Contribute your work to the TourneyBot project

[MadMatt]

TourneyBot is becoming a fairly complex project (and I'm not talking (just) about the code here). New feature requests are being submitted almost daily now, some old ones are still to be implemented and there are many things people interested in helping out can contribute, thus freeing up a bit more of my time to focus on improving TourneyBot.

I have started this topic to maintain a list of current tasks that I would greatly appreciate some help with. If you want to help with any of them, please let me know. Since TourneyBot is a volunteer project of the FIBS community, I cannot offer you anything for your effort except mention your name in the official TourneyBot credits. Of course, helping out will also greatly increase your chances of being appointed as a Tourney Director should you ever want to become one.

So, if you have time and the needed expertise, I invite you to contribute to this FIBS improvement project. Your efforts will be greatly appreciated!


Tasks:

• Write an HTML page to explain how to communicate with TourneyBot using different FIBS clients. Explanation: A lot of people have trouble communicating with TourneyBot. They don't really understand what it means to send a tell to TourneyBot. A web page explaining how to use their client to communicate with the bot would be a great way to alleviate that problem. I still need this written for every FIBS client except MacFIBS and JavaFIBS 2001. Required skills: HTML programming preferred, but if you can't code HTML, a plain text file will do and I will convert it to HTML myself. Also you need enough knowledge of your client to use TourneyBot with it.

Right, I'll add some more in a couple of days. Please don't hesitate to contact me if you want to help with one of these tasks.

Completed tasks:

• How to communicate with TourneyBot using MacFIBS. Credits: jinnate. You can see it here.
• How to use TourneyBot with JavaFIBS 2001. Credits: amarganth. You can see it here.

[jinnate]

i'm working on a page explaining how to communicate with tourneybot using macfibs and i'll email it as soon as i'm done.

[MadMatt]

i'm working on a page explaining how to communicate with tourneybot using macfibs and i'll email it as soon as i'm done.

Good girl, jinnate! Any takers for other clients?

[MadMatt]

Jinnate finished the MacFIBS part of the Using clients with TourneyBot How-To. Thanks, jinnate! I've put it on a page here.

[jinnate]

my schedule's been a little busy lately, but i'm still playing with the idea of doing some research for the double elim stuff. i probably wouldn't be able to devote much time to it until june.

[amarganth]

I'll do it for JavaFIBS. Because of my poor english, I'll need someone, who can check my writings.

amarganth.

[JLee]

I'll be happy to work on the double elimination format right away.  I've got a format setup already for a "progressive consolation" -- which is probably the canonical method of running a double elimination tournament.  But I'll research other methods as well.

JLee

[alef]

I'm still confused over the Double Elimination format. The one I've been using for the postcard tournaments has the problem that opponents sometimes end up playing each other twice prior to the final. From what I can tell this is unavoidable unless the bracket system is dynamic, meaning that you don't have automatic movements depending on just the past win and loss but factor in the previous matches at each stage, which I don't think the php script we currently have can easily handle.

[JLee]

I'm still confused over the Double Elimination format. The one I've been using for the postcard tournaments has the problem that opponents sometimes end up playing each other twice prior to the final. From what I can tell this is unavoidable unless the bracket system is dynamic, meaning that you don't have automatic movements depending on just the past win and loss but factor in the previous matches at each stage, which I don't think the php script we currently have can easily handle.

This is not such a problem, because if you do a DE bracket correctly, you twist the placements of the players bumped down from the winner's bracket.

Completely avoiding rematches is essentially impossible anyway.

Your first 3 postcard tournaments had the canonical DE format -- your 4th tournament deviated in a way that I think is suboptimal.

JLee

[alef]

The changed format in the 4th tournament was to remove a round from the 2nd bracket following advice from dorbel, who sounds quite knowledgeable on the format. Problem before was that if you remain unbeaten you can win the final in just six matches, whereas if you lose the first match you then have to win ten -- a major difference. In tournament four this has been reduced to nine, still a lot more but an improvement, and also I believe the same system used by the Bracket Maker site previously recommended on this board.

[JLee]

The changed format in the 4th tournament was to remove a round from the 2nd bracket following advice from dorbel, who sounds quite knowledgeable on the format. Problem before was that if you remain unbeaten you can win the final in just six matches, whereas if you lose the first match you then have to win ten -- a major difference. In tournament four this has been reduced to nine, still a lot more but an improvement, and also I believe the same system used by the Bracket Maker site previously recommended on this board.

The problem is that the advancement is not uniform.  The suggestion made is in a good spirit, but I strongly disagree with its usage.

Basically every double elimination tournaments are run in the standard method, which is how you ran Tournaments #2 and #3, and that's the way any DE tournament should be run.

It puts a premium on winning in the main flight, as it should.  It SHOULDN'T necessarily be easy to advance through the loser's bracket to the finals.  You should LIKE the idea that it's harder to advance through the 2nd bracket.

Besides, do you really think sacrificing a uniform system is worth making it so that you have to win "only" nine rather than ten?  That's hardly a difference, but you've sacrificed a great deal of uniformity, and you've sacrificed it in a part of the tournament where it's getting near the end, and it makes a big difference.  This isn't an improvement.

Please return the postcard tournaments to the standard DE format, rather than this nonuniform setup.

JLee

[MadMatt]

Is this "uniform" system described anywhere? The literature on double elimination systems seems to be scarce on the internet.

[MadMatt]

Many thanks to amarganth, who explains how to use TourneyBot with JavaFIBS 2001.

If anyone else would like to write a similar document for their client, please don't hesitate to contact me.

[JLee]

Is this "uniform" system described anywhere? The literature on double elimination systems seems to be scarce on the internet.

I have not found a formal description of the uniform system.

However, I poked around on the net the other day, looking at different DE brackets for various games/sports.  With extremely few exceptions, they are ALL identical.

Hence, I have described this as a "uniform" DE system.  The backgammon world will call this type of event a "progressive consolation".

If you would like, I could give you a precise general formulation of how to assemble the brackets.

JLee

[JLee]

Another system proposed elsewhere on this board is to create two separate brackets, and have the winners of the brackets meet.  Each player is entered once in both brackets, so it's ostensibly double elimination.

There are a couple of obvious problems:

* How do you arrange the pairings so as to minimize the probability that two players will meet in both brackets (this seems like a good thing to avoid).

* Can this still work in the event that there are not 2^n players?

JLee

[MadMatt]

Is this "uniform" system described anywhere? The literature on double elimination systems seems to be scarce on the internet.

I have not found a formal description of the uniform system.

However, I poked around on the net the other day, looking at different DE brackets for various games/sports.  With extremely few exceptions, they are ALL identical.

Hence, I have described this as a "uniform" DE system.  The backgammon world will call this type of event a "progressive consolation".

If you would like, I could give you a precise general formulation of how to assemble the brackets.

JLee

That would be helpful. I have not been able to come up with a general algorithm on how to assemble the brackets, so what I'm doing right now is generating the brackets using some bracket making software, then hardcoding that info into the bot. This is time consuming and not very elegant.

[MadMatt]

Another system proposed elsewhere on this board is to create two separate brackets, and have the winners of the brackets meet.  Each player is entered once in both brackets, so it's ostensibly double elimination.

There are a couple of obvious problems:

* How do you arrange the pairings so as to minimize the probability that two players will meet in both brackets (this seems like a good thing to avoid).

* Can this still work in the event that there are not 2^n players?

JLee

These are not main concerns with such system.

Obviously, while the issue of playing the same person in both brackets cannot be entirely avoided, it could still be alleviated somewhat with a couple of checks.

The second issue is non-existant, as TourneyBot handles byes and non 2^n brackets very well and this would just be two such brackets. This is not the case with a true double elimination system.

The main problems with it is the simplicity of the tourney and time. I imagine it would not be very easy to explain to an average FIBSter just how the two brackets go together, especially in situations where they would have to play the same person in both. Secondly, as the two brackets are paralel, it means there would be a lot of waiting. Not a problem for longer tourneys, but for the usual quickies it would be quite nightmarish.