Bot Contest
Here I'll be posting information on various Bot contests that challenge and test a Bot's AI and realism. Feel free to post comments and updates on contests, as well as announcements for new contests.
Posts 190 - 202 of 4,091
Posts 190 - 202 of 4,091
View Contest Winners in the Hall of Fame.
...and Chris - the challenge has obviously gotten us all pretty enthusiastic. I, for one, don't think you've sold your soul in any way. These types of concerns are going to happen in any kind of contest, so when you add the fact that the chatterboxchallenge competition involves such new and unpredictable contestants... it requires a whole new playing field! I mean, when I try to talk to someone from the regular world about my bot-building experiences, sometimes they'll get it, but most of the time they'll just sort of look at me with that "uh-oh, he believes in aliens" frown... so there are bound to be people - even judges - that just don't get it!
I'm pretty impressed with all the work you've put into uniting bot builders from around the world. Unfortunately, the humans involved behind this competition are from the Nintendo and Sega generation, so we're sometimes a little tough to satisfy to say the least! Way to pull it off! And it's by addressing issues like these that the contest will steadily evolve from year to year.
BTW, next year... I'm winning!
I'm pretty impressed with all the work you've put into uniting bot builders from around the world. Unfortunately, the humans involved behind this competition are from the Nintendo and Sega generation, so we're sometimes a little tough to satisfy to say the least! Way to pull it off! And it's by addressing issues like these that the contest will steadily evolve from year to year.
BTW, next year... I'm winning!
Jon, I hope you realize that I wasn't serious with my comments regarding my score that I place on the contest site. I was simply trying to add a little humor to what I considered a disappointing performance.
The 10 questions were a way to compare the bots and to narrow the field. No judge is going to personally talk to 58 bots. In addition it is not the only thing we are looking at. I didn't spent the last 10 days of my life talking to 50+ bot and logging them for nothing. This was a major undertaking. Things to think about:
1) Most of the bots I had to cut and paste the conversation. I couldn't rely on the bot owner to send me his logged copy because of the possibility it could be doctored. With some you couldn't even do that. I had to type it out line by line.
2) Some bots are painfully slow to reply. A simply conversations seemed to last forever.
3) I had to personally download 15+ Bots to my hard drive. Many of them you had to read the readme files to see how the fricky thing works.
I would like to have done more and give the judges more to work with but when you are dealing with so many bots you can't do much more. As we narrow the field down to 10 then that is a workable number and the judges will be able to speak with them on a one on one bases.
I have this idea I will like to throw out and see what people think. After the judges select the ten bots for the finals we the bot owners will have the opportunity to select two additional bots. My idea here is with 58 bots and only 10 being select there is going to be people who think they should be in the top ten. With the 2 additional picks we can go and select two more deserving bots. I think this is a good idea to try to include 2 more bots that people feel might have been slighted as in your case. The question is how do we select the two bots and what criteria do we apply. It is late here so I need to go but more about this later. I would like to hear what everybody thinks.
Chris
The 10 questions were a way to compare the bots and to narrow the field. No judge is going to personally talk to 58 bots. In addition it is not the only thing we are looking at. I didn't spent the last 10 days of my life talking to 50+ bot and logging them for nothing. This was a major undertaking. Things to think about:
1) Most of the bots I had to cut and paste the conversation. I couldn't rely on the bot owner to send me his logged copy because of the possibility it could be doctored. With some you couldn't even do that. I had to type it out line by line.
2) Some bots are painfully slow to reply. A simply conversations seemed to last forever.
3) I had to personally download 15+ Bots to my hard drive. Many of them you had to read the readme files to see how the fricky thing works.
I would like to have done more and give the judges more to work with but when you are dealing with so many bots you can't do much more. As we narrow the field down to 10 then that is a workable number and the judges will be able to speak with them on a one on one bases.
I have this idea I will like to throw out and see what people think. After the judges select the ten bots for the finals we the bot owners will have the opportunity to select two additional bots. My idea here is with 58 bots and only 10 being select there is going to be people who think they should be in the top ten. With the 2 additional picks we can go and select two more deserving bots. I think this is a good idea to try to include 2 more bots that people feel might have been slighted as in your case. The question is how do we select the two bots and what criteria do we apply. It is late here so I need to go but more about this later. I would like to hear what everybody thinks.
Chris
Hey Chris, nice to see you here. Don't worry, I don't think anybody fails to understand and appreciate the hard work you've been doing. I certainly couldn't do 58 chats and maintain the kind of patience you had.
Hope you don't take anything I've said as anything but constructive.
I actually DO think 10 questions is a good barometer, though I think the bigger problem with the scoring is not that the questions were bad (obviously any set of disconnected questions is going to be puzzling to a bot tuned to conversation rather than interrogation), nor even that judges didn't all get the jokes, but that the scoring directions are vague. No ill reflection on you Chris, I truly can't believe you've done all this on your own -- I could not (and would not have tried to) have done such a good job if I were on my own organizing the whole thing.
You don't think Oraknabo's number 4 was solid?
Hope you don't take anything I've said as anything but constructive.
I actually DO think 10 questions is a good barometer, though I think the bigger problem with the scoring is not that the questions were bad (obviously any set of disconnected questions is going to be puzzling to a bot tuned to conversation rather than interrogation), nor even that judges didn't all get the jokes, but that the scoring directions are vague. No ill reflection on you Chris, I truly can't believe you've done all this on your own -- I could not (and would not have tried to) have done such a good job if I were on my own organizing the whole thing.
You don't think Oraknabo's number 4 was solid?
Welcome Chris! It's good to have you here.
I dont think any of us would care so much about the scoring if the contest wasnt such a cool, exciting thing. Thanks for making it possible, and thanks for all your work.
Regarding the selection of the additional two: I'd say to take the two top scoring bots by popularity (public vote) that DIDNT make it into the top ten. That way a Bot who everyone loves but who isnt such a great question-answerer can move on to the final round.
The question I have for you is this: How are the Conversation and 10-Question scores going to be put together? Will they simply be added, meaning the 10-Question score is worth 5x as much? Or is there some other scheme to putting the two numbers together?
I dont think any of us would care so much about the scoring if the contest wasnt such a cool, exciting thing. Thanks for making it possible, and thanks for all your work.
Regarding the selection of the additional two: I'd say to take the two top scoring bots by popularity (public vote) that DIDNT make it into the top ten. That way a Bot who everyone loves but who isnt such a great question-answerer can move on to the final round.
The question I have for you is this: How are the Conversation and 10-Question scores going to be put together? Will they simply be added, meaning the 10-Question score is worth 5x as much? Or is there some other scheme to putting the two numbers together?
Yeah, if you would have had more time to spend on fine tuning the instructions and info about the contest, I think you would have saved yourself a lot of trouble... I know you tried your best, but I found it hard to figure out how all the contest procedures went. I'm sure it'll be better and better every year!
Taking the two most popular bots that don't make the cut sounds like a great idea to me (both for the contest and my own selfish reasons)
I think your list is gold. I'd pretty much pick the same ones, though I think I'd put Elbot and Markbot higher, and I would rank Alice much lower. I've been able to stump her a bunch of times and gotten a LOT of less-than-grammaticaly-correct responses from her.
If you can't rely on the popular vote, and have to work with the jury rankings, I'd recommend one last 10 question round before the 10 are chosen.
If you can't rely on the popular vote, and have to work with the jury rankings, I'd recommend one last 10 question round before the 10 are chosen.
Here's my list based on the online conversations.. As I'm working so much on the Forge, I dont have the time to talk to them individually, so this list might not be the most accurate:
Top 10: ChatBot, Dogh'd, Elbot, Eugene, Gizzle, Hex, Hillbilly Hank, Jabberwacky, Liddora, MarkBot, and Oraknabo.
Extra Two: Alice, Talk-Bot
Chris, what of the scoring question? Repost: How are the Conversation and 10-Question scores going to be put together? Will they simply be added, meaning the 10-Question score is worth 5x as much? Or is there some other scheme to putting the two numbers together?
Top 10: ChatBot, Dogh'd, Elbot, Eugene, Gizzle, Hex, Hillbilly Hank, Jabberwacky, Liddora, MarkBot, and Oraknabo.
Extra Two: Alice, Talk-Bot
Chris, what of the scoring question? Repost: How are the Conversation and 10-Question scores going to be put together? Will they simply be added, meaning the 10-Question score is worth 5x as much? Or is there some other scheme to putting the two numbers together?
And in all of this, poor, ignored Steve is sitting in a corner, taking a beating from public scores... It's funny how no one even talks to him, and judges on first appearances. When I submitted Steve, he was only 2 days old (because I wanted to make the deadline) and no, he 'didn't know squat' for the ten questions, but his intelligence has raised almost 4X that of what it was... *Steve continues to sit in a corner, lonely*
Thank you! Gizzle's a little flaky, owing to the psychology of a transplanted alien people say looks like a viking ant, but I think he's fun to talk to.
My list probably looks similar, but I'm going to have a conversation with each of them before posting...
My list probably looks similar, but I'm going to have a conversation with each of them before posting...
Though if the criterion is effort, gotta give points to Cheez ... how many lines does he ave jbryanc?
It's funny, as soon as I dropped out of first place, the activity in my transcripts went into a sharp decline.
» More new posts: Doghead's Cosmic Bar