Aliases and anonymity online

[EdLee]

My point is that I don't believe that various changes in someone's true rank/strength based on time settings, state of mind, or just their physical health counts as sandbagging if the situation occurs where they're a few stones stronger than the average.

Could you give a concrete example of what you consider to be sandbagging ? (post 28)

[Abyssinica]

My point is that I don't believe that various changes in someone's true rank/strength based on time settings, state of mind, or just their physical health counts as sandbagging if the situation occurs where they're a few stones stronger than the average.

Could you give a concrete example of what you consider to be sandbagging ? (post 28)

Being a 3 dan on Tygem (And having 3d on average strength given your up and downswings) and registering for the AGA and entering a tournament as a 5 kyu.

[EdLee]

Being a 3 dan on Tygem (And having 3d on average strength given your up and downswings) and registering for the AGA and entering a tournament as a 5 kyu.

For simplicity and clarity, could we stick to just one rating system, say AGA 3 dan and AGA 5 kyu -- is this OK ?

Suppose your above hypothetical person plays at AGA 3 dan level given slow time settings,
and the same person plays at AGA 5 kyu level in blitz.

Two questions for you:
- Do you think that is possible ( a swing of AGA 3 dan and AGA 5 kyu, from slow vs. blitz ) ?
- If the AGA 3 dan and AGA 5 kyu difference is true, given the time settings, then in your above example, why is it sandbagging ?

[mhlepore]

Oh, mhlepore, Abyssinica,

Could you give a concrete example of what you consider to be sandbagging in Go, and the reasons.

From what you've written here so far, I'm not certain you believe there is such a thing as sandbagging in Go, or do you ?

Sure, there's sandbagging, but it has to do with how you represent yourself. Adopting the philosophy of "playing your rating" is in most cases not sandbagging.

A 10 year old child taking lessons from a professional may be improving so quickly that her rating lags behind her true strength. This ten year old knows she has a good chance of winning her games at the tournament next month, but I wouldn't call playing her rating sandbagging. While I will likely lose when playing such a child, I accept it, because I think it would be a bad strategy overall to say "anyone under 4 feet tall taking lessons from a pro should self-promote, so as to avoid sandbagging."

On the other hand, my wife played at the Congress this year, and it was her first rated game in nine years. We know she was not 16k, so she entered as 9k, which ended up being about right. Had she entered as 16k, this would have been sandbagging in my opinion, because while she hadn't played an AGA game in 9 years, she had thousands of online games under her belt to gauge her true strength.

Unless she had bacon for breakfast.

[xed_over]

Suppose your above hypothetical person plays at AGA 3 dan level given slow time settings,
and the same person plays at AGA 5 kyu level in blitz.

- Do you think that is possible ( a swing of AGA 3 dan and AGA 5 kyu, from slow vs. blitz ) ?

No. (and I don't believe you do either)

And the only way totally prove that would be to have separate rating systems for the different time settings. There might be some measurable difference, but not 8 stones. And I don't believe whatever the difference, it would be enough to warrant separate rating systems (but without actual data, we'll never know).

[mhlepore]

...
Since you freshly returned from the US Open, let's use it as an analogy.
AGA, like KGS and IGS, also has only one rating that does not differentiate between time settings.

For slow settings -- say US Open, (90 mins. + 5 x 30 secs), or (90 mins. + 5 x 1 min.), etc. -- suppose you real level is AGA 1 dan. You know this.
For faster settings -- say (30 mins. + 5 x 30 secs) like at Cotsen's -- further suppose your real level is AGA 3 kyu. That's 2-3 stones difference. You also know this.

( In case you're wondering if that's an exaggerated example, it is not. I easily drop 2 to 3 or even more stones from slow to blitz. )

...

Ed,

Is this you:http://agagd.usgo.org/player/16081/??

If so, it seems you have a 56% win rate at Cotsen Open games, and a 55% win rate in Go Congress US Open games. Are these percentages close because you played below your true rating at Cotsen tournaments, and at your rating at US Congress games?

[EdLee]

There might be some measurable difference, but not 8 stones. And I don't believe whatever the difference, it would be enough to warrant separate rating systems (but without actual data, we'll never know).

Yes, lots of data points would be very helpful.

I can only provide one data point, mine; see next post.
At least 2, 3, or even more stones drop from slow to blitz.

[EdLee]

If so, it seems you have a 56% win rate at Cotsen Open games, and a 55% win rate in Go Congress US Open games. Are these percentages close because you played below your true rating at Cotsen tournaments, and at your rating at US Congress games?

Hi mhlepore,

You want to get up close and personal ?
No problem. Very happy to discuss.

Yes, that's me.
Since you're so happy to dig up the records, which is freely available to the public, why don't you go ahead and dig up every one of my rated AGA games (or, as many as are a available). Then, tell us what you can gleam from the data.

As you do the digging, here are some of the questions you may want to answer:

(a) How many rated games can you find ?
(b) Given (a), what would you say about the statistical reliability ?
(c) How many Cotsen rated games can you find ?
(d) Given (c), what would you say about the statistical reliability ?
(e) How many rated US Open games can you find ?
(f) Given (e), what would you say about the statistical reliability ?
(g) How much elapsed time(s) between these rated games ?
(h) Given (g), what would you say about the statistical reliability ?

You asked; now, please do the work.

[mhlepore]

If so, it seems you have a 56% win rate at Cotsen Open games, and a 55% win rate in Go Congress US Open games. Are these percentages close because you played below your true rating at Cotsen tournaments, and at your rating at US Congress games?

Hi mhlepore,

You want to get up close and personal ?
No problem. Very happy to discuss.

Yes, that's me.
Since you're so happy to dig up the records, which is freely available to the public, why don't you go ahead and dig up every one of my rated AGA games (or, as many as are a available). Then, tell us what you can gleam from the data.

As you do the digging, here are some of the questions you may want to answer:

(a) How many rated games can you find ?
(b) Given (a), what would you say about the statistical reliability ?
(c) How many Cotsen rated games can you find ?
(d) Given (c), what would you say about the statistical reliability ?
(e) How many rated US Open games can you find ?
(f) Given (e), what would you say about the statistical reliability ?
(g) How much elapsed time(s) between these rated games ?
(h) Given (g), what would you say about the statistical reliability ?

You asked; now, please do the work.

Sounds like fun, but most of the work is done, and this thread is getting a little too chippy for me. But I'll end by saying I agree with you about the statistical reliability, or lack thereof, in using rated games to draw meaningful conclusions. Which is precisely why if I were a tournament director, I would question a person who says they are 3 stones stronger at longer rated games, or better after the Wednesday day off at the Congress, or whatever, when there is not enough data to take and run with.

[EdLee]

it seems you have a 56% win rate at Cotsen Open games, and a 55% win rate in Go Congress US Open games. Are these percentages close because you played below your true rating at Cotsen tournaments, and at your rating at US Congress games?

Sounds like fun, but most of the work is done, and this thread is getting a little too chippy for me.

Hi mhlepore,

No, most of the work is not done. Not even close.
I would say over 90% of the work is not done -- and answering the questions is only a starting point.

Instead of discussing the issue at hand, you happily bring up some data,
some personal data -- that even you agree is statistically unreliable --
and without any further actual investigation, throw up some accusatory questions
that appear to me awfully like an ad hominem.

May I refresh your memory: Forum rules, guideline #1.

___________________________________________________________________________________________________________________________
EDIT: After the above, I received a private message from mhlepore.
And I replied back. The tone, so far, is very friendly in the 2 private messages.

Please move along; Nothing to see here.

[daal]

Kind of ironic that a thread about anonymity in the go world should veer off in into a personal discussion about someone who chose to use his real name on the forum. On the one hand we see that using one's real name makes one more accountable, on the other, it opens the user up to lines of inquiry anonymous users don't have to trouble with. The (Western) go world is a small one, and many of the names behind the aliases are known to more than a few - all the more reason to be generous with the number of aliases people should be allowed to use.

[jeromie]

I choose to use my real name online (my user name is my first name, but it's rare enough that it wouldn't be difficult to trace it back to my full identity, especially since I also report my location), but I think that anonymity should remain an option in most online settings as a matter of principle. There are a lot of good reasons someone might wish to remain anonymous on the web, and a number of those reasons are applicable to go servers and forums. In places that permit anonymity, revealing a person's identity (or threatening to do so) would be a major breach of trust.

As for the discussion around sandbagging, I think that the term carries an important connotation of intent. Anyone who states their level below what they know it to be in order to increase their chance of winning or with a desire to humiliate their opponent is sandbagging. But there can be many reasons to play a game below your stated level that are not sandbagging: servers that only allow a maximum strength at registration, ignorance of your own true rank, tournaments that rely on established ratings rather than user reported ranks, and significant improvement while an account was dormant all come to mind. Server admins and tournament organizers can't accurately judge a person's intent, so they have to set policies that will lead to the overall best health of the play environment and deal with exceptional cases when they arise.

[Kirby]

Since you freshly returned from the US Open, let's use it as an analogy.
AGA, like KGS and IGS, also has only one rating that does not differentiate between time settings.

...

Yes, you can argue, as you have, that this is not your problem -- it is merely AGA's fault/inadequacy for not differentiating between time settings.

Really ?

Why would it be my problem? I haven't broken any rules. I played at my rank and won a prize. So what?

But even going further: Is it even a problem? If the answer is yes, then the rules and/or rating system should enforce this, no?

For the record, bacon is delicious.

[wineandgolover]

What a bizarre, and unnecessarily acrimonious thread.

Here are my thoughts. Flame away if you desire.

I have no problem with people having multiple accounts on servers. I have one I for when I am more serious, and another I use when I'm rusty or trying blitz, or drinking.

I can't think of a good reason for a user to have multiple accounts on a bulletin board. Just be who you are, and post what you think, even if the server allows anonymity. If what you have to say is so controversial that you are tempted to post with another account, rethink posting it at all. Do people really think it's a good idea to allow a second account to post objectionable material (racist, sexist, homophobic, whatever) and a third (and fourth) to upvote it?

I would favor a real-name policy on this BB.

Regarding banning, assuming a banning is warranted, of course admins should ban the user, not the account. The account didn't do something wrong, the user did.

Regarding sandbagging, I won't fault a player for playing at their listed rank in AGA tourneys. The way the system was set up at this year's Open, you had to fill out a form requesting a different rank, and justifying it to the TD. I totally understand people not wanting to go through the hassle. I also totally understand somebody grumbling about the rating system and saying, "Fine, I'll show you." The current rating system is flawed, but that isn't the players' fault. And if everybody requests a rating jump due to low sigmas and slow adjustments, than it's the same as if nobody jumped. Finally, until the AGA (and other associations) come up with a way to factor in game length, I don't think it's reasonable to force the players to do so.

FTR, I agree with daal about the irony in this thread. That was staggering!

[Anonymong]

I can't think of a good reason for a user to have multiple accounts on a bulletin board. <snip> I would favor a real-name policy on this BB.

One's own lack of imagination of appreciation of people's situations and cultures doesn't mean other people might not have a good reason. Perhaps Salman Rushdie plays Go and doesn't want to make it easier for himself to be assassinated. Or someone in an official capacity in the go world wants to post in a clearly non-official voice. Another common group of victims of this kind of thing are minority groups who don't conform to your cultural expectations, say a transgender person who is known in the go world as a woman but who's legal name is still a male one, or a native american with a 'strange' name. That you are happy to reveal your identity doesn't mean everyone else is, just look at the grief Facebook's policy causes.

If what you have to say is so controversial that you are tempted to post with another account, rethink posting it at all. Do people really think it's a good idea to allow a second account to post objectionable material (racist, sexist, homophobic, whatever) and a third (and fourth) to upvote it?

Do you think my original post was controversial/objectionable? I have my reasons for making this account, one of which was to test the self-restraint of the admins here versus elsewhere. Another was to focus discussion on the issue itself rather than myself. I do agree though that sock-puppetry behaviour should not be allowed (such as upvoting one's own comments, posting glowing reviews on your own books/teaching/etc).

Regarding banning, assuming a banning is warranted, of course admins should ban the user, not the account. The account didn't do something wrong, the user did.

What if the punishment is not a ban which disappears after a day or two, but a permanent black mark on the user's profile that they are a bad/naughty person. Should this punishment apply to all aliases? Or perhaps this form of eternal reputation-damaging punishment is the problem, not the spreading of it to all aliases?