Some more information from the NXTlog Sumo contest is up, including a picture of the "on deck" area and the actual results of the competition, heat-by-heat. As you can see, MG_sumo by mpego beat all comers, including a rematch against the robot it first came up against, Wedge Bot. For those who might not have seen it before, the style was a double elimination: if the robot lost a match in the first bracket (the upper tree in the diagram), it was entered in the 2nd bracket, for a second shot at the title. That way there were no eliminations due to a single "lucky match", and everybody had a second chance. The list of robots is linked to their NXTlog entries (with .rbt and LDD files), so if you like you can build your own version just like the Robotics Posse did and test your creation against the "best of the best" from the NXTlog event.
I'd actually recommend looking over the entries of the 29 finalists for at least two reasons. First, there are a wide variety of strategies and building techniques in the list - they're not all perfect (& some of them didn't work, I'm certain, as intended), but there's a lot to learn from other people's designs. Second, these represent some well-documented robots on NXTlog, which was very important in this contest - after all, if the judges couldn't build them, they couldn't test them. And if the program didn't function, they couldn't run them. Testing I suspect was critical, as some of the programs I looked over clearly either hadn't completely understood the rules (no three second delay was a common problem), or hadn't tested the robot (the program, as written, just didn't work).
Some of the robot were remarkably simple, durable, and functional - DozerBot by LEgoFan65536 was a good example, and a lot of us just picked it up and looked at it due to the contstruction (OK, and the fact that it looked kinda "cute"). Some, like Intelligent Sumo by rjbsfarm, were a blast to watch even if they didn't do quite as well as intended (here, Intelligent Sumo did pretty much what it was supposed to, but couldn't run fast enough to avoid some of the more aggressive opponents. It's also a great example of documentation without an LDD file - the pictures were clear, ordered, and made construction possible. I've got some video of it going against Not-So-Blind-Bandit that is just great, I'll try to get some together later. Not-So-Blind-Bandit by poetryinmotion used a novel sound-based system that didn't seem to function perfectly for detecting the opponent, but it was well-tested and stayed on the ring even when it missed its opponent, allowing it to do fairly well (John Brost built this one at home, so it had by far the nicest color scheme of the day, a red-&-NXT-white shell that just made it a nice looking robot overall). And then there were some that while they had a cool weapon, like NXTardigrade nosepuncher sumo, it didn't function well in the ring... but we had a great time retasking it to punch Jay in the nose repeatedly :).
Hopefully some of us or LEGO will be able to pull together the time to put up some video soon. We do have some, it just takes time to edit it... and it's a lot of fun to watch.
--
Brian Davis
I'd actually recommend looking over the entries of the 29 finalists for at least two reasons. First, there are a wide variety of strategies and building techniques in the list - they're not all perfect (& some of them didn't work, I'm certain, as intended), but there's a lot to learn from other people's designs. Second, these represent some well-documented robots on NXTlog, which was very important in this contest - after all, if the judges couldn't build them, they couldn't test them. And if the program didn't function, they couldn't run them. Testing I suspect was critical, as some of the programs I looked over clearly either hadn't completely understood the rules (no three second delay was a common problem), or hadn't tested the robot (the program, as written, just didn't work).
Some of the robot were remarkably simple, durable, and functional - DozerBot by LEgoFan65536 was a good example, and a lot of us just picked it up and looked at it due to the contstruction (OK, and the fact that it looked kinda "cute"). Some, like Intelligent Sumo by rjbsfarm, were a blast to watch even if they didn't do quite as well as intended (here, Intelligent Sumo did pretty much what it was supposed to, but couldn't run fast enough to avoid some of the more aggressive opponents. It's also a great example of documentation without an LDD file - the pictures were clear, ordered, and made construction possible. I've got some video of it going against Not-So-Blind-Bandit that is just great, I'll try to get some together later. Not-So-Blind-Bandit by poetryinmotion used a novel sound-based system that didn't seem to function perfectly for detecting the opponent, but it was well-tested and stayed on the ring even when it missed its opponent, allowing it to do fairly well (John Brost built this one at home, so it had by far the nicest color scheme of the day, a red-&-NXT-white shell that just made it a nice looking robot overall). And then there were some that while they had a cool weapon, like NXTardigrade nosepuncher sumo, it didn't function well in the ring... but we had a great time retasking it to punch Jay in the nose repeatedly :).
Hopefully some of us or LEGO will be able to pull together the time to put up some video soon. We do have some, it just takes time to edit it... and it's a lot of fun to watch.
--
Brian Davis