RogerBW's Blog

Letter Tycoon, by Brad Brooks, was published in 2015 and is now quite hard to find. But I was recently introduced to it on BoardGameArena, and it turns out I'm quite good at it.

To summarise the game: you have a hand of seven letter cards, and three more make up a common pool. You aim to make the highest-scoring word out of them, which means the longest ones rather than the ones with the more obscure letters. You can "patent" one of the letters in your word (if it hasn't already been claimed), after which any time it's used in another player's word you get an extra point. You then discard any letters left in your hand, then refill both your hand and the pool. When one player has bought a sufficient value of patents, the game is won by the player with the highest score.

With that level of sophistication it's basically a game of vocabulary size, anagram manipulation, and a bit of luck. (It turns out that losing at Scrabble to my wife has made me quite decent at the first two.) But the cunning bit is the way you're encouraged to play low-value letters by giving special powers to whoever patents the cheapest of them:

letter	effect
B	Double score if word starts and ends with a vowel
J	Double score if word is at least half vowels
K	Double score if word has only one vowel
Q	May replace a letter before playing
V	May play two separate words
X	May duplicate any one letter
Z	May add a terminal S to the word

(Playing a Q as part of a word also doubles the value of that play, whether or not you patent it.) I'll call these patents the "specials".

Although the rulebook describes a system with "cash" and "stock" awards and you spend "cash" to buy patents, this is just a means of stopping someone hoovering up high-value patents in early turns; your overall score doesn't go down when you buy them. Therefore the real cost of buying a special patent is the number of points you lose by playing a word with that letter in it rather than a higher-scoring word without. The value of the patent is the number of extra points it gets you in later turns.

Unless it's close to the end of the game, I tend to feel that any special is worth taking for a point or three: even if it's not going to make many points for you, it denies them to the opposition. The main exception is the case in which you have it in your power to end the game by taking a high-value patent, and are far enough ahead to be confident of your victory. It's very rare to have a meaningful choice between two specials; if they're both in your hand, try to save the second one for next turn, while if one of them's in the common area you take that first. Still, there are a few occasions on which one may have a choice between patents – for example, if there are two in the common area and your opponent can get the other one.

By the official rules all doublers multiply: "aqua" with B and J patents would be a basic 2 points, but doubled three times to make 16. This makes the B and J combination particularly powerful, because many words which are at least half vowels also start and end with vowels. Even if you use the BGA variants (where doublers either add, so that Q+B+J means ×4, or don't stack at all, so that only one counts) these are specials that can often be played to enhance a long word.

In particular, that B+J combination is nearly a guaranteed game winner. (Indeed, if one's playing against opponents with a reasonable vocabulary, a great part of one's final score depends on the luck of who happened to draw those two letters first. I wonder about a variant in which one could draft them.)

K is nice to have, but mostly applies to short words (two-letter words can't be played at all) and doesn't often combine. Using an old copy of OSPD2+, not wildly different from the word lists used on BGA, I see:

letters	score	B	J	K	Q	B+J	B+Q	J+Q	K+Q	B+J+Q	total
3	1	71	254	789	3	71		1	2		978
4	2	190	2052	2167	16	190	1	13	3	1	3904
5	3	349	1229	2727	75	307	2	36	3	2	8635
6	5	582	5597	1479	162	578	8	118	2	8	15225
7	7	911	2399	460	300	586	19	115	1	19	23097
8	8	1133	7975	147	421	995	22	266		22	28407
9	9		7		2			1			202
10	10		7								48
11	11										18
12	12										1

(There are also a few words with no vowels in them at all, even including Y, which I assume don't qualify for the K-patent.)

and of course there are no B+K or J+K combinations. Note in particular the non-linear score increases for 6 and 7-letter words; that Q-words tend to be vowel-heavy, so as it happens any word in this dictionary which qualifies for B+Q also qualifies for B+J+Q; and that even without that the majority of B words also qualify for B+J.

So on this analysis I'd rate J as the most valuable, and B well above K, for its long-word bias even if combinations aren't available.

Q and X are the hand-manipulating specials. I find them nice to have, but they often only get me a point or four in a game. They can drag a horrible hand out of no-score-at-all territory, though. Z is similar, especially given its destructive interaction with B and often J: doubling is usually worth more than the 1-2 points for an extra letter. I'd probably still put them all above K, though.

And V… really feels like a trap. Again, it's useful for getting some score out of a horrible hand, but because of those score steps, it's much better to play a 10-letter word than to play two 5s. (Especially if multipliers are involved.) And if I am dragging three or four points out of a bad hand, I'm usually using all the available vowels to do it.

As for discarding, the other major choice in the game, I generally retain only 1-2 high-value cards (unless I have a special I'm planning to play next turn), because that way I see more new cards and have a better chance of drawing one of the specials. (Also a better chance of getting an all-vowel or all-consonant hand of course.)