In order to illustrate some concepts from elsewhere in this guide with live, real-game examples, I have decided to play a best of 7 series against the Words With Friends® AI bot. I also just think it would be fun to do so.
I've taken the best of 7 concept from other competitions like the World Series or the NBA Finals. There are also different types of games, and hopefully we can see that by playing through a few different examples. For each move, I will write a few sentences about the state of the board, my thought process, and my final move.
A couple of disclaimers are in order.
With that being said, let's proceed to the summary for Game 1.
18 points. Winning 18-0.
There are, unfortunately, no 7-letter words in our initial rack. However, we can play MOROSE for a score of 18.
The decision to play to the left is more likely to lead to an open game, with longer words.
We could place either the M or the O on the Double-Word square, which is aligned with the TL and TW squares above and below. I am choosing to play the O in this square to reduce the odds of our opponent having an 8-letter word which could hit the TW square. There are fewer 8-letter words which begin or end with an O, vs. an M.
21 points. Winning 39-11.
The WWF AI has played COSS. Since there no two-letter words with C, this move will probably block the action somewhat towards the top-center of the board.
The rack we have drawn contains only a single vowel. This constrains our play and leaves us with a decision to make.
One option would be to simply exchange the rack. However, we have some decent consonants and can probably make a word for 20-30 points.
The highest-scoring option would be to play KHAN, as pictured. However, this leaves us vulnerable to a huge reply from the opponent. They could, with the right letters, hit a TL and TW on the same turn and run up a massive score. This is particularly true since we would be placing a K in the relevant column.
Instead, we can play KOHL for 21 points. It is in a safer place, and it also unloads three consonants. Since we only have one vowel, the A, it would be better to only play consonants.
There is one downside to this move. KOHLS is a valid word, and the S would be on a TL space. But I am proceeding with it in light of its other advantages.
0 points. Winning 39-24.
The WWF AI has played DEL, and scored 13 points in the process.
The AI has also split the left side of the board very distinctly into a top and bottom half. Since there is no letter that can prefix DEL, or MOROSE, there is now effectively a wall to the left of those two words.
There is also a wall where the C in COSS and the K in KOHL meet. No words can join into that area. If things continue as they have gone so far, this looks to be a relatively closed game.
We are still heavier on consonants than we would like to be, and nothing in our rack can make a longer word than five letters. In fact, from a scoring standpoint, one of our best moves nets us only 12 points (playing ENTIA into the DW).
It's a tough decision, but in this case I have decided to swap the entire rack. We cannot gain any kind of points this turn, and any move we make is likely to open the board up significantly for our opponent's next move. Hopefully we can draw something to make a longer word on our next turn.
61 points. Winning 100-36.
The WWF AI has played RUSH for a score of 12 points. Meanwhile, our draw has worked out well for us. We picked up a blank and an S, and the letters can make a 7-letter word -- BRAWEST.
There are actually two places to stick this. One is adjacent to RUSH for 61 points. The other is vertically, to also make KOHLS, for 77 points. On points and elegance, the decision would favor the second option.
We have no idea, however, what our opponent might hold or how they will play it. If we played the 77-point option, we would be sticking our B right in the danger zone between a TL and a TW. With the right letters, our opponent could easily make a 100-point move or more!
After deliberating, I have decided to play the 61-point version, because it is better defensively.
30 points. Winning 130-81.
The WWF AI has played GOX for 45 points. This was an excellent use of the DL square below the E in BRAWEST. The X on that square counts for 16 points in EX, and another 16 points in GOX, for 32 points total. Putting a high-value letter on a premium square, such that it plays in two words at once, is rarely a bad choice.
Our letters are again unbalanced, with only two vowels. On the other hand, we have drawn a Z. However, we are still in a tough spot -- using these letters alone, there is no word longer then four letters that we can play.
There is also no obvious place on the board to maximize the value from our Z. The only two-letter word that Z plays with is A, and the lone A on this board is well-covered. After looking at the available words, along with this board, however, we find that the word ADZ can play off the end of RUSH, into a DW square. This gives us 30 points without exposing too much to the opponent (there is a chance he could bonus into the D from the TW at top right).
30 points is more than enough to warrant playing (as opposed to exchanging), and that is what we do here.
29 points. Winning 159-102.
The WWF AI has played FUTZ for 21 points. As hoped for, extending the board with ADZ did not turn out to be too dangerous for us.
We still have an awkward assortment of letters, but at least the vowel-consonant distribution is evening out. We can also expect that the game will be very closed as we continue. The C in COSS, the Z in FUTZ, the EX (which has no 1-letter suffix), and the DEL/MOROSE to the left, all block off significant parts of the board. It would be almost impossible to play a 7-letter word on this board, even if a clean one was sitting right on our rack.
Fortunately, we are also winning. Not by a ton, but by enough that we don't need to take wild chances in opening up the board right now.
It would be nice to play a word from the DW down to the TL by KOHL, but there is no 5-6 letter ending in S that we can play. We also have nothing that makes a 2-letter word with the F in FUTZ, which could be extended to the TW at top-right.
After searching for awhile, the best move I can find is to play EVES for 29 points. This is a decent score for a move, and doesn't give up too much in terms of position. While we open up the TW squares in bottom-left, we also occupy the TL and leave only three empty squares to build a word.
NOTE: Looking at this later, it would have also been possible to play THESE, hitting the TW, and making EF into FUTZ. This would have scored 32 points and been a better strategic play. But I will leave my initial analysis above as is.
56 points. Winning 215-141.
The WWF AI made a decent move for 39 points, prepending BR to the EVES that we just played, to make BREVES.
Since we now have letters which can make a 2-letter word with the F in FUTZ (OF and EF both work), that is where our attention should be directed. Not only are we likely to make our best score by using the TW square (1,12), but we cannot afford to risk a move here by our opponent.
Fortunately, there are a decent number of 5-letter words we can make which end in E or O. Our best choice puts the H on the TL square, for it to be multiplied by nine. We will play OUPHE here to do so. This will get us a very solid score of 56 points.
One thing to be aware of is that this might speed the action up for the next couple of turns. As explained earlier, the board has been pretty locked down. However, our move will let our opponent play across the top three rows of the board without obstruction. It is possible that we in turn could play a long word -- either adjacent to our opponent, or down the right side of the board.
16 points. Winning 231-175.
The WWF AI has played
If we had an S, our opponent's play would open up the TW and the right-column of the board for us. At first glance, it might appear that our opponent took a real chance here.
However, if we look more closely at the board, all five S tiles have now been played. A glance at the Tile Bag confirms this. Since no other letter besides S can be played at the end of EQUAL, our opponent's move actually closes the upper-right part of the board. Only a blank can open that area of the board at this point. Without a blank, neither player can touch the TW square in that area, and neither player is likely to give the other player an opportunity to do so.
As for our own letters, they do not leave us with a long word to play. At this point, we have two options. One would be to use the V in BREVES and try for a TW, which could be done with VAIN or VITA. This would give a 27-point score, which is decent.
However, we are currently winning. Our opponent will only get back into the game if they have a monster word to play. One way that could happen would be for the opponent to play a 7-letter word from the TW in upper-left to the square above the E in EQUAL. This would hit two TL squares and could lead to a massive score. We might suddenly find ourselves 25 or 30 points behind.
Therefore, we will sacrifice a few points and play across the top row. This is a defensive play that we are making, to close that row and prevent a possible catastrophe.
The specific word we play is NADA, which puts the N and D on TL squares and gives us a score of 16.
30 points. Winning 261-202.
The WWF AI has played RIF for 27 points. Immediately we notice that this opens up the TW square above the R.
We start looking for 5-letter words without a 1-letter prefix to play here. This will keep the other TW in that area closed from our opponent. They will then be unable to play into the space to the right of our word. Essentially, we can use the threat of playing a TW to forestall a number of moves from our opponent.
A specific word that will accomplish this goal, while providing a decent score, is PAGER. There is no 1-letter prefix to PAGER, and the P and G will provide a decent score of 30.
12 points. Winning 273-256.
Our AI opponent has responded with TANNERY, playing all seven letters and scoring 54 points. Suddenly, the game is almost tied.
Our own rack consists of six vowels and one consonant. If we had a better set of letters, it would be a simple matter to use the TL square above the Y, in tandem with either TW square in the lower-right, and rack up a strong score with our response.
However, with the letters we do have, our options are very limited. If it were early in the game, we would probably just exchange and hope for something better. At this point in the game, though, with two different TW squares open, it is simply too risky to give our opponent a second shot to score big. If we swapped tiles, and they had a strong wrong which used the TL/TW square combination, they could suddenly have a significant lead.
Therefore, we could again focus on playing defense here. We can't make a great word, but we can at least occupy some of the premium squares so that our opponent can't blow the game open.
It is in this type of situation where a broad Words With Friends® vocabulary becomes extremely important. It would be terrific if there were even a single 4-letter word that would play from the TL-TW row above the Y in TANNERY. Unfortunately, our letters can only make one 4-letter word, ETUI, and the EY that would be required is invalid.
Another option is to play a 3-letter word from Y down to the TW square in lower-right. This would block that square, but would also open up the entire bottom row. If our opponent played another 8-letter word there, hitting premium squares, the game might be effectively over.
With all of that being said, let's try to play something along the top of TANNERY that will muddle that section of the board, and hopefully keep the game even until we can pull some better letters. This is where familiarity with 2 and 3-letter words is essential.
After looking at a couple of options, we can play OUT and combine it with OY. This will block one TW square, and keep the damage under control if our opponent uses the other one (by at least preventing them from putting a good letter on the TL and having its value be multiplied by nine).
Normally, we would be begging our opponent to play OUTS on their turn with another long word, which would devastate us. However, if you'll remember from earlier, there are no S tiles left to be played. Since no other letter is a suffix to OUT, we are actually blocking the TW square permanently with this move! Knowing which tiles are still to be played can be very important in a situation like this.
Having said all of this, I am now playing OUT for 12 points. The point value is, needless to say, terrible. But this should hopefully prevent the WWF AI from getting a huge score, and maybe we can draw something better next turn.
NOTE: In retrospect, there was still a blank tile outstanding when we played OUT, which makes this move considerably more risky than implied in my analysis above.
27 points. Winning 300-269.
Our opponent has played MA for 13 points. We don't know exactly what letters they have, but it looks like our blocking maneuver from last turn worked.
Unfortunately, we continue to have a lopsided letter distribution. We have three I tiles, two E's, and only one consonant. It looks like this game will be going to the wire.
We have a few options here. We could just swap our tiles. We could play something like VIAL from the V and get a decent score. Or we could look for something which might utilize any number of DW squares while combining with the words on the board.
At this point in the game, we can look at the Tile Bag and make a reasonable guess as to what we might receive. There are 24 tiles left (17 in the bag, and 7 on our opponent's rack). The only high-value letter out is the J. There is still one blank. And, there are 9 vowels and 15 consonants.
Early in the game, it might make sense to chance the Tile Bag. At this point though, even if we drew the perfect 7-letter word, it might be impossible to play it. So many areas of the board are now closed off or blocked that we would need an absolutely perfect draw for it to pan out.
Furthermore, we can make 27 points by playing VIAL into the TW score. This is a high enough score that it doesn't really make sense to swap and take a 0. It also blocks a TW square from being used by our opponent.
The one risk from this move is that our opponent has an 8-letter word, beginning with L, that would stretch all the way to the TW square. If this happens, we are probably in trouble. But, with no A's left, plus a J, C, and V still out, we can take the chance that the AI probably can't play a bonus word on its next turn.
17 points. Winning 317-284.
Our opponent has played ODE for 15 points. This is a pretty good outcome for us, given some of the fears from last turn.
We still lack an obvious good move, and have an awkward distribution of letters. It is at this point that we can really start to look for premium squares and a series of short words we can make, to consolidate our narrow lead and keep the board closed.
The other option would be to swap tiles, but it is a little late in the game to make that worthwhile.
First of all, let's decide where we don't want to play. There are a couple of areas that are well-blocked, in which the first player to "activate" will invariably come out worse for the deal.
One such area is along the right side of the board between EQUAL, FUTZ, and OUT. The first player who plays into this area might gain a DW or TL square on a short word, but they will be giving up the chance for their opponent to bonus out with a 7 or 8-letter word which hits the TW square. This would obviously be a disaster, and thus this area is something of a no-man's land at the moment.
It would also be a questionable decision to play into the area between the TW/TL squares between PAGER and MA. Those are almost unreachable at the moment, but the first player who changes that equation risks a much stronger response from their opponent.
Playing anything in the bottom of TANNERY is also risks opening the TW square more, and in any case we don't have good letters at the moment to combine with that word.
There is nothing great to play on the B in BREVES or the L in VIAL with what we have.
There is still the risk of our opponent playing a huge word off of the L in VIAL. In fact, that is practically the only significant risk we have right now. There are a very few words, like EMPANADA, which might combine with NADA in the top row, but the letters for our opponent to do so have already been played.
After searching for awhile, playing YE adjacent to VIAL could serve us well. It doubles the Y in two different words, and also makes it more difficult for our opponent to use the L in VIAL in devastating fashion.
12 points. Winning 329-299.
Our AI opponent has played END for 15 points. Without becoming too overconfident, we seem to have a good lead right now, so long as we can prevent our opponent from playing a premium word.
We also continue to lack the letters to make a great score for ourselves, so let's see if we can't get a moderate score while continuing our scorched-earth strategy. We could do something like TITHE on the T in OUT, but in return for scoring 20 points we would open up a huge new part of the board for a counter-play. Let's see what else we could do instead.
We know there is a still a J and a blank out there. An obvious move we can see which would get our opponent back into things, would be to play JOYED into the TW in lower-right. This would be a 50 point play or more. Let's do something to block that. Unfortunately, there is no word we can play on the OY ourselves to use that square, so we must try something else.
If we play ETH, we are giving ourselves the best chance on defense. It will leave the stem HOY pointing towards the TW square. The three words which our opponent might then play to use the TW are SHOYUS, HOYAS, and HOYLE. However, two of these are impossible given the letters in play. There are no A or S tiles left, and SHOYUS and HOYAS both use two of those. Even with a blank, our opponent would be one tile short.
If our opponent has a blank and an E, they could play HOYLE. This would be worth 24 points, which would not be terrible for us at this point in the game. Alternatively, we could always hope to draw those tiles ourselves. With 13 tiles left, there's about a 29% chance that our opponent has both the E and blank in their rack.
So we play ETH. Again, as a defensive play to close this game out.
17 points. Winning 346-311.
Our AI opponent has played TIN for 12 points. We now know exactly what letters they have, and neither player will be drawing any more letters for the rest of the game. We are definitely in the endgame.
TIN opens the TL square by I. Our opponent has the tiles to more than take advantage of this. For instance, using their blank along with J to make JET would score 36 points for them, and they would be in the lead.
Our opponent's move does not open up the TW square. There is no one letter that can prefix TEND and form part of a word there, so our sole focus should be using the TL square by TIN and denying it to our opponent.
Our highest-value letter is G, so let's put it on the TL. Then we can play the T, making GIT and MET. This will score us 17 points and forestall the danger of a word like JET.
We will now be left with nothing but five vowels, but our AI opponent will be facing the opposite problem with its consonants. Let's play GIT and hope that our 35-point lead will be enough to see us through to the end of this game.
18 points. Winning 364-322.
Our AI opponent has played WIT for 11 points, trying desperately to open the TW square and somehow get back into this.
Our analysis is pretty similar to last turn. We need to preserve our lead and prevent any heroics from our opponent. In this case, that means blocking the TW square.
An obvious way to do this is to play WOE. It is the only word we can make which will play to that square, and it makes our highest possible score of 18 points.
From here on out, we will use our I tiles to make small words until the game is over. We should be far enough ahead to win, with the position we have.
6 points. Winning 370-329.
Our AI opponent has played GI/IN for 11 points. By now, we seem to have the game pretty much under control. Now we are simply looking for good places to play our I tiles until we can close this out.
6 points. Winning 376-226.
Our opponent played CEL for 7 points. There's not much left to say here. We're just running the game down and should win it next turn.
5 points. GAME OVER. I won 391-333.
Our opponent played EVE for 7 points. We'll just play our last I and close out the game. Since our opponent got stuck with a J late in the game, our score will get a nice bump.
It didn't matter in this specific case, but getting stuck with a J or a Q in the last turn can be devastating in a close game. Without the points from the J, I would have won by 38 points. The J added 10 points for me, and subtracted 10 points from my opponent, increasing the margin of victory to 58 points. A 20 point-swing would definitely be sufficient to change the outcome of many games.
A Q can be especially devastating, because this letter is often hard to play without a U. It can become very important to know that QI, QAT, and QIS are valid words, in order to avoid being hung with a huge loss of points.
I am now winning the series 1-0 against the Words With Friends® AI. This game was a pretty closed, tactical affair, with a relatively low final score.
A lot of this was dictated by letter distribution. During a few key turns in the middle-game, I consistently had nearly all vowels, and had to play defensive words to block premium squares.
Since this was the first time I've played the WWF AI opponent, I was worried after a few moves that it was going to be a yawner. I took a huge early lead and the opponent didn't seem to be making any strong plays. However, it got right back into it with TANNERY, and really made me think about how to get through the last few turns.
Hopefully this game has been a good example of how to play defense and deny premium squares. I also hope that the next game is a little more free-flowing and offensive. See how subsequent games in this series went by continuing to Game 2.
The words in our dictionary are meant to match, as closely as possible, the words used in Words With Friends®. We have no affiliation with Words With Friends, and offer this site for entertainment purposes only.