Selection of the finalists

[omichel]

Today at 15:00:00 CEST, we will select the best 32 teams from the leader board who will participate to the finals.

I will post here the ranking of these 32 selected teams which will be used to organize the finals.

Next week on September 26th between 15:00:00 CEST and 18:00:00 CEST, we will freeze the code of the competitors and use this code to run the finals. Please do not commit any change in your code during this time frame as there is no guarantee it will be taken into account.

[omichel]

Here are the 32 finalists:

[kibergrad]

@omichel It appears that Team Humanoid is currently in the first position, and HeavyRotation is not in that position after the queue has ended and is no longer accepting any further triggers. Is that correct?

I recall that you waited during the last ICRA event for the queue to conclude.

[omichel]

No, the deadline was 15:00:00, regardless of the queue. And the queue is not blocked at the moment, so that the ranking may still evolve. However, the ranking retained to organize the finals is the one of today 15:00:00 as displayed above. This doesn't mean that HeavyRotation has won the finals. It simply means that HeavyRotation will play its first finals game versus ADWIK while Team Humanoid will play its first final game versus T, etc.

[kibergrad]

Thanks for the clarification.

[omichel]

Reminder: the code freeze is going to happen in one hour from now. Please do not push any change to your repository between 15:00:00 CEST and 18:00:00 CEST.

[omichel]

Due to some problem caused by an update today of the NVIDIA CUDA package on the server machine, we are extending the code freeze deadline by one hour, e.g., until 16:00:00 CEST.

[omichel]

All the controllers have been downloaded. We are going to run the final games next week.

[omichel]

The finals table is available here. The first round will be run on October 2nd, 2023. Stay tuned!

[omichel]

I did a mistake for the 1/8, 1/4 and 1/2 final games which is now fixed here.

[kibergrad]

Hello @omichel,
I hope you're doing well. I wanted to bring up a potential discrepancy in the results-iros2023.md page regarding the format of the upcoming finals. It seems that there might be conflicting information. The result.md and readme.md mentions that the finals will be in a best-of-five format. However the results-iros2023.md states it will be in a best-of-three format. If it's convenient, could you please clarify which format will be used for the finals?

I understand changing the format now might not be possible as the 1/16 finals are approaching and a set of five will significantly increase the duration of finals. I appreciate your assistance in resolving this matter, and I hope it doesn't bring too much trouble to you.

Thank you for your understanding and efforts.

[omichel]

Oh yes, you are right. We should run the best-of-five format. I will fix this in result-iros2023.md.

[omichel]

I just fixed it. Let me know if you spot any other issue with it.

[kibergrad]

Sure, I will.
Thanks for your time.

[AndrejOrsula]

Thank you very much for the effort required to organize this competition yet again!

Good luck to all the participants in the finals! 😃

[omichel]

The first round is now complete, check the results at https://github.com/cyberbotics/wrestling/blob/main/result-iros2023.md
Please let me know if you spot any major problem as I didn't visually check everything.

[omichel]

I don't see you controller crashing in the first game. It keeps moving all the time. At which time do you believe it crashed?

[adwik-wiz]

@omichel at 1:16 after standing up , it should have kept moving , but it didnt move just the hands were moving

[MJavadZallaghi]

I noticed that my robot “MechaCheetahs” is frozen in one of the games at the second 25, While We have won the game, but “robot being frozen” is not part of our robot control strategy!
I just wanted to mention this issue for your reference in the future.
Thanks in advance.

[omichel]

@adwik-wiz: Yes, I noticed that, but the problem here is clearly in your controller, not in our infrastructure as your controller is live and didn't even crashed.

[omichel]

@MJavadZallaghi: In your case, it could be two things: either your controller crashed for some reason (which could be a bug in your code, or some external reason) or it didn't crashed but stopped sending commands to the robot for some reason (strategy or bug in your program). Since you won the game, I didn't inspected the reason for this apparent crash. I will investigate only if a controller loose a game while there is a suspicion of crash for it. Only in the case where I identify it crashed for some external reason, I will re-run the game.

[kibergrad]

@omichel at 1:16 after standing up , it should have kept moving , but it didnt move just the hands were moving

Something similar has happened during the past with my controller. And most of the time it happens just after standing up. I don't know the reason but yes, this happens quite sometimes @omichel

[adwik-wiz]

@omichel
Will the matches where the controllers froze be re-run, considering other participants are experiencing the same issue?

[omichel]

I don't believe that in this case there is a frozen controller problem, because the robot is clearly not crashed and sending regularly motor commands to the arms of the robot. It is rather a bug in the controller code that in our infrastructure.

[omichel]

So, clearly we won't re-run that games.

[adwik-wiz]

It's ok if you feel so, but its quite strange on my part as i never saw the bot previously acting like this on webots.
Thanks for considering my issue.

[omichel]

The second round (1/8 finals) is now complete, check the results at https://github.com/cyberbotics/wrestling/blob/main/result-iros2023.md
Please let me know if you spot any major problem as I didn't visually check everything.

[MJavadZallaghi]

I checked some matches, and every thing was OK.

[omichel]

The 1/4 finals is now complete, check the results at https://github.com/cyberbotics/wrestling/blob/main/result-iros2023.md.

[omichel]

The semi-finals are now complete! Check it at https://github.com/cyberbotics/wrestling/blob/main/result-iros2023.md.

[danielebg]

Hi @omichel in the 1/4 finals the game 3 of uditech vs HexHowells, From minute 1:38 the HexHowells robot leaves the platform and at one point its counting time does not advance. Was there any error in that simulation?

Thank you

[omichel]

This is not a bug, but a limitation of the KO algorithm. It seems that when the two robots were down, out of the platform, HexHowells managed to get higher than uditech, even though it was out of the platform thus stopping its KO countdown temporarily, while uditech kept falling having its countdown running more often than HexHowells. I agree this may seem a bit strange, but if one of the robots had clearly recovered and reached a stable position on the platform, it would certainly have won that game. Obviously, this was not the case and I admit that the determination of the winner is a bit fuzzy in such a case.

[omichel]

And the winner is 🎉 MechaCheetahs!
Silver medal goes to Toddler and Bronze medal goes to HexHowells.
Congratulation to the winners and to all the participants who really did a great job.
I would like to address a special mention to ijnek who was the only participant using a programming language other than Python, e.g., ROS2.
All the games are available on https://github.com/cyberbotics/wrestling/blob/main/result-iros2023.md.

[MJavadZallaghi]

OMG!
I'm really happy that our team “MechaCheetahs” won this amazing competition.
All the teams performed well.
Especial thanks to you, dear @omichel for supporting us in this awesome competition.

[omichel]

I just reset the main webservice so that you can continue, if you want, to participate to this online competition, although we have no plan yet for another round of finals with a prize.