[foota]

I noticed the game feels pretty hard because of the letter distribution, if you want to make it easier you might try using the english distribution, which can be found at https://en.wikipedia.org/wiki/Letter_frequency#Relative_freq...

For those curious the rough differences (in game - in english, out of 100) are:

e: -4.7 t: -2.5 c: 2.4 h: -2.3 p: 2.3 u: 1.9 b: 1.6 m: 1.6 l: 1.5 g: 1.4 r: 1.4 k: 1 x: 1 z: 0.8 o: -0.7 n: -0.6 f: 0.5 j: 0.5 v: 0.5 y: 0.5 d: 0.4 a: -0.3 q: 0.2 s: -0.2 w: 0.2 i: 0.1

edit: this might go too far towards making it easy, but you could also try and group common pairs of letter near each other.

[arethuza]

As a one time practitioner of "good old-fashioned" AI I had thought it was ETAOIN SHRDLU

https://en.wikipedia.org/wiki/Etaoin_shrdlu

https://en.wikipedia.org/wiki/SHRDLU

[StavrosK]

That's close enough, and it'll vary based on the dataset. In the provided link, it's ETAOIN SHRDLCU, which is almost the same.

[abvishek]

Actually I tried to implement same. check code here

https://github.com/AbhishekCode/word-tetris/blob/master/src/...

I guess my implementation is still not perfect. I looking for ways to improved this. Thanks for pointing this out.

[tikwidd]

You could try giving a choice of two or three letters that the user selects to make it a bit easier to form words. Or the block could cycle between several letters as it drops and the letter gets decided when it hits the bottom (maybe some blocks could do this?).. maybe wildcard letters too? And it would be fun to turn up the fall velocity on more common letters!

It would also be neat to score words with less common letters higher like scrabble, or map your dictionary to a distribution in an English language corpus and score them based on rarity.