Tournament manger software frustration

[SteveP]

JeremyB, yah, by 3 points. That CERTAINLY doesn't make the rankings algorithm in the software broken. In fact, I highly doubt it's broken. The tournament tiebreaker system and the ELO ranking system AREN'T the same. The tiebreaker system takes ALL games into account for that tournament. The ELO algorithm only takes your current ranking into consideration.

So, just because you may have finished higher in the tiebreaker doesn't mean you finished higher in the ELO ranking. Do the math. I'm absolutely positive you'll find NO ERROR in the ELO algorithm in the software.

 

[DaytonGymLeader]

The ELO ratings aren't calculated by TMS, but most likely on a server somewhere in the Nintendo Data Center.

 

[jdb728]

JeremyB, yah, by 3 points. That CERTAINLY doesn't make the rankings algorithm in the software broken. In fact, I highly doubt it's broken. The tournament tiebreaker system and the ELO ranking system AREN'T the same. The tiebreaker system takes ALL games into account for that tournament. The ELO algorithm only takes your current ranking into consideration.

So, just because you may have finished higher in the tiebreaker doesn't mean you finished higher in the ELO ranking. Do the math. I'm absolutely positive you'll find NO ERROR in the ELO algorithm in the software.

Well, don't you think that the Tourney rankings system, and the world wide rankings systems should work together, since I mean, does it make a whole lot of sense for the tourney software to say that "Player 1" did better then "Player2" in the tourney, but when it comes to Global rankings, because a different system is used, according to it, "Player 2" did better then "Player 1"?
I don't know, it may just be me, but shouldn't they rank players both via the same system?


Also, why is tie breaker ALWAYS opponent difficulty, since if 2 people are tied for a spot, shouldn't it go to head to head matchup(to see who won the game, if they played against each other earlier)?

 

[SteveP]

jdb, the math for the tiebreakers and the math for the ELO rankings are different. I can't explain it any better. For example, let's say that after 6 rounds, 1 player is undefeated and 3 players are 5-1. Assuming that all players started at the same ELO base ranking of 1600, the 6-0 player will have the highest ELO ranking. In round 7, the undefeated player (Player A) plays a 5-1 player (Player B), someone with a lower ELO ranking, and looses. The ELO system rewards the underdog (Player B) and penalizes the undefeated player (Player A) more drastically than if 2 undefeated players had played in round 7. That means there's a significant chance that the underdog's ranking (Player B) shot past the undefeated player's ranking (Player A). I assume that's close to what happened at your tournament. Now, if the 7th-round winner (Player B) had lost earlier to someone who went 5-2 (or worse) in the tournament, then the 7th-round loser (Player A) would have a higher tiebreaker, thereby having a higher tournament ranking. However, earlier loses are NOT taken into consideration in the ELO ranking system.

Just remember this general rule. Loosing late in a tournament (when your ELO ranking is higher) results in a more drastic decline in your ELO ranking than if you'd lost earlier (when your ELO ranking is lower). However, everyone who has played in a swiss tournament knows that it's better to loose late in a tournament than to loose early, that is, if you want to finish higher in the tournament. That may seem odd, but that's just how it works.

Hey, my son once had a 2100+ DCI ranking. When he won games locally, he usually only got 1-2 DCI points per win. But, when he lost, he'd often loose 10-20 DCI points per lose. At one point, my son had to win at least 9 out of 10 games to keep his ranking from dropping.

Welcome to world of the ELO ranking system.

 

[jdb728]

SteveP, I realize where you're comin' from, but the thing is, there weren't any Worldwide rankings for any of these players, seeing as how none of them played in any sanctioned tourneys(with nintendo) before the Dragon Prerelease, so the whole point about previous rankings is flawed(IMO).
I realize the tiebreakers may be different(I still think it should be head to head matchup as the first tiebreaker).
Thanks for reading.

 

[NoPoke]

Well, don't you think that the Tourney rankings system, and the world wide rankings systems should work together, since I mean, does it make a whole lot of sense for the tourney software to say that "Player 1" did better then "Player2" in the tourney, but when it comes to Global rankings, because a different system is used, according to it, "Player 2" did better then "Player 1"?
I don't know, it may just be me, but shouldn't they rank players both via the same system?

...................)?

The ELO rating system takes no account of awarded BYES. Yet a BYE is a full win in a tournament. In a short tournament first round byes do influence the final standings. The discrepancy is inherent and unavoidable. That said it is not unreasonable to expect that the tournament winner will have a higher ELO ranking than lower placed players who also started from a similar ELO rating. There are several ways in which this discrepancy can ocurr. The ELO system is non-linear while the tournament points system is perfectly linear and there is the issue with byes.

You would have to provide all the tournament data for anaylsis before it can be established if the result is in error. Until the tournament is examined in detail we are stuck with speculation as to why the tournament winner ended up with a lower ELO rating. Personally I'll speculate that it is down to BYEs.

 

[old man]

... but the thing is, there weren't any Worldwide rankings for any of these players, seeing as how none of them played in any sanctioned tourneys(with nintendo) before the Dragon Prerelease, so the whole point about previous rankings is flawed(IMO).

Just so you understand, A player enter a POP tournament for the first time starts with a rating, (score) of 1600. They do not have had to play in any other tournaments. This also applies to each catergory of tournament, ie. limited, open & modified.

 

[NoPoke]

Doesn't look like it is down to BYEs

If I've understood the Portland tournament correctly all the players in question Won 6 and lost 1. With a K=32 tournament this would give each player an ELO rating of 1680 give or take a couple of points. Which appears to be the case. So there is nothing obviously wrong with the result. The only way to be certain is to carry out the full rating system for the tournament. Think I'd better get back to that little piece of database software I was working on

There are some discrepancies in the ratings pages... if you examine the unlimited ratings and pick a top player you are given two different ELO ratings.

 

[SteveP]

jdb, like oldman said, you DO start with a ranking, even though you've never played before. Plus, ELO rankings are recalculated EACH TIME you play a game. They DON'T take into account any previous games like the tiebreaker system does. Plus, like NoPoke said, byes are NOT calculated into the ELO ranking like they are for the tiebreaker system.

The ELO system is NOT flawed. That's how it works. After looking at your Oregon tournament online, the player who ended up with the highest ELO ranking, but finished lower in the tournament rankings, did EXACTLY as my example.

 

[jdb728]

jdb, like oldman said, you DO start with a ranking, even though you've never played before. Plus, ELO rankings are recalculated EACH TIME you play a game. They DON'T take into account any previous games like the tiebreaker system does. Plus, like NoPoke said, byes are NOT calculated into the ELO ranking like they are for the tiebreaker system.

The ELO system is NOT flawed. That's how it works. After looking at your Oregon tournament online, the player who ended up with the highest ELO ranking, but finished lower in the tournament rankings, did EXACTLY as my example.

As far as I know there weren't any byes in the tourney(But I very well could be wrong), and I don't remember sayin' it's flawed, I just don't think it makes very much sense for them to use different systems for ranking(That's just IMO, you're free to disagree if you wish).

 

[ScythKing]

jdb, like oldman said, you DO start with a ranking, even though you've never played before. Plus, ELO rankings are recalculated EACH TIME you play a game. They DON'T take into account any previous games like the tiebreaker system does. Plus, like NoPoke said, byes are NOT calculated into the ELO ranking like they are for the tiebreaker system.

The ELO system is NOT flawed. That's how it works. After looking at your Oregon tournament online, the player who ended up with the highest ELO ranking, but finished lower in the tournament rankings, did EXACTLY as my example.

While the ELO system may not be flawed - there are issues.
Jeremey is right - I looked at the ratings for the players at my pre-release and I have trouble understanding the ratings - actually I sorta understand but sorta don't. Here's an example:

Rashad Rizer - played one round, lost, and then had to drop. Ranked 1584.
Ashle Graves - played two rounds lost both, then dropped. Ranked 1570.
Zachary Wilson - 0-3-0-0. Ranked 1568.

These guys are all in the same age group. This certainly shows that you should be a quitter if you don't think you can hack it. If you persevere you'll pay.
Great lesson....

 

[GordonKane]

I need to find out from Jimmer if this is the same software he distributed to the Magi-Nation Directors (when Jimmer worked for 2i).
The double bye issue was present in that version too (as well as a bad habit of repairing the same people around rounds 3, 4, & 5 that had the lowest scores and had met in a prior round -- and they get repaired (especially bad when the people getting repaired don't say anything).
Does the software generate *.TRN files to be uploaded to the website? If so, it probably is the same software.

I know the Tourney Finder / Player Finder has the same User Interface as the one on the Magi-Nation site.

If the programs are the same, I should be able to help ya'll troubleshoot some issues (since I've been using the software since July 2001).

 

[ScythKing]

I need to find out from Jimmer if this is the same software he distributed to the Magi-Nation Directors (when Jimmer worked for 2i).
The double bye issue was present in that version too (as well as a bad habit of repairing the same people around rounds 3, 4, & 5 that had the lowest scores and had met in a prior round -- and they get repaired (especially bad when the people getting repaired don't say anything).
Does the software generate *.TRN files to be uploaded to the website? If so, it probably is the same software.

I know the Tourney Finder / Player Finder has the same User Interface as the one on the Magi-Nation site.

If the programs are the same, I should be able to help ya'll troubleshoot some issues (since I've been using the software since July 2001).

Same stuff. Even says licensed from Interactive-Imagination, the Magi-Nation folks (which I like that game BTW). My problem seems to revolve around two issues - it will allow duplicate ID's and when a player drops it shows him in the list as dropped but occupying a table which apparently causes valid players to be "dropped" from the list.

 

[SteveP]

Scythking, so, what are you saying? You're ranking should go down for games you didn't play in a tournament (because you dropped or something)?

Personally, I'd like to see a tournament scoring system like Decipher's where you get a point for just playing when you loose. But then again, you can't draw in Decipher games.

 

[DaytonGymLeader]

Personally, I'm to the point of calling the software "Intollerable." We ran in to the Bye issue again today. The top player in 15+ gets a Bye, then doesn't get the points. Kept him out of first place. Not kosher under any respect. All my other Bye players did receive 3 points. Any idea if there is a "Special Editor" version of this software that will let you tinker with tournament values?

 

[farbsman]

After round 5 of my tournament the software decided to stop working. This caused all kinds of delays and problems with people not wanting to wait for pairings or results. A lot of parents and players left mad because of it. I know however why the software had problems. When I did pairings for round 6, 2 players with the same last name ended up getting combined into one player and the other player was dropped out of the tournament. Luckily the judges that were working the event for me were able to do everything the long way to at least get the tournament finished. It took 20-30 minutes to do pairings and almost 45 minutes to figure out standings the long way.

In my opinion, if Nintendo and Pokemon USA paid money for this program, they got ripped off. The people that programed it should be fired. I will not run another tournament with this software until it is corrected or replaced. Parents getting mad at you and players quiting the game because software doesn't work is nuts. If you can't get the software right, it shouldn't have been released. My question is, why didn't Nintendo hire the same people that did DCI reporter to make there program????

 

[jdb728]

After round 5 of my tournament the software decided to stop working. This caused all kinds of delays and problems with people not wanting to wait for pairings or results. A lot of parents and players left mad because of it. I know however why the software had problems. When I did pairings for round 6, 2 players with the same last name ended up getting combined into one player and the other player was dropped out of the tournament. Luckily the judges that were working the event for me were able to do everything the long way to at least get the tournament finished. It took 20-30 minutes to do pairings and almost 45 minutes to figure out standings the long way.

In my opinion, if Nintendo and Pokemon USA paid money for this program, they got ripped off. The people that programed it should be fired. I will not run another tournament with this software until it is corrected or replaced. Parents getting mad at you and players quiting the game because software doesn't work is nuts. If you can't get the software right, it shouldn't have been released. My question is, why didn't Nintendo hire the same people that did DCI reporter to make there program????

Just a quick question, would it be allowed for a TO to use DCI reporter while running a sanctioned event, or do you have to use the messed up Interac, umm... Nintendo one?

 

[farbsman]

Just a quick question, would it be allowed for a TO to use DCI reporter while running a sanctioned event, or do you have to use the messed up Interac, umm... Nintendo one?

We are required to use the one nintendo supplies. It is in the Premier TO contract. Not sure about other tournaments though.

 

[Pokeplayer]

Well,

I will have to admit that from all the lessons learned to date about the problems with the TMS software, I was able to complete my main tournament and side tournament without a single program crash...

I made sure that I typed in players names and last name with out any numbers or dashes or anything but letters. I saved it constantly...

I used the Alt Print screen, pasted the image into Paint/Bitmap, erased the parts of the screen I did not want, and printed it out on my printer. This took about five - ten minutes to do as I had to copy the screen three times because the pairings window in TMS would not get larger when I attempted to make it bigger)

Overall it went well.....However, I woudl like to see it improved to print pairings, eliminate program crash for typing in numbers/forget to type in last names.........

See Ya.....

 

[jdb728]

We are required to use the one nintendo supplies. It is in the Premier TO contract. Not sure about other tournaments though.

Another quick question, if you're required to use the software, and then it crashes, and you have to do the final couple rounds without it, are the results of the tournament void(since obviously the software wasn't used through the whole thing)?