Two Letter Scrabble Words: The Must-Know List

Two letter Scrabble words are short, valid words like QI, ZA, JO, XU, and AA that let you score with awkward tiles and open up the board. They're the single most useful thing a beginner can learn, and every strong player knows them cold. There are 107 of them in the North American dictionary and 127 in the international one. This guide gives you the full list, tells you which to learn first, and shows you how a couple of two-letter words can turn one turn into five scoring words at once.

What Are the Two Letter Scrabble Words?

They're exactly what they sound like: legal words that are only two letters long. Some are obvious (AS, IT, ON, TO), and some look like typos until you check (QI, ZA, XU, HM). Every one of them is a real, playable word in the official Scrabble dictionaries, and that mix of familiar and strange is what makes the list so worth learning. Here's the full set from the North American list:

That's all 107 of them, and every single one is also legal in Collins, so nothing on that grid can get challenged off the board anywhere in the world. Which of those two lists you're actually playing under is the thing to pin down, and our Scrabble dictionary guide walks through how to tell. Not sure a combo is legal in your edition? Run it through the word finder before you commit.

How Many Two Letter Words Are There?

Here's where it pays to know your dictionary, because the count isn't the same everywhere. According to the NASPA Word List, curated by the North American Scrabble Players Association and used for tournament play in the United States and Canada, there are 107 valid two-letter words.

Play internationally and the number goes up. The Collins Scrabble Words list from Collins Dictionary, the standard almost everywhere outside North America, contains 127 two-letter words, and the page says so in as many words: "You will really help your game by learning these 127 playable two-letter words." That's exactly 20 more than the NASPA list. So the real answer to "how many" is 107 or 127, depending on which official source your game follows. Agree on the dictionary before the first tile goes down and you'll save yourself a lot of arguments. Our Scrabble dictionary guide explains the difference in full.

One thing that makes this easier than it sounds: the 107 is a clean subset of the 127. Nothing is legal in North America but illegal internationally. Learn the NASPA list and you've learned the safe core of both.

A word of warning if you also play Words with Friends, because that's a third dictionary and it trips people up. WWF doesn't use either list. It runs on its own lexicon derived from ENABLE, and Zynga has never published it, so there's no official page to check the way there is for NASPA and Collins. You'll see confident two-letter counts for WWF quoted around the web. Most trace back to a Scrabble edition that's years out of date, and none of them cite a source, so we're not repeating a number here. What's safe to say is that the overlap is large but not total, and the only reliable check is the game itself, which will reject an invalid play before it costs you anything. If you move between the two games, our Words with Friends vs Scrabble comparison covers where else the rules diverge.

Which Two Letter Words Does Collins Allow That NASPA Doesn't?

Exactly twenty, and NASPA publishes the list itself. Its How Collins differs page names the "20 extra two letter words in CSW24" as:

These are the ones that cost people games. Play ZE or UR in a North American club event and it comes straight back off the board. Play them in London or Melbourne and they're fine. Two of them are worth real money: ZE and ZO both score 11 base points, the same as QI and ZA, which means Collins players have four ways to dump a Z or Q in two squares instead of two.

Which Letters Have No Two Letter Word at All?

Two of them, and knowing which changes how you hold your rack. The same NASPA page that lists the Collins extras spells it out: adding CH means "the letter C therefore now has a two letter word, CH, but there are still no two letter words containing a V."

So the position is this. In North American play under the NASPA list, neither C nor V appears in any legal two-letter word. Switch to Collins and C gets exactly one, CH. V gets nothing, anywhere, in either dictionary.

That single fact explains more about the V than any tile-value chart does. Every other letter in the bag has at least one two-square escape hatch. Stuck with a J you can play JO. Stuck with an X you have AX, EX, XI and XU. The V has none of that, which is why it sits on racks for turn after turn and why experienced players treat it as something to spend rather than something to keep.

Three practical consequences worth carrying into a game.

• Don't save a V for a better spot. There often isn't one. The shortest common words that use it need three or four squares and a helpful board, so the longer you hold it the more likely you are to still be holding it at the end.
• Two Vs is a real problem, not bad luck. If you draw both, plan to shed one even at a mediocre score. The alternative is carrying it to the endgame, where unplayed tiles get deducted from your total and handed to your opponent.
• The C is only awkward in North America. If you play Collins, CH gives you the same escape hatch every other letter has. If you play NASPA, treat the C a bit like the V, though it's far easier to place in three squares.

And here's the flip side, which is the more useful half. Because a V can never form a two-letter word, a square directly beside an existing tile is a square a V can never legally occupy in a parallel play. When you're reading a tight board and wondering what your opponent might be able to do in a gap, the V is one thing you can rule out.

None of this makes the V unplayable. It makes it a tile with a shorter shelf life than the rest, and the two-letter list is exactly what tells you so.

Which Ones Should You Learn First?

Don't try to swallow all 107 at once. Start with the words that solve real problems at the board. Two groups matter most.

The high-value tile dumps. These let you offload your hardest tiles for big points:

Word	Base points	Why it earns its place
QI	11	Play the Q with no U needed
ZA	11	Dumps the Z with just a common A
XI / XU	9	Two easy homes for the X
AX / OX / EX	9	X plus a vowel, hooks almost anywhere
JO	9	The cleanest way to unload a J
KA / KI	6	Gets rid of a stubborn K

The vowel dumps. When your rack is all vowels and nothing works, these save the turn: AA, AE, AI, OE, OI, and AB. A short valid play always beats passing or swapping when you don't have to.

One caveat on how you drill them, because the research here is counterintuitive. Moxley, Ericsson and Tuffiash surveyed competitive Scrabble players for a 2017 paper in Psychological Research and found that analysing your own games and practising anagrams were significant positive predictors of Scrabble skill, while "the hours of studying vocabulary alone was a negative predictor of SCRABBLE skill." Casual games outside competitive play didn't correlate with rating improvement at all. So don't just stare at the list. Drill it as anagram practice, then go back over a real game and find the two-letter plays you missed. That's the version of study that actually moves your rating.

Why Do Two Letter Words Matter So Much?

Because they punch far above their size. A two-letter word looks like it can only score a handful of points, but the real value is threefold. First, they let you use your ugliest tiles: QI and ZA are worth 11 base points each, the highest of any two-letter word on the NASPA list, and landing that Q or Z on a premium square can push a tiny word past 20 or 30 points. (Collins players get two more at 11, ZE and ZO.)

Second, they open closed boards. When tiles are packed tight and there's no room for a long word, a well-placed two-letter word creates a fresh hook for your next turn. And third, they're the key to parallel plays, which is where the real scoring lives. That's worth its own section. If you want the wider picture, our guide to winning at Scrabble puts it all together.

How Do You Actually Score a Two Letter Word?

Every word your play forms gets scored separately, and a premium square you cover counts in full for each of them. That second half is the rule that makes short words dangerous, and it's the one most casual players have never been told.

NASPA's Official Tournament Rules, Player Edition, still the current edition, sets it out in two sentences: "When the main word makes additional cross-words, they are each considered to be part of the play for the purposes of scoring the play and determining its validity. The score of the play is the sum of the scores of its words, plus a 50-point 'bingo' bonus if all seven tiles are used." Then the per-word method: "The score of a word is the sum of the scores of its letters, each multiplied according to any letter bonus squares newly covered, then finally multiplied according to any word bonus squares newly covered."

Three things fall out of that, and each one changes how you play the list.

One tile can score twice. Say there's an A sitting to the right of an empty double letter square, and another A directly below that same square. Drop a single Z on it and you've made ZA across and ZA down. The Z doubles to 20 in both, so each word scores 21 and the play totals 42. One tile, one square, 42 points, because the premium is applied inside every word that runs through it. That's the whole engine behind parallel play, and it's why the next section is worth reading slowly.

Word squares multiply the entire word, not just your tile. Land ZA on a double word square and it's not 11 plus a bit. It's (10 + 1) doubled, so 22. On a triple word square that same two-tile play is 33 points. A two-letter word sitting on a triple can out-score a sloppy seven-letter one.

"Newly covered" is doing a lot of work. Read that phrase again, because it's the trap. Premium squares only pay on the turn a tile first lands on them. Play through a double letter square your opponent covered three turns ago and you get nothing extra for it. So when you're hunting a spot for the Q or the Z, you're not looking for premium squares. You're looking for empty ones.

Worth knowing what the tiles are actually worth before you do any of this arithmetic:

Play	Where it lands	Score
ZA	No premium	11
ZA	Z on a double letter	21
ZA	On a double word square	22
ZA	On a triple word square	33
ZA + ZA	One Z on a double letter, both directions	42

Two footnotes that catch people out. The rules note that "the centre square counts as a double word square," so an opening two-letter play is automatically doubled. And blanks are a trap here: the rules state plainly that "blank tiles always score zero points," so a blank standing in for your Z gives you a legal ZA worth 1 point rather than 11. Useful for opening a lane, useless for scoring, which is why blanks are worth saving. Our guide to blank tile strategy covers when to spend one.

One last line in that quote deserves attention, because it isn't about scoring at all. Cross-words count as part of the play "for the purposes of scoring the play and determining its validity." That's the rule behind the warning further down: get one pair wrong and the whole play is invalid, not just the bad word.

How Do You Use Them for Parallel Plays?

A parallel play is when you lay a word directly alongside a word already on the board, so that every pair of touching letters forms a valid two-letter word. Say STONE sits on the board and you play HOMES right beneath it, letter under letter. You don't just score HOMES. You also score SH, TO, OM, NE, and ES running down the columns, six words from one move.

Check that stack for a second, because it shows you the trap. Every column has to work. Slide HAZE under STONE instead and the middle column gives you OZ, which isn't valid in either dictionary, so the whole play is challengeable. One dead pair kills the move.

That's why the list matters so much. A single parallel placement can score your main word plus three, four, or even five short words at once, routinely turning tiles that would score 15 or 20 points into a 40 to 70 point turn. But it only works if every one of those two-letter combos is legal, which is exactly why strong players know them by heart. When you know them, you can scan a crowded board and spot the parallel in seconds.

And that scanning ability is measurable, which is a genuinely odd thing to be able to say about a board game. Hargreaves, Pexman, Zdrazilova and Sargious tested competitive Scrabble players against age-matched controls and published the results in Memory & Cognition in 2012. On general cognitive ability the two groups were no different. But the players showed what the researchers called vertical fluency: they had markedly less trouble judging whether a word was real when it was printed vertically instead of horizontally, which is exactly the direction you read a parallel play down a Scrabble column. The paper's conclusion is that "visual word recognition is shaped by experience, and that with experience there are efficiencies to be had even in the adult word recognition system." Learning these words doesn't just give you more options. It changes how fast you see them.

When Is Playing a Two Letter Word the Wrong Move?

Everything above argues for these words, and it should. But a guide that only tells you to play them will make you worse at some point, because a two-letter word has costs that don't show up in the score you just wrote down. Three situations are worth knowing.

When it opens a lane you can't use. NASPA's glossary keeps the distinction plain: an open board is "a board where there are many places for new words may be added," and a closed board is one "where there are few places where new words may be added." Almost every two-letter word you play nudges the board toward open. That's fine when your rack is good, because openness rewards whoever can use it. It's a gift when your rack is junk. If you're sitting on four vowels and a V while your opponent just drew from a bag full of blanks and S tiles, prying the board open for 9 points is not a bargain.

The move you're looking for in that spot has a name too. NASPA defines blocking as making "a play that prevents the opponent from making an important or high-scoring play." Sometimes the best two-letter word on the board is the one that seals a triple lane for six points rather than the one that scores eleven and unseals another.

When the leave is worse than the score. Your leave is what stays on your rack afterward, which NASPA defines as the tiles remaining "after a play is made and before any replacement tiles are drawn." Here's the trap. The score is printed on the board where you can admire it, and the leave sits in your hand where nobody grades it. So beginners optimise the number they can see. Dumping ZA for 11 feels great, and it can still be the wrong play if it strands you holding Q, U, V and two I tiles going into a closed board.

Except in the endgame, where the maths flips hard. This is where twos stop being handy and start being decisive, and the official arithmetic is bigger than most casual players realise. Under the NASPA tournament rules, rule V.A says the game ends when one player "has successfully played all of his/her tiles, and the bag is empty." Rule V.G then says that if you go out, you "increase your score by double the value of your opponent's unplayed tiles." Double. And if nobody goes out, each player's score is instead reduced by the value of their own leftovers, so being stuck costs you twice over.

Run that on a real board. Your opponent is holding the Q and can't place it. You have two tiles left and a legal two-letter word. That play might score 5, and then swing another 20 on top from doubling their stranded Q. A word you'd ignore in the midgame becomes the best move in the game.

Worth flagging that this doubling is the tournament rule. Plenty of kitchen-table games score the ending differently, so check which version you're playing before you count on it.

Knowing which tiles are still out there is what makes that judgement possible, and NASPA has a term for that as well: tile tracking, meaning "to count the tiles played so as to determine which tiles are as yet unplayed." You don't need to track a whole game to benefit. Just start noticing the Q and the blanks.

One last rule that quietly rewards knowing your twos. Rule V.B, the six-zero rule, lets the game end after "a sixth successive zero-scoring play from passes, exchanges, challenges, or illegal plays." Passing your way out of a bad rack isn't free. Often the cheapest way to avoid a scoreless turn is exactly the small word you'd otherwise dismiss.

So the habit to build is one question, asked before you place the tiles. Not just what does this score, but what does it leave me and what does it open for them.

Which Two Letter Words Look Valid But Aren't?

Knowing what's off the list saves you as many points as knowing what's on it. A challenged word costs you the turn, and in a parallel play one bad pair kills the whole move. These are the ones that catch people.

V has nothing at all. Look back at the grid above and you'll notice something: not a single word starts with or ends in V. It's the only letter in the alphabet with no two-letter word in either dictionary. VA, VE, VI, and VO all feel like they should work. None of them do. If you're holding a V, plan on a longer word, because there is no two-tile escape hatch.

C has nothing in North America. Same story on the NASPA list: no CA, no CO, no CE. Collins players get exactly one, CH, which is on the twenty-word list above. So a C is nearly as awkward as a V, and more so if you're playing NWL.

The Z traps. ZA is valid everywhere and it's the one people know. But AZ, EZ, IZ, UZ, and OZ are not words in either dictionary, and OZ is the one that burns people most often because it looks so obviously fine. Collins adds ZE and ZO. NASPA does not. If you're playing in North America, ZA is your only two-letter Z.

IQ is not a word. It reads like it must be, it would be worth 11 points, and it is not on either list. QI is the Q word you want, and it's the only two-letter Q play in both dictionaries.

A quick way to hold all of that: think of V and C as the two letters with no short exit, and remember that Q and Z each have exactly one two-letter home in NWL, QI and ZA. Everything else your brain suggests around those tiles is probably a challenge waiting to happen.

The broader habit worth building is checking rather than assuming. If a pair feels plausible but you can't picture it on the grid above, it's usually not there. Run it through the word finder before a game, not during one.

What Happens If Someone Challenges Your Two Letter Word?

Depends entirely on which challenge rule you're playing under, and this is the bit that flips the whole calculation. The section above treated a challenge as pure downside, something to avoid. That's only true if you're wrong. If you're right, a challenge is one of the best things that can happen to you, and knowing the list is what turns other people's doubt into your points.

Two rule sets cover almost all serious play.

Double challenge. This is the standard in NASPA-sanctioned play in the United States and Canada, the same territory where the 107-word list applies. Challenge a play, and if the word is good, you forfeit your next turn. That's a brutal price. It means that against anyone who knows the rules, playing a legal-but-weird two-letter word is close to free: either it stands, or someone hands you an entire turn for doubting it.

Points per word. Collins play works differently, and NASPA spells it out on its own How Collins differs page: "If your opponent makes a play, you challenge, and their play is good, instead of losing your turn, the opponent gains points. Typically, this is 5 or 10 points per word challenged." No lost turn, just a points transfer, which is why Collins players challenge far more readily than North American ones.

Now read those four words again: per word challenged. That's where two-letter words quietly change the maths.

Go back to the parallel play from earlier, the one where HOMES under STONE scored six words at once. Under a points-per-word rule, if someone challenges that play and every word holds, they're not paying a flat 5 or 10. They're paying it six times over. A challenge that fails against a big parallel play can hand you 30 or 60 points without you touching a tile. The same knowledge that lets you build the play also makes the play expensive to doubt.

And the reverse is the warning. A challenge covers the whole play, not one word in it, which is why a single dead pair takes the entire move off the board no matter how good the other five words were. You don't get partial credit. That's the real reason the false friends in the previous section matter: OZ isn't just a wasted word, it's a wasted turn and everything you built around it.

Three things worth doing with that:

• Ask which rule is in force before you sit down. Same as agreeing the dictionary. Double challenge makes you bolder, points-per-word makes your opponent bolder, and playing one while assuming the other is how people lose games they understood.
• Stop apologising for the strange ones. If you know QI, XU or HM is legal, play it without the nervous glance. The glance is what invites the challenge, and under double challenge the challenge is a gift.
• Never bluff a two-letter word you haven't verified. Under double challenge you'd lose the turn plus the whole parallel play. Under points-per-word you'd lose it anyway. There's no rule set where guessing pays, which is the entire argument for learning the 107.

Casual kitchen-table games usually run no penalty at all, where a wrong word just comes back and play carries on. Fine for a Sunday. But it's also why home-game habits travel badly, because it's the one setting where guessing costs you nothing and so nobody ever learns the list.

When Does the Official List Actually Change?

Not often, and never mid-season. The North American list is maintained by the NASPA Dictionary Committee, chaired by John Chew, and each edition gets a fixed switchover date so tournaments aren't playing two rulebooks at once.

NASPA's own word list page records the recent run: NWL2018 was released electronically in December 2018 and took effect on March 1, 2019. NWL2020 came out in October 2020 and took effect on January 6, 2021. NWL2023, described by NASPA as "the official lexicon governing competitive play in the United States and Canada," was released in October 2023 and took effect on February 29, 2024. That's the edition in force now.

What this means for you is simple. The two-letter list is stable enough to be worth memorising, and roughly every three years you should check whether a handful of new ones showed up. Anything you learned five years ago is still legal. You might just be missing a couple.

Collins runs on its own clock, and it's the one most of the world plays. If you're in the UK, Australia, Singapore, or anywhere outside North America, the NASPA schedule above isn't yours. NASPA's page on Collins tracks the other lineage: the list was "first published in 2007 by Collins, and updated in 2012, 2015, 2019, 2021, and 2024."

The switchover dates are worth knowing if you play rated games, because they don't line up with the American ones at all:

• CSW15 took effect on 1 September 2015.
• CSW19 took effect on 1 July 2019.
• CSW21 took effect on 1 December 2021.
• CSW24 took effect on 1 January 2025, and NASPA notes that tournaments already in progress on that date, early birds included, used the new edition immediately.

Collins published the 2024 edition on 14 August 2024, a few months before it became playable, and describes the list as one that "constantly evolves just as language does." That gap between publication and effect is the thing to watch. A word appearing in a new Collins edition doesn't mean you can play it that weekend.

And here's the reassuring part for anyone staring at 127 words. Collins' own announcement of the 2024 edition leads on hooks, modernisms, and long words worth chasing. The two-letter list simply isn't where these updates do their work. Whichever dictionary you play, the twos are close to the most durable thing you can commit to memory, which is exactly why they repay the effort more than any other list on the board.

Do These Words Work in Words With Friends?

Mostly, but not reliably, and the exceptions are the ones that will cost you a turn. If you learn this list and then take it straight into Words With Friends, expect a few rejections you won't see coming.

The reason is structural. Everything above this point is about two lists: the NASPA Word List for North American Scrabble and Collins for everywhere else. Words With Friends uses neither. Zynga maintains its own word list for the game, shaped by what its players ask for rather than by a tournament committee.

So a word that's fine at a Scrabble club can bounce in the app, and something the app happily accepts can be challenged off a tournament board. Neither list is wrong. They're answering different questions, and only one of them is governed by the rules covered further up this page.

Two practical consequences:

• Don't assume a rejection means you misremembered. If a two you're confident about gets refused in the app, the likely explanation is the list, not your memory. Check it against the dictionary you actually play under before you unlearn anything.
• Keep the two vocabularies separate in your head. If you play both, treat them as overlapping rather than identical. The overlap is large enough to be useful and small enough to embarrass you.

None of which makes this list less worth learning. It's the foundation for both games and the great majority of it carries across. Just don't treat a Scrabble two-letter sheet as authoritative inside an app that never agreed to use it. Our guide to Words With Friends versus Scrabble covers the rest of the differences, board and scoring included.

How Many Do You Actually Need to Learn?

Fewer than 107, and the real figure is small enough to be encouraging. Medill Reports Chicago, published by Northwestern University, put this to several experts in a May 2023 piece on memorising the list, and the arithmetic they landed on is the most useful thing written on the subject.

Start by subtracting what you already own. The average English speaker uses about 35 of these words in ordinary conversation, things like AS, IN and SO, which you will never have to sit down and study. That leaves 72.

Then notice how lopsided the remainder is. Of the 107, 67 contain an E, an I, or an O. The list clusters hard around three vowels, so working through "what can follow O" in one sitting covers far more ground than picking words off the grid at random.

Doug Shaw, a maths professor at the University of Northern Iowa who published his own guide to the list, doesn't hedge about how much it matters. He told Medill "it is crucial to the point where I won't play anyone who doesn't have them memorized." Strong words for what turns out to be 72 words of homework.

How Do You Memorize the List?

By testing yourself, not by reading the list again. That sounds like a small distinction. It isn't, and it's the difference between the studying that works and the studying that doesn't.

Dunlosky, Rawson, Marsh, Nathan and Willingham reviewed ten common learning techniques for Psychological Science in the Public Interest in 2013, volume 14, pages 4 to 58, and graded each one. Practice testing and distributed practice, meaning quizzing yourself and spreading sessions out over time, both earned high utility ratings because they "benefit learners of different ages and abilities." Rereading and highlighting landed at the bottom. The authors write that they "do not consistently boost students' performance, so other techniques should be used in their place."

Which quietly explains the odd finding from earlier in this guide. Moxley and colleagues found hours spent studying vocabulary was a negative predictor of Scrabble skill, and that reads like a paradox until you notice it's measuring hours, not method. Staring at a word grid is rereading. It feels productive and it's the low-utility technique. Covering the grid and trying to recall every word that starts with O is practice testing, and it's the high-utility one. Same list, same twenty minutes, very different result.

So build it as a quiz from day one. Cover the grid, pick a letter, write down everything you think is valid, then check. The words you miss are the only ones you actually need to work on, and you'll find that out in about thirty seconds. That self-check is what "quiz, quiz, quiz" means when the experts in the Medill piece say it.

Order matters too. John Chew, chief executive of the North American Scrabble Players Association, recommends starting with the J, Q, X and Z words: AX, EX, JO, OX, QI, XI, XU and ZA. That's eight words covering the four tiles most likely to strand you, so it's the highest return per minute on the whole list. From there work through the vowel dumps, then group the rest by leading letter, which makes the outliers stick. A leads a big cluster while J gives you only JO.

Then space it out. Five words a week that you actively quiz yourself on will beat all 107 crammed in one evening, because distributed practice was the other technique that scored high. And play them. Chester Santos, who won the USA Memory Championship in 2008, suggests hanging the odd ones on a mental journey, and a two-letter word you've actually played in a real game gets that anchor for free. By the time 50 of them are fluent your scores climb in ways you can't trace to any single move. For more quick wins, our Scrabble tips for beginners is a good next read.

Does Natural Talent Decide How Good You Get?

Mostly not, and the study this guide already leans on says so more directly than the bit quoted earlier.

The paper by Jerad Moxley, K. Anders Ericsson and Michael Tuffiash appears twice above, once for the finding that vocabulary study is a negative predictor of skill. But that was a side result. The paper's actual subject is gender differences in Scrabble performance, published in Psychological Research in 2017, and what it found there is the most encouraging thing in this whole guide.

A real gap that stopped being a gap

They surveyed competitive players twice. Study 1 covered 133 people, 85 men and 48 women. Study 2A covered 122, this time 59 men and 63 women.

There was a clear performance difference between the groups, and it wasn't small: d = -0.74 in the first study and d = -0.69 in the second.

Then they controlled for how much people had practised, and the researchers report that "after controlling for practice differences, gender effects became statistically insignificant." The men in the sample had spent "significantly more time studying."

The gap was a practice gap wearing a costume.

What actually moved the needle

The same analysis separates the activities that predict skill from the ones that don't, and the split is sharper than most players would guess.

• Scrabble-specific solitary practice predicted skill, with a standardised coefficient of 0.46. Working on words, on your own, deliberately.
• General vocabulary study went the other way, at -0.45. Learning words in general is not the same activity as learning the words this game rewards.
• Playing was not enough. The "amount of time spent playing SCRABBLE was not" significantly predictive of skill. Nor was tournament play on its own.
• Starting young helped. Earlier starting ages predicted higher skill, and earlier tournament entry predicted improvement.

Their headline number ties it together: "Over half of the variance in SCRABBLE performance was accounted for by measures of starting ages and the amount of different types of practice activities."

Over half. Not from verbal ability, not from a knack for words, but from when you started and what you did with the time.

Which is good news for the list on this page

Three things follow, and they change what you should do with the 107 words above.

First, playing more games will not get you there. That's the finding most people resist, because playing feels like practice and is far more fun. But time at the board didn't predict skill in this sample. Sitting down with the two-letter list did the kind of thing that did.

Second, "I'm not really a word person" is carrying less weight than you think. When over half the variance sits in starting age and practice activities, the story about innate ability has a lot less room to work with than the story most players tell themselves.

Third, the starting-age result is the one honest catch. You can't go back and take up Scrabble at nine, and it would be dishonest to pretend that finding away. What you can change is the other half of the equation, which is the half this article is about.

Two caveats worth holding. This is a survey of competitive players, so it shows what correlates with skill rather than proving what causes it, and practice hours were self-reported, which is a famously generous form of measurement. But the direction is consistent with the Dunlosky review quoted in the section above: what you do beats how long you sit there.

What Should You Study After the Twos?

The threes. And you don't have to hunt for a separate list, because the people who run competitive Scrabble in North America put both lists on one sheet.

NASPA's Youth SCRABBLE Resources page sends players off to download "study sheets with the 2- and 3-letter words, common bingos, and other useful words to learn." Notice what's sitting together in that sentence. The twos and the threes aren't two separate homework assignments. They're one.

The reason is hooks. A three-letter word is very often a two-letter word wearing one extra letter, so once you know AE is good, the useful question stops being "is AE a word" and becomes "what can I hang on it." NASPA's sheets answer that on the same line. The association describes its own two-and-three-letter file as a "List of two- and three-letter words acceptable according to OCTWL2, with additional hook information," and on the page each word carries its hooks right beside it: a single letter where only one works, a numeral where several do. So you're not memorising two lists. You're memorising one list and its exits.

That page also gives the least glamorous advice in the hobby, and it's correct. "If you print these on two sides of a brightly colored paper and then laminate them, the cheat sheets will last for several years." Bright paper because you will put it down somewhere. Laminated because you will spill something on it.

But check the edition before you memorise a word of it.

Go looking for a NASPA study sheet and you can easily land on a PDF hosted on the association's own wiki, titled "Two- and Three-Letter Words." It looks official because it is official. It's also from another era. NASPA's file record shows it was uploaded in April 2012, and the first page of the document says it lists the words acceptable "according to the Official Club and Tournament Word List (2nd edition, published 2006)." The same record notes that no pages on the wiki link to it any more, which is why a search engine is now the only realistic way to reach it.

Three editions of the North American list have shipped since that one, as the section above lays out, and the sheet knows nothing about any of them. A 2006 list will still teach you most of the twos, because the core barely moves. But "most" is the whole problem. The words that changed between editions are exactly the words that draw a challenge.

So read the edition line the way you'd read a use-by date. Open the sheet, find where it names the word list it was built from, and if that isn't the list you actually play under, go and get the current one. The list on this page covers both dictionaries. And if you'd rather think about words by what they earn you than by alphabet, our guide to the highest scoring Scrabble words takes that angle.

What Else Do People Ask?

How many two letter words are valid in Scrabble?

It depends on the dictionary. The NASPA Word List used in North America has 107 valid two-letter words. The Collins Scrabble Words list used internationally has 127, which is 20 more. So the honest answer is 107 or 127 depending on which official word source your game follows. Both include the high-value staples like QI and ZA.

Is QI a valid Scrabble word?

Yes. QI is valid in both the NASPA and Collins dictionaries, and it's one of the most useful words in the game because it lets you play a Q without needing a U. It scores 11 base points before any bonus squares, so landing the Q on a premium square can turn a tiny word into a big score.

What is the highest scoring two letter Scrabble word?

QI and ZA are tied for the highest base score at 11 points each, since both use a 10-point tile (Q or Z) plus a 1-point tile. They're the top two on the NASPA list used in North America. Collins players have two more at 11, ZE and ZO, which are not valid in NASPA play. On a bare board any of them is worth 11, but place that Q or Z on a double or triple letter square and a two-letter word can suddenly score 20 to 30 points or more.

Are two letter words allowed in Scrabble?

Absolutely. Any word in the official dictionary is legal, and that includes every valid two-letter word. There's no minimum length rule beyond the two-tile floor for a play. Two-letter words are not only allowed, they're one of the most important tools skilled players use, especially for scoring several words at once.

Which dictionary should you use for two letter words?

Use whichever your game or club follows. In North America that's the NASPA Word List, curated by the North American Scrabble Players Association. Most of the rest of the world uses Collins Scrabble Words. Casual home games often just use a standard dictionary. Agree on the source before you start so a challenge never turns into an argument.

Word validity varies by edition. The NASPA Word List (used in North America) and Collins Scrabble Words (used internationally) don't share identical lists. Word counts cited here reflect the current editions and can change when the lists are updated. Always verify with the official dictionary for your game. Scoring rules quoted here are from the NASPA Official Tournament Rules, Player Edition effective 20 January 2017, which is still the current edition at the time of writing, though a revision was out for consultation. Casual and app versions may score or challenge differently. Not affiliated with Hasbro or Mattel.