FIBS Board backgammon forum

TourneyBot performance in 2013

Author Topic: TourneyBot performance in 2013  (Read 3753 times)

Offline Tom

  • Moderator
  • Fibsboarder ++
  • *
  • Posts: 522
TourneyBot performance in 2013
« on: February 26, 2014, 06:26:18 PM »
During Tuesday's Three Five and again today there were shouts from rif_raf, a regular FIBS TPR tournament player, that TourneyBot code is badly broken, poorly written and needs to be fixed. The disparaging comments began Sunday during the Bloody Mary when he was timed out. The comments were over the top. They have been for some time.

Considering the effort we have put into the code and system design, it is insulting and deserves a proper response.  Many of us have devoted countless hours, directly and indirectly, toward providing a stable and accurate tournament management system. Since Tom began supporting the bot in 2008, many others have contributed numerous ideas to the development effort, especially socksey, sixtie (sixty_something), diane, KMA and aviator to name but a few. Shortly thereafter, Ron (sixtie) and Tom began collaborating on code testing, maintenance, new features and continued support.

That collaboration continues actively today. Even though sixtie is no longer on-line or visible as much as before, the phone calls and emails between us are numerous each week. No TPR or TourneyBot decision is made without mutual agreement. We are proud of the system we've developed. We think most of you enjoy it, many on a daily basis.

Just to put the issues in perspective, let's review the bot's performance for 2013 to see just how broken and in need of fixing TourneyBot has been for well over a year. You decide what you think.

In 2013, there were 1,429 regularly scheduled tourneys resulting in 14,205 tournament matches or rounds played. There were 15,634 players overall in those tournaments or an average of 10.9 players per event. Of these players, there were 471 unique FIBS players with 204 unique winners.

During the course of the year, there were only 5-10 detailed reports of possible "Incorrect" timeouts, the bone of contention frequently raised in shouts by players, especially those with concerns over the dreaded TIMEOUT message. When a detailed report is made via email or a FIBS message, we take it very seriously. Each one is investigated carefully and fully. This can take anywhere from a few minutes to a few days. Timeout issues are one of the more subtle and complex programming issues. We investigate each report of a possible error in the fullest way possible. This can involve days of back and forth investigating, detailed questions, followup with involved parties and very careful scrutiny. Of the 5-10 reported errors, ONLY 1 of them was an error by the bot. The problem was very subtle to find, but it was fixed ASAP.

For the sake of argument let's say ALL 10, not 1, were bot errors. We can accept that even a perceived error, while technically not a coding or design error, is nonetheless perceived by players as a serious problem. So, we accept 10 as the number of reported errors. Only 10 clock errors in 14,205 matches yields an error rate of 0.000702978 or 0.07%. Hello? That's a perceived error of ONLY 7 in every 10,000 matches. But remember, ONLY 1 actual error was detected or 0.007 actual errors. Folks, that's 7 errors in 100,000! Double those numbers, triple them push it by 10x for unreported errors. Is even 7 errors in 1,000 (10x the perceived errors reported and 100x the actual technical errors) really that alarming?

Broken? Needs fixing? Bad coding? Judge for yourself. Ironically, the same player involved in the over the top criticism, rif_raf, is the player involved in most of the carefully investigated timeout error reports. In ALL cases his timeouts were the result of him playing other non-tournament matches when assigned a match by TourneyBot or complaining that his opponent was not on-line when indeed they are. His unwillingness to change his behavior, i.e. not play non-tournament matches during a tournament, or learn to properly refresh his player list, RAWWHO, continues to result in clock difficulties for him. For us and other tournament players, his absurd comments in shouts about TourneyBot are at the least annoying, at the worst destructive of the credibility which we have worked hard to establish and justly deserve. Fortunately, we have thick skin. We also have a sense of humor. It is always enjoyable to see someone making a fool of themselves in shouts.

We are always looking for ways to improve clock handling, after all the goal is 0 errors. With only 1 in 2013, maybe 2014 will be our year of technical perfection! We would also like to achieve zero perceived errors. The key for us reaching that goal is to receive specific reports when an error is suspected. These reports often result in expansion of details that are logged during clock events which ultimately yield further improvements in design. Additionally, improved messaging for all players involved are also added regularly when perceived problems persist.

Participation in FIBS TPR tournaments has been growing annually for several years now. We expect that growth to continue and are preparing and planning for it with every design and coding change. During 2014, several new features in TourneyBot, TPR and other new developments will be announced.

Thank you all for your continued support and constructive comments! Thank you critics as well. It keeps us on our toes. While we have often wasted hours chasing down erroneous error reports, it does provide a good laugh as well. After all, expressing one's opinion, regardless of how absurd, in FIBS shouts is part and parcel of what makes life at FIBS and TPR tournament play so much fun.

Tom Moulton, TourneyBot Admin and Coding Czar
Ron Blake, TourneyBot Admin and TPR Czar

FIBS Board backgammon forum

TourneyBot performance in 2013
« on: February 26, 2014, 06:26:18 PM »

Offline jeeves

  • Fibsboard Executive VIP Donor 2013
  • Fibsboarder +
  • *
  • Posts: 80
Re: TourneyBot performance in 2013
« Reply #1 on: February 26, 2014, 08:31:58 PM »
A very interesting read, and one that goes to show the progress you guys have made in relatively recent times.

I hope that rif_raf hangs his head in shame  :cry: if or when he gets to read this, but that I doubt.  Unfortunately he is not alone in making disparaging comments, and these so-called 'old school' players/moderators are misinterpreting these incidents (at best), or deliberately misinforming and misleading FIBS users (at worst).  My guess is it's geared to achieving a desired outcome, namely spreading their influence across every tournament.  Imagine the furore if all Tournament Moderators were to be active for every tournament!  :ohmy:

Anyway, thank you both for highlighting how successful tourneybot is, and I wish you every success in achieving a 0% error rate for 2014.  :yes: :thumbsup2:

FIBS Board backgammon forum

Re: TourneyBot performance in 2013
« Reply #1 on: February 26, 2014, 08:31:58 PM »

Offline socksey

  • Global Moderator
  • Expert
  • *
  • Posts: 2,916
Re: TourneyBot performance in 2013
« Reply #2 on: February 26, 2014, 10:24:17 PM »
  Imagine the furore if all Tournament Moderators were to be active for every tournament!  :ohmy:

I think you mean "furor", and my recent suggestions that all Tournament Directors be named in all tourneys was not to cause furor, and not for all TD's to be active in all tourneys.  My suggestion was designed so that if the main TD of any tournament was not there, then perhaps a backup would be in the case of a possible wrong timeout.

socksey



Imagination is more important than knowledge. - Albert Einstein

Offline dorbel

  • Advanced Fibsboarder
  • *****
  • Posts: 1,150
Re: TourneyBot performance in 2013
« Reply #3 on: February 27, 2014, 02:13:51 AM »
Furore is the English spelling.

How many wrong time outs are there that require human intervention? Tom seems to think that the number is vanishingly close to zero. I play 500+ tourney matches a year and have never been incorrectly timed out that I can recall. I think that there did use to be more problems in the already dim and distant past, but recently no.

It may or may not be a good idea to have a back-up human td. The intervention of a fallible human isn't guaranteed to produce a better result IMO, not least because humans can be seen, correctly or not, as being partial, leading to a lack of confidence in decisions and further opportunities for furore. Tourneybot doesn't have any friends, whatever riffy may think. I'll take my chances with the bot thanks.

FIBS Board backgammon forum

Re: TourneyBot performance in 2013
« Reply #3 on: February 27, 2014, 02:13:51 AM »

Offline jeeves

  • Fibsboard Executive VIP Donor 2013
  • Fibsboarder +
  • *
  • Posts: 80
Re: TourneyBot performance in 2013
« Reply #4 on: February 27, 2014, 07:01:05 AM »
if the main TD of any tournament

I think you mean TM......Tourneybot is the only TD these days

Offline socksey

  • Global Moderator
  • Expert
  • *
  • Posts: 2,916
Re: TourneyBot performance in 2013
« Reply #5 on: February 27, 2014, 02:10:53 PM »
Semantics jeeves.  Tournament Director has been used since I have been doing this which has been years.  The recent sixtie shout diatribe indicating that he was not TD of Bloody Mary's was his way of copping out when he wasn't able to stop a timeout in time having not been notified soon enough.  Tourneybot is not the only TD these days.

socksey



When you sit with a nice girl for two hours, you think it's only a minute. But when you sit on a hot stove for a minute, you think it's two hours. That's relativity. - Albert Einstein


Offline jeeves

  • Fibsboard Executive VIP Donor 2013
  • Fibsboarder +
  • *
  • Posts: 80
Re: TourneyBot performance in 2013
« Reply #6 on: February 27, 2014, 07:51:24 PM »
which has been years

Yeah, that's the trouble with the Old and Bold Brigade, resistance to change




P.S.  Tom's right, so I'll end it here and let you have the last word because you're like a dog with a bone and before it starts to get personal and you tell me to look in the mirror or something like that

“Change is inevitable. Growth is optional.”

― John C. Maxwell
« Last Edit: February 28, 2014, 06:44:39 AM by jeeves »

Offline Tom

  • Moderator
  • Fibsboarder ++
  • *
  • Posts: 522
Re: TourneyBot performance in 2013
« Reply #7 on: February 27, 2014, 09:25:29 PM »
Hey guys, let's not let this get out of hand.

The important points of this thread:

1) TourneyBot has been doing a very good job with the clock handling
2) It is not perfect, but is getting better each year
3) We are comitted to improving the clock handling and adding other features on an ongoing basis

Tom

FIBS Board backgammon forum

Re: TourneyBot performance in 2013
« Reply #7 on: February 27, 2014, 09:25:29 PM »

Offline KDP

  • Fibsboarder plus
  • ****
  • Posts: 257
Re: TourneyBot performance in 2013
« Reply #8 on: February 28, 2014, 12:37:03 AM »
thank you tom, although its probably not said often enough, there are plenty of us who appreciate the hard work and long hours you put in  making fibs a better place to play bg.

Offline socksey

  • Global Moderator
  • Expert
  • *
  • Posts: 2,916
Re: TourneyBot performance in 2013
« Reply #9 on: February 28, 2014, 02:46:46 AM »
Yeah, that's the trouble with the Old and Bold Brigade, resistance to change

Resistant to change is not what I'm about at all.  Perfection is what I strive for and I have always applauded Tom in his efforts.  I just find it frustrating when I see a possible wrong timeout, the named TD is not online, I am, and I am helpless.  If you find fault with that, then perhaps you should look in the mirror and see where the real problem resides. 

socksey, one of the "Old and Bold Brigade



If my theory of relativity is proven successful, Germany will claim me as a German, and France will declare that I am a citizen of the world. Should my theory prove untrue, France will say that I am a German, and Germany will declare that I am a Jew. - Albert Einstein




Tags:
 

anything