Does anyone want to participate in a Computer Scrabble Olympiad, to be conducted over the Internet in early July? We have in mind about 100 games per...
sheppardco@...
Mar 15, 2002 5:35 am
398
In case anyone didn't see it: I have always admired computer programs that simulate the evolution of lifeforms (for example, see the free "Darwin Pond" at ...
Thanks for the reference to Darwin Pond--I plan to check it out. I was a MUMPS programmer for almost 20 years and still use it regularly for help in puzzle...
I recently did a word generator in Java using a variation on Markov/NGram models. It intentionally tries to generate unseen words by throwing out words which...
... Add to that throwing out anything that can be found on Google or Alta Vista, and you have a product there that you can sell to marketing companies! They're...
I had my computer program play itself in a million games (actually 999,800) this afternoon. It took approximately 39 minutes to do on several processors. I...
Top 10 combined scores? Games with most bingos? Have any record of what the board looked like after that 241-167 game? Must've passed out.... Interesting to...
Here are two of the five 8 bingo games from the million. I'd post more, but it takes about 10 minutes to grep the data to find the right game :) 0 2 7 8 7...
wow, 5 in 1M? I would think it would be much higher. I'm sure I've played less than 5k games in my life and can specifically recall 2 8s and a 9. All were...
... I postulate that it will go up dramatically when I put on SOWPODS and calculate better 2-6 tile leaves. I think the present leave heuristics may prevent...
... A couple of other questions not answered: - What is the correlation between winning & losing score? - How many ties? - Percentage of first-player wins &...
My theory on why Brian came up with so few 8/9 games is that humans probably look for bingos and computers probably play more smaller plays that score a great...
John Babina
john@...
Mar 28, 2002 5:29 am
410
... I'm not sure I understand the question ... 4587 I suspect this would be quite a bit different if my Scrabble program had an exhaustive endgame solver for...
... I'm looking for the statistical correlation coefficient. On one hand, since the average combined score is 835 (I think), the more points the winner gets,...
To: "Alan Frank" <alf@...> ... 50,000. My thought is that the odds of this occurring for you would be very different than the computer-computer...
John Babina
john@...
Mar 28, 2002 1:58 pm
413
I've been looking forward to seeing the statistics Brian posted for a long time, and I'm excited by all the response it's triggered. I too was surprised by...
I know this is a very late response to this post, but I think it's time for me to implement one of these as well. I computed the million games, but I'm...
... Maybe more intelligent replies to the opening will reduce it to 55%. Maven-Maven games also show a 56% edge to the first player. First player has an edge...
sheppardco@...
Mar 29, 2002 5:02 am
416
I found Brian Sheppard's response to this thread to be most reasonable. I would like to add a few brief thoughts: - I also believe the heuristics do not...
Steven Gordon
sagordon2@...
Mar 29, 2002 2:16 pm
417
Well, my idiocy prevails on this one. I sent an odd looking game from the million to John O'Laughlin and he did an independent analysis. ... It starts out...
... OK, I have my data in front of me. From 131,000 Maven-Maven SOWPODS games, I have the top bingo is word #178: ETAERIO. Here are all of the bingos in the ...
sheppardco@...
Apr 2, 2002 3:16 am
420
... Maven does something similar. I calibrated the length penalty (which is an array of 7 bonuses that apply to exchanges (set to zero by definition), and ...
sheppardco@...
Apr 2, 2002 3:28 am
421
... Better get started, or they will change the dictionary before you finish. :-) Seriously: why is this better than generating the same results on the fly? It...
sheppardco@...
Apr 2, 2002 3:34 am
422
... Inferences. If you know every rack that generates a given opening move, you know your opponent's possible leaves. Granted, this assumes your opponent ...
I said I believed the ultimate program should have an opening library. The use of the word ultimate was meant to imply that I did not believe it was the...
Steven Gordon
sagordon2@...
Apr 2, 2002 2:42 pm
424
Some time ago I ran my program against itself for just over 2.4 million games. (2425227, to be exact.) Although I didn't make notes at the time about...
Having problems with message search?
Fill out this form to ensure your group is one of the first to be migrated to the new message search system.
