Now that it's all over ... a few observations! Feel free to use the message board to discuss ... these are some personal observations for better or for worse! PLEASE send in your questionairre if you haven't done so yet ... =Fred 1. encoding of commonly appearing letter triplets (or doublets, or quartets) helped a good deal to focus the guessing. 2. a most useful "trick" was making good use of the scores of previous guesses to limit the "next" guess ... as follows: a) I'm going to make a guess - there are a large number of possible guesses to choose from b) IF my guess is to be a winner, then it MUST yield the identical scores that my previous guesses yielded. If it doesn't, then it couldn't have been the target word - so perhaps I'm better off with another guess! 3. One should take care when "borrowing" or creating letter frequency tables. Depending on the source, very different results could be derived. Since the test words were to be taken from public sources like the net, perhaps using a dictionary to derive letter frequencies is not the best approach! Perhaps it would be better to use a source like a novel, or random web pages, or a newspaper ... take the letter triple "THE" for example. I would guess that this triple appears much more often in the real world sources than in any dictionary. Carrying the notion further, since this problem uses only 4-9 letter words, shouldn't the frequencies be derived something like this: a) collect some public documents from the web (or similar source) b) select only the 4-9 letter words c) NOW derive your letter, doublet, triplet, frequencies Enhancing the notion further - perhaps there are separate tables for each word length. Perhaps separate tables for the beginning or ending of words of length N. 4. All of that is a wonderful idea - but the entries had to be of limited length. Encoding of these frequency tables tended to be the most difficult "coding" challenge - aside from the fundamental "algorithm" challenge upon which the POTM is based. Since this isn't the "obfuscated C contest", several entries were allowed to compete which exceeded the 25K limit - if (IMHO) it would be possible to get it to 25K by removing comments, shortening variable names, using #defines, or packing their tables using strange and lovable techniques. This is about the only forum in my life where I get to make absolute decisions on matters like these!!! So much for my observations ... I'll look for YOUR opinions on the POTM bulleting board ... check it out: http://www.sitepowerup.com/mb/view.asp?BoardID=100152 Thanks for listening - if you haven't sent in YOUR program description, please send it as soon as possible! =Fred