 200 combos and behind 12. The betting that follows decides how many of those 182 are still with you by the river.

Set 3: the price and the bluffs

1. The river pot is $60 and your opponent bets $40. How often does your call need to win?
2. Your opponent's river value range is 15 combos and they bet the size of the pot. How many bluff combos make your bluff catcher break even?
3. They bet three-quarters of the pot with 10 value combos and 2 bluffs. Call or fold a hand that beats only the bluffs?

Answers to set 3

1. 28.6%: you call $40 to win a $100 pot, so 40 ÷ 140.
2. 7.5. A pot-size bet needs a third of their betting range to be bluffs: 7.5 of 22.5 combos.
3. Fold. Bluffs are 2 of 12 combos (16.7%), and a three-quarter-pot bet needs 30%.

Drills by player type

Combos and pot odds work the same against everyone. Ranges don't. A solver chart can score your read against a chart; the drill that matters at the table scores it against the player, because the most common reading mistake is the right range for the wrong opponent.

So take one action and run it against all seven types. Here is one with an answer key: the rules Soul Read's opponents follow on the turn when they face a bet from a player who looks solid. You bet, they call. Which of these hands can still be in their range?

Type	Middle pair, half-pot bet	Middle pair, pot-size bet	Bottom pair, half-pot bet	Flush draw, pot-size bet
Nit	Folds	Folds	Folds	Folds
TAG	Calls	Folds	Folds	Folds
Reg	Calls 75%	Calls 45%	Calls 45%	Folds
LAG	Calls	Calls	Calls	Calls
Fish	Calls	Folds	Folds	Folds
Calling station	Calls	Calls	Calls	Folds
Aggressive fish	Calls	Calls	Calls	Calls

"Middle pair" here includes top pair with a weak kicker. Two things come out of the table. A nit's turn call of a half-pot bet rules out middle pair entirely, while the same call from a LAG or an aggressive fish keeps everything down to bottom pair. And bet size moves the range as much as the type does: the TAG and the fish call half pot with middle pair but fold it to a pot-size bet, so when they call the bigger bet, middle pair is gone. Against a bettor they see as tight, the app's players fold one size sooner; against a LAG or aggressive fish they call one size later.

What to drill against each type:

• Nit: folding. Write the range a nit raises the turn with and count how much of it an overpair beats. Usually very little.
• TAG: river bluff catching. TAGs bluff the river less than they should, so drill folding bluff catchers when the price is close.
• Reg: the price table. Against a balanced player, the bet size tells you how many bluffs to expect; set 3 is the drill.
• LAG: both halves of a polarized range. Count the value, count the bluffs, compare with the price.
• Fish: splitting the range by how they entered. Their raises are a tight top slice; their limps are everything else.
• Calling station: your own value range. List the worse hands that call three streets, since that's how you get paid.
• Aggressive fish: calling down. Count their bluffs on each street; one pair goes up in value.

A progression with pass marks

Five stages. Move on when you pass. Drop back when your log says you've slipped.

1. Counting. All three combo sets, every answer right in under ten seconds each, three sessions running. Then build your own boards.
2. Starting ranges. Given a type and a seat, write the opening range and estimate its size. Pass when your size estimate is within five percentage points ten times in a row. Check against a grid that counts combos.
3. One street. Flop reads only: commit after the flop action. Pass at 50 logged reads with a hit rate between 70% and 85%.
4. Whole hands. Commit after every street and score the river read. Pass at 50 hands inside the band, with a lower average river width than your first 50.
5. By type. Split the log by player type. Pass when every type you've logged 20 or more reads against is inside the band. The type you miss most against is next month's focus.

A week that works

• Every session: drills 1, 2 and 4. They happen in hands you've already folded.
• Every session you play seriously: drill 3, once an orbit.
• Twice a week, ten minutes: drill 9 or 10.
• Once a week, twenty minutes: a combo set or a replay, then score the log.
• Once a month: total the log by player type and pick the next focus.

For a whole hand read from the first action to the river, see how to put someone on a range. The hand reading cheat sheet puts the counting rules, the price table and the seven types on one printable sheet.