Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

V:TES ranking/rating/event calendar website remodeling project started

9 views
Skip to first unread message

Johannes Walch

unread,
Jan 10, 2010, 11:24:46 AM1/10/10
to
Our little development group of volunteers has started the project of
programming an independent ranking/rating site.

The goal of the project is to create a *working* system, especially with
the feature to upload Archon files without doing a lot of manual
entering. It will run on a different domain (name TBD) and the
White-Wolf page will just link to it once we are done and have imported
the old data.

If there are more volunteers familiar with web programming or HTML
design send me an e-mail and I give you directions how to access our
development forum.

Once we have a working prototype we will do a public beta test where all
of you can check the site out and give us input for improvement.

Johannes Walch
European V:TES Coordinator

_angst_

unread,
Jan 10, 2010, 4:19:44 PM1/10/10
to

<3 :)

Emiliano Imeroni

unread,
Jan 11, 2010, 2:03:11 AM1/11/10
to
On 10 Gen, 17:24, Johannes Walch <johannes.wa...@vekn.de> wrote:
> Our little development group of volunteers has started the project of
> programming an independent ranking/rating site.
>
>
> Johannes Walch
> European V:TES Coordinator

Great news!

Salem

unread,
Jan 11, 2010, 3:07:50 AM1/11/10
to
Johannes Walch wrote:
> Our little development group of volunteers has started the project of
> programming an independent ranking/rating site.
>
> The goal of the project is to create a *working* system, especially with
> the feature to upload Archon files without doing a lot of manual
> entering. It will run on a different domain (name TBD) and the
> White-Wolf page will just link to it once we are done and have imported
> the old data.

are you going to use the current rating/ranking formulas, or something new?

either way, this is great news :)

can you also make it so you don't need to have linked your vekn id to a
WW id? i don't think any of my players ever got a ww id, and we pretty
much stopped submitting results around the time that became necessary, too.

--
salem
(replace 'hotmail' with 'gmail' to email)

James Coupe

unread,
Jan 11, 2010, 4:00:23 AM1/11/10
to
Salem <kell...@hotmail.com> wrote:
>Johannes Walch wrote:
>> Our little development group of volunteers has started the project of
>>programming an independent ranking/rating site.
>> The goal of the project is to create a *working* system, especially
>>with the feature to upload Archon files without doing a lot of manual
>>entering. It will run on a different domain (name TBD) and the White-
>>Wolf page will just link to it once we are done and have imported the
>>old data.
>
>are you going to use the current rating/ranking formulas, or something new?
>
>either way, this is great news :)

From what Johannes has said - here and elsewhere - the initial intention
is to provide a 'drop-in' replacement (except it works better), so I
imagine it would be keeping the existing formulae.

However, assuming the raw data is kept in an SQL database of some sort,
you can reasonably easily start constructing other searches for fun
stats (e.g. which players have the most VPs over the 18 months) or
trying out new formulae for improved rankings.

--
James Coupe
PGP Key: 0x5D623D5D YOU ARE IN ERROR.
EBD690ECD7A1FB457CA2 NO-ONE IS SCREAMING.
13D7E668C3695D623D5D THANK YOU FOR YOUR COOPERATION.

Johannes Walch

unread,
Jan 11, 2010, 4:27:19 AM1/11/10
to Salem
Salem schrieb:

> Johannes Walch wrote:
>> Our little development group of volunteers has started the project of
>> programming an independent ranking/rating site.
>>
>> The goal of the project is to create a *working* system, especially
>> with the feature to upload Archon files without doing a lot of manual
>> entering. It will run on a different domain (name TBD) and the
>> White-Wolf page will just link to it once we are done and have
>> imported the old data.
>
> are you going to use the current rating/ranking formulas, or something new?

The current formula, nothing wrong with that.

> can you also make it so you don't need to have linked your vekn id to a
> WW id? i don't think any of my players ever got a ww id, and we pretty
> much stopped submitting results around the time that became necessary, too.

Yes, there will be no need to have a WW id then. The identifier will be
the VEKN id of the player. Of course princes will have their prince id
along with a log-in to enter and report events.

Forhead

unread,
Jan 11, 2010, 6:36:26 AM1/11/10
to

This is fantstic. You should have all the cred in the world!

Isak Bjärmark Esbjörnsson
Prince of Örebro

Smiling Jack XXX

unread,
Jan 11, 2010, 10:28:21 AM1/11/10
to

> Yes, there will be no need to have a WW id then. The identifier will be
> the VEKN id of the player. Of course princes will have their prince id
> along with a log-in to enter and report events.


Of course, BUT what about the players who are not princes and register
tournaments? How will they report their results?

Brum

unread,
Jan 11, 2010, 11:05:22 AM1/11/10
to

WW likes it's pyramid.
In Lisbon we are 1 Prince and 3 Primogen. Sometimes one of us does the
demos, another judges, another gets the venue, another organizes the
tournaments, etc.
We have been more or less equal, but the Prince has special tasks that
the others cannot do, namely the WW interaction and results posting.
I think it's important for WW to keep some control over who can do
qualifying tournaments, releases, etc.

My advice is this: apply to be a Prince if you don't have one handy.

Tiago

Johannes Walch

unread,
Jan 11, 2010, 12:44:30 PM1/11/10
to
Brum schrieb:

That�s it. Up till today only princes can register tournaments and thats
the way it will be. You can of course share the login/password (which is
only used for that) with your helping hands if need be (up to your local
way of running things).

Johannes

Juggernaut1981

unread,
Jan 11, 2010, 4:47:51 PM1/11/10
to

Do what we've done before...
Run the tourney then email/call a Prince somewhere else and go "Here
are the results dude, can you upload? Thanks"

Johannes Walch

unread,
Jan 11, 2010, 4:56:32 PM1/11/10
to
Forhead schrieb:
> Isak Bj�rmark Esbj�rnsson
> Prince of �rebro

I am just putting together the people.
Also be aware that this will probably take some months to complete.

Johannes Walch

unread,
Jan 11, 2010, 4:57:27 PM1/11/10
to
Juggernaut1981 schrieb:

Especially since now it is really only upload and not enter the results
in stupid form for an hour.

Frederick Scott

unread,
Jan 11, 2010, 6:16:56 PM1/11/10
to
"Johannes Walch" <johanne...@vekn.de> wrote in message
news:4B4AEEF7...@vekn.de...

> Salem schrieb:
>> Johannes Walch wrote:
>>> Our little development group of volunteers has started the project of programming an independent ranking/rating site.
>>>
>>> The goal of the project is to create a *working* system, especially with the feature to upload Archon files without doing a lot
>>> of manual entering. It will run on a different domain (name TBD) and the White-Wolf page will just link to it once we are done
>>> and have imported the old data.
>>
>> are you going to use the current rating/ranking formulas, or
>> something new?
>
> The current formula, nothing wrong with that.

Of course there is - the thing that I've point out all along:

1) Attendence matters significantly to the formula.
2) Attendence != skill.

Ergo, 3) The numbers don't measure skill.

Although there might be a very rough relationship between this formula and
skill, it's easy to suspect that skilled players who aren't prolific at
attending VTES tournaments are vastly underrated and therefore at least
some who have attended many tournaments may be somewhat overrated by
comparison. One prime example is Brian Moritz in the U.S. Northwest, for
whom the challenge of attending tournaments on a level that most European
players can is vastly higher.

We've had the philosophical argument about why this silly system needs to
be what it is. The real reason is that a true skill rating system of
any sort would be much more difficult to implement and it's fair to say
that any such system would still likely have problems representing skill
as an accurate number. That's fine. But please don't make such
statements, "...nothing wrong with the current formula". For my money,
there's practically nothing right with it.

Fred


Salem

unread,
Jan 12, 2010, 5:57:53 AM1/12/10
to

or just apply to be a prince yourself. it's not like it costs anything
or legally obliges you to do anything. if there's already a prince in
your area, just append some compass point to the city name in your
application, or choose a sub-urban section. eg: apply to be prince of
North Boston, or prince of the Bronx, or something.

Johannes Walch

unread,
Jan 12, 2010, 7:44:40 AM1/12/10
to
Frederick Scott schrieb:

What I meant is that the general unhapiness with the current system is
not about formula but rather about the non-functional website. Most
players seems to be at least ok with the way the formula works.

The system is very similiar to the Tennis ATP System (which was a big
inspiration), lacking the factoring of opponents. And there it is the
same, if you play no Grand Slams you are not going to be #1. To be on
top of the list you have to play the big events, if they are hard to
attend to for you (not Fred, generally) personally that is bad luck. You
either need to move to Europe or skip ambitions about being #1 ;-)

Johannes Walch

unread,
Jan 12, 2010, 7:51:17 AM1/12/10
to
Frederick Scott schrieb:

> Of course there is - the thing that I've point out all along:
>
> 1) Attendence matters significantly to the formula.
> 2) Attendence != skill.
>
> Ergo, 3) The numbers don't measure skill.

In a 10 player tournament you have a statistical chance of 50% to go to
the finals. In a 100 player tournament your chance is 5%. Being Top 5%
requires more skill than being Top 50%. A player who is consistently
going to Top 5% is obviously a great player. If you can only play 10
player tournaments due to personal reasons (location etc..) and you
always are in the final you are a regular Top 50%. You might be a great
player but there is no way to tell since your competition does not allow
for correct analysis. If you win every one of them that is a better
indication, but then your rating will be quite good even with small
tournaments. But that is very unlikely given the dynamics of a final.

Another way to argue is that in large tournaments the challenge is to go
into the final, in small tournaments the challenge is to win the final.
Almost impossible to put this into a fair "rating" system since
priorities are different.

Archibald Zimonyi

unread,
Jan 12, 2010, 8:46:33 AM1/12/10
to
Why does not White Wolf update their site? Any reason for this? No
interest? No money?

Archie

Johannes Walch

unread,
Jan 12, 2010, 9:03:25 AM1/12/10
to
Archibald Zimonyi schrieb:

> Why does not White Wolf update their site? Any reason for this?

There is always a reason.

> No interest? No money?

A combination plus No time/resources I�d say.

Meej

unread,
Jan 12, 2010, 10:27:17 AM1/12/10
to
On Jan 12, 7:51 am, Johannes Walch <johannes.wa...@vekn.de> wrote:
> Frederick Scott schrieb:
>
> > Of course there is - the thing that I've point out all along:
>
> > 1) Attendence matters significantly to the formula.
> > 2) Attendence != skill.
>
> > Ergo, 3) The numbers don't measure skill.
>
> In a 10 player tournament you have a statistical chance of 50% to go to
> the finals. In a 100 player tournament your chance is 5%. Being Top 5%
> requires more skill than being Top 50%.

That depends on who your competition is. If your local playgroup
would at least half be in the top 5% in many major tournaments of that
size, but it's only 10 people, you're still going to be ranking in
that caliber of players.

> A player who is consistently
> going to Top 5% is obviously a great player.

Not much argument there - it'd be very hard to amass a 100-player
grouping that's filled with enough lousy players to make that
statement false. It's the other end of the spectrum that has
problems.

> If you can only play 10
> player tournaments due to personal reasons (location etc..) and you
> always are in the final you are a regular Top 50%. You might be a great
> player but there is no way to tell since your competition does not allow
> for correct analysis.

If the system included the rankings of your opponents into the
factoring, it could.

An example: I play in Boston. Now, I'm not saying I'm particularly
great. But the local playgroup, for which we regularly get up to only
about 10 players for most tournaments (despite the valiant efforts of
our Prince, who's doing all he can to promote things), is the home
playgroup of Ben Peal, Ben Swainbank, and Matt Hirsch (all in the Hall
of Fame), we frequently see Jesse Cross-Nickerson (USNC 2009), and Jon
Scherer (USNC 2008) used to be up hereabouts. The competition is stiff
around here. Consistently getting into the finals of those small
tournaments is no easy feat; consistently anything is no easy feat.
There are folks around here who frequently do well against top-tier
players but can't make it to enough major tournaments for the system
to gauge their skill based on number of opponents as well.

That's what Fred is talking about, from what I can tell.

> If you win every one of them that is a better
> indication, but then your rating will be quite good even with small
> tournaments. But that is very unlikely given the dynamics of a final.

*Nod* very unlikely.

Frederick Scott

unread,
Jan 12, 2010, 4:30:09 PM1/12/10
to
"Johannes Walch" <johanne...@vekn.de> wrote in message
news:hihqro$j8o$1...@news01.versatel.de...

> Frederick Scott schrieb:
>> "Johannes Walch" <johanne...@vekn.de> wrote in message
>> news:4B4AEEF7...@vekn.de...
>>> Salem schrieb:
>>>> are you going to use the current rating/ranking formulas, or
>>>> something new?
>>>
>>> The current formula, nothing wrong with that.
>>
>> Of course there is - the thing that I've point out all along:
>>
>> 1) Attendence matters significantly to the formula.
>> 2) Attendence != skill.
>>
>> Ergo, 3) The numbers don't measure skill.
>>
>> Although there might be a very rough relationship between this formula and
>> skill, it's easy to suspect that skilled players who aren't prolific at
>> attending VTES tournaments are vastly underrated and therefore at least
>> some who have attended many tournaments may be somewhat overrated by
>> comparison. One prime example is Brian Moritz in the U.S. Northwest, for
>> whom the challenge of attending tournaments on a level that most European
>> players can is vastly higher.
...

> What I meant is that the general unhapiness with the current system is
> not about formula but rather about the non-functional website. Most
> players seems to be at least ok with the way the formula works.

OK, I guess that's fair enough as a clarification for purposes of this
thread.

> The system is very similiar to the Tennis ATP System (which was a big inspiration), lacking the factoring of opponents. And there
> it is the same, if you play no Grand Slams you are not going to be #1. To be on top of the list you have to play the big events,
> if they are hard to attend to for you (not Fred, generally) personally that is bad luck.

As I've pointed out many times in discussions about this issue, comparing
a professional sport and an amateur tabletop game in this respect is
thoroughly flawed. If a person wants to become a million-dollar-a-year
famous professional athlete, sure, stirring attendance into the system is
fine. Issues like travel cost and time spent traveling to and playing
tournaments are fairly meaningless in _THAT_ context. It's fair to assume
that anyone who truly cares about their ATP rankings will attend as many
tournaments as possible. Given that they do, the influence of attendance
is canceled out and only skill is really represented.

But get a grip! This is an amateur game and, as such, it's not sane to
fly all over the world to attend tournaments. A few people do some of that
but the vast majority have modest limits to the travel they're able and/or
willing to do in order to attend tournaments. A rating system which takes
this into account would be much more reasonable for the vast majority of
the players. Having "bad luck" significantly influence the rating system
makes the rating system bad.

Fred


Frederick Scott

unread,
Jan 12, 2010, 4:45:44 PM1/12/10
to

"Johannes Walch" <johanne...@vekn.de> wrote in message news:hihr85$jef$1...@news01.versatel.de...

There are a couple of flaws in this reasoning, taken as an attempt to
defend to the existing system as accurate:

1) No one said you had to use tournament results as your indicator (or
only indicator) of skill. One past system used only games. And games
are always games - the same everywhere, no matter the size of the tournaments
in which they're held. Opponent skills may differ in a given game; the
solution was to correct for skill by using the opponents' existing ratings
to calculate new ratings. (But, of course, this is what made that system
so difficult to implement in a process: order of games played mattered.)
Of course, even when comparing a 50-player tournament to another 50-player
tournament, the skills in the opposing players may also differ greatly (for
instance, in different parts of the world). This, in fact, is another flaw
in the existing system I didn't even bring up. The existing system assumes
all tournaments of a given size are exactly valued exactly the same, which
is a poor assumption in some cases.

2) If you must use tournament results as part or all of your rating system,
the value of the finish position can be adjusted to the size of the tournament.
This is in fact what the existing system does. The problem (apart from the
possibility that the existing system values the finishing positions of the
various-sized tournaments incorrectly) is that it treats the output scores as
something to add to the players' ratings rather than to average them into the
ratings in a way which is attendance-neutral (or as attendance-neutral as
possible).

Fred


librarian

unread,
Jan 12, 2010, 4:57:04 PM1/12/10
to
Frederick Scott wrote:
> "Johannes Walch" <johanne...@vekn.de> wrote in message news:hihr85$jef$1...@news01.versatel.de...
>> Frederick Scott schrieb:

>

> 1) No one said you had to use tournament results as your indicator (or
> only indicator) of skill. One past system used only games. And games
> are always games - the same everywhere, no matter the size of the tournaments
> in which they're held. Opponent skills may differ in a given game; the
> solution was to correct for skill by using the opponents' existing ratings
> to calculate new ratings. (But, of course, this is what made that system
> so difficult to implement in a process: order of games played mattered.)
>

I never understood why this was so hard to implement. Just
have a date field in the database, and as results come in,
adjust everything appropriately.

And maybe have a cut-off date for when results can be entered.

best -

chris

The Lasombra

unread,
Jan 12, 2010, 5:49:48 PM1/12/10
to
On Tue, 12 Jan 2010 05:46:33 -0800 (PST), Archibald Zimonyi wrote:

>Why does not White Wolf update their site? Any reason for this? No
>interest? No money?

No response from players.

See also:
http://groups.google.com/group/rec.games.trading-cards.jyhad/msg/89c45da151e95d09

Frederick Scott

unread,
Jan 12, 2010, 7:25:04 PM1/12/10
to
"librarian" <auct...@superfuncards.com> wrote in message
news:hiir7l$sr7$1...@news.eternal-september.org...

Spot on! In fact, I believe I suggested something very similar to this,
once upon a time when we were still debating the merits of the various system.
It would still require a lot more processing time to back up and recalculate
things whenever a severely post-dated report came in. But given the advances
in computer performance these days, I doubt it matters any more. And to
whatever extent it does, the problem diminishes every day.

> And maybe have a cut-off date for when results can be entered.

Right. So, as time passes, you could discard individual reports if you liked.
Whether you choose to or not, at least this would limit the amount of
reprocessing.

Fred


Johannes Walch

unread,
Jan 18, 2010, 3:52:16 AM1/18/10
to
Frederick Scott schrieb:

> As I've pointed out many times in discussions about this issue, comparing
> a professional sport and an amateur tabletop game in this respect is
> thoroughly flawed. If a person wants to become a million-dollar-a-year
> famous professional athlete, sure, stirring attendance into the system is
> fine. Issues like travel cost and time spent traveling to and playing
> tournaments are fairly meaningless in _THAT_ context. It's fair to assume
> that anyone who truly cares about their ATP rankings will attend as many
> tournaments as possible. Given that they do, the influence of attendance
> is canceled out and only skill is really represented.
>
> But get a grip! This is an amateur game and, as such, it's not sane to
> fly all over the world to attend tournaments. A few people do some of that
> but the vast majority have modest limits to the travel they're able and/or
> willing to do in order to attend tournaments. A rating system which takes
> this into account would be much more reasonable for the vast majority of
> the players. Having "bad luck" significantly influence the rating system
> makes the rating system bad.

I get you point, but I looked around a bit and could not find a system
which does take into account what you describe. Do you have any hint
where to copy or which formula could be used?

Johannes Walch

unread,
Jan 18, 2010, 4:46:29 AM1/18/10
to
Frederick Scott schrieb:

> One past system used only games.

I hope you are not implying by this that the old system was better than
the new one. The old one was seriously flawed (e.g where you could win
the largest tournament ever and your rating would actually go DOWN, this
has happened). Probably it was more fair in terms of attendance (how
many and big tournaments you can attend) but it was pretty easy to get
into the Top10 by playing some local newbie tournaments and winning all
the games. I think we can all agree that the Top10 (even 50) of the
current system has only decent players in it. Also it was punishing you
for having bad results in a tournament which I don´t like. This means if
you want to have top rating you can´t play a tournament once in a while
with a fun deck (because you have 8 good ones in the rating allready)
you gotta bring those tried and true archetypes all the time. Makes the
game very boring.

> And games
> are always games - the same everywhere, no matter the size of the tournaments
> in which they're held.

Not true IMHO. In very large tournaments (e.g EC) the competition gets
very stiff, independently from the individual opponents. Everybody
(well, almost, there are always a few guys playing for style) figures
that to go into that final you have to bring something top notch. So you
are not about to see the very experimental decks you tend to see in
local 10 Player tournaments, you will sometimes face 5 Tier-1 decks on
the same table. Hard to win those games, I tell you. When you are an
experienced player and bring a semi-broken archetype into a local 10
Player newbie tournament you are bound to win 3 tables with it.

Frederick Scott

unread,
Jan 22, 2010, 8:13:18 PM1/22/10
to
"Johannes Walch" <johanne...@vekn.de> wrote in message
news:hj17g0$vj3$1...@news01.versatel.de...
> Frederick Scott schrieb:

>> This is an amateur game and, as such, it's not sane to
>> fly all over the world to attend tournaments. A few people do some of that
>> but the vast majority have modest limits to the travel they're able and/or
>> willing to do in order to attend tournaments. A rating system which takes
>> this into account would be much more reasonable for the vast majority of
>> the players. Having "bad luck" significantly influence the rating system
>> makes the rating system bad.
>
> I get you point, but I looked around a bit and could not find a system which does take into account what you describe. Do you have
> any hint where to copy or which formula could be used?

I don't recall if you were around when we were having the discussions about
rating, ranking, the "rewards system" (i.e. the current system), and all that
stuff. Basically, the one I know about is the original Elo system that VEKN
originally copped from chess qualifies. It was a system for constantly
recalculating each player's rating after each game from: 1) his previous rating,
2) the rating of the other players who played the game, and 3) the outcome of
the game. It qualifies because it doesn't matter that much how active a player
is. But had its own drawbacks:

1) Elo is a system intended for rating two-player games. There's different
ways around this, but they're kind of complex and arguably artificial. The
original Elo system used in VEKN compared the outcome of each pair of players
in a game separately and held that one player "beat" the other player if he
finished that game with more VPs than the other player (or they drew if they
finished with the same number of VPs). Hence if you played in a five-player
game, it was calculated as if you'd played four simultaneous games.

2) In order to make sense of rating people like this, there has to be an
artificially chosen coefficient to weight how much of your new rating was
derived from your previous rating and how much was derived from the results
of the latest game. In the original ELO system, the coefficient VEKN used
was terrible! It weighted the most recent results far too heavily and hence,
everyone's ratings constantly bounced all over the place. In the course of a
couple tournaments, you could go from a VTES god to a dithering idiot. That
could have been fixed by choosing a better coefficient.

3) It matters in which order the results are entered. At the time, the only
solution was seen to be, "We can't update any ratings beyond the point of the
first missing report." Since then, as now, missing reports were a common
occurrence, this made the system problematic to administer. As Chris points
out, you can actually calculate as if the missing tournaments never happened
and then just go back and recalculate results from that point if a tournament
result from the past suddenly shows up. But the software would have to be
written to support this and it might take a fair amount of time to recalculate
something from several weeks back, depending on the computer, the software tools
used to write the software, and the skill of the programmer himself. Including
a cutoff date after which results will no longer be accepted would be helpful -
and I believe one already exists in the extant rating system.

4) For a number of reasons, the rating still wouldn't be that accurate:
- There's lots of luck in VtES: card drawing, metagame luck, politics,
opponents choosing to play fun decks vs. killer decks. I'm sure people
could think of others.
- The ratings of my opponents used to recalculate my rating might not be
very accurate at any given time. For instance, new players must be
given arbitrary "average" ratings (a number in the exact middle of the
pack) that they may not deserve. Or a player who hasn't played in three
years might have a higher rating than he can support at the current
time due to changes in the metagame, forgetting the skills, etc.

The first set of reason apply to the existing system. The second set do not.

With its various drawbacks, I'm still sure Elo would ultimately be a much more
accurate system, at least for players who have a fair number of games under
their belt. It would clearly take a lot of work to impose, however.

Fred


Frederick Scott

unread,
Jan 22, 2010, 8:22:21 PM1/22/10
to
"Johannes Walch" <johanne...@vekn.de> wrote in message
news:hj1alm$15q$1...@news01.versatel.de...

> Frederick Scott schrieb:
>> One past system used only games.
>
> I hope you are not implying by this that the old system was better than
> the new one. The old one was seriously flawed (e.g where you could win
> the largest tournament ever and your rating would actually go DOWN, this
> has happened).

See my comments about the coefficient issue in my other post. Basically,
yes, you're right that it old system was seriously flawed. The flaw you're
talking about could have been easily fixed. Why no one ever saw the problem
and fixed it totally confuses me. But whatever the case, it is not an issue.

>> And games
>> are always games - the same everywhere, no matter the size of the tournaments
>> in which they're held.
>
> Not true IMHO. In very large tournaments (e.g EC) the competition gets
> very stiff, independently from the individual opponents.

The fact that the EC competition gets stiff is a function of how much practice -
in or out of tournaments - the players get. You could create a 110-player
tournament in an Alaskan village with 110 newly taught players and the point
rewards would be just as high. I gurantee you, the competition wouldn't be
very good.

That large tournaments tend to attract excellent players is not, of course,
completely coincidental. At the same time, there's nothing which requires
any connection between the two and so it's a bad idea to assume there
is. At the same time, my point is not to ignore quality of competion (which
the current formula certainly does); it's to take it into account in
recalculating ratings.

Fred


JH

unread,
Jan 23, 2010, 4:41:42 AM1/23/10
to
On Jan 23, 3:13 am, "Frederick Scott" <nos...@no.spam.dot.com> wrote:
> With its various drawbacks, I'm still sure Elo would ultimately be a much more
> accurate system, at least for players who have a fair number of games under
> their belt.  It would clearly take a lot of work to impose, however.
>
> Fred

If a player would disappear from the scene and later reappear,
wouldn't a regular ELO system preserve his rating throughout this time?

Johannes Walch

unread,
Jan 23, 2010, 7:27:43 AM1/23/10
to Frederick Scott
Frederick Scott schrieb:

> "Johannes Walch" <johanne...@vekn.de> wrote in message
> news:hj1alm$15q$1...@news01.versatel.de...
>> Frederick Scott schrieb:
>>> One past system used only games.
>> I hope you are not implying by this that the old system was better than
>> the new one. The old one was seriously flawed (e.g where you could win
>> the largest tournament ever and your rating would actually go DOWN, this
>> has happened).
>
> See my comments about the coefficient issue in my other post. Basically,
> yes, you're right that it old system was seriously flawed. The flaw you're
> talking about could have been easily fixed. Why no one ever saw the problem
> and fixed it totally confuses me. But whatever the case, it is not an issue.
>
>>> And games
>>> are always games - the same everywhere, no matter the size of the tournaments
>>> in which they're held.
>> Not true IMHO. In very large tournaments (e.g EC) the competition gets
>> very stiff, independently from the individual opponents.
>
> The fact that the EC competition gets stiff is a function of how much practice -
> in or out of tournaments - the players get. You could create a 110-player
> tournament in an Alaskan village with 110 newly taught players and the point
> rewards would be just as high. I gurantee you, the competition wouldn't be
> very good.

No you couldn´t and that is the point. If a large number of people come
together in V:TES it is a difficult event, it has always been that way.

Frederick Scott

unread,
Jan 23, 2010, 11:58:46 AM1/23/10
to
"JH" <jhat...@tenerdo.org> wrote in message
news:b7f0bd5a-d587-4468...@e37g2000yqn.googlegroups.com...

On Jan 23, 3:13 am, "Frederick Scott" <nos...@no.spam.dot.com> wrote:
> > With its various drawbacks, I'm still sure Elo would ultimately be a much more
> > accurate system, at least for players who have a fair number of games under
> > their belt. It would clearly take a lot of work to impose, however.
>
> If a player would disappear from the scene and later reappear,
> wouldn't a regular ELO system preserve his rating throughout this time?

Sure - for better or worse.

Fred


Frederick Scott

unread,
Jan 23, 2010, 12:07:33 PM1/23/10
to

"Johannes Walch" <johanne...@vekn.de> wrote in message
news:4B5AEB3F...@vekn.de...

>>>> And games are always games - the same everywhere, no matter
>>>> the size of the tournaments in which they're held.
>>>
>>> Not true IMHO. In very large tournaments (e.g EC) the competition gets
>>> very stiff, independently from the individual opponents.
>>
>> The fact that the EC competition gets stiff is a function of how much practice -
>> in or out of tournaments - the players get. You could create a 110-player
>> tournament in an Alaskan village with 110 newly taught players and the point
>> rewards would be just as high. I gurantee you, the competition wouldn't be
>> very good.
>
> No you couldn�t and that is the point.

Yes I could. Blatant assertions to the contrary are irrelevent - the games at
such an event would certainly be far easier to win for anyone of a given
skill level than games at the EC championship. The tournament would certainly
be far easier to win for someone of a given skill level than the EC.

> If a large number of people come together in V:TES it is a difficult event,
> it has always been that way.

I don't think it's very hard to get my point. Like any other event in the universe,
winning any VTES tournament of any size depends on the skill of the opposition
as well as your own skill.

Fred


James Coupe

unread,
Jan 23, 2010, 5:41:58 PM1/23/10
to
JH <jhat...@tenerdo.org> wrote:
>If a player would disappear from the scene and later reappear,
>wouldn't a regular ELO system preserve his rating throughout this time?

Possibly.

Some ELO-derived systems reset on a periodic basis - either on a known
cycle or after defined events.

--
James Coupe
PGP Key: 0x5D623D5D YOU ARE IN ERROR.
EBD690ECD7A1FB457CA2 NO-ONE IS SCREAMING.
13D7E668C3695D623D5D THANK YOU FOR YOUR COOPERATION.

0 new messages