Life In 19x19 :: Psychological Survey of Myers-Briggs Type Indicators

Author:

Uberdude [ Tue Jun 21, 2016 7:17 am ]

Post subject:

Re: Psychological Survey of Myers-Briggs Type Indicators

To examine the validity and variability of the gostyle test, I had it analyze several of my KGS accounts. Uberdude is my main account for slow, serious games (though recently it went ? and I went on tilt and played many silly loses) and got this (rounding away absurd decimal places):
Territoriality Moyo Territoriality: 2.7 ± 1.4 Territory
Novelty Classic Novelty: 7.4 ± 1.6 Novel
Aggresivity Calm Aggresivity: 8.7 ± 1.4 Fighting
Thickness Safe Thickness: 4.2 ± 1.5 Shinogi

Silly blitz account:
Territoriality Moyo Territoriality: 2.5 ± 1.4 Territory
Novelty Classic Novelty: 7.0 ± 1.6 Novel
Aggresivity Calm Aggresivity: 8.6 ± 1.42 Fighting
Thickness Safe Thickness: 3.8 ± 1.5 Shinogi

Another account (usually blitz I think, maybe sometimes more sensible):
Territoriality Moyo Territoriality: 4.6 ± 1.4 Territory
Novelty Classic Novelty: 4.5 ± 1.6 Novel
Aggresivity Calm Aggresivity: 7.3 ± 1.4 Fighting
Thickness Safe Thickness: 4.0 ± 1.5 Shinogi

So #1 and #2 quite similar, and #3 more territorial and less novel, despite me expecting #2 and #3 to be closer. Then again I've not played much recently and can't exactly recall how much I have played with different styles/personas on the different accounts.

P.S. Don't understand the Thickness metric, as the labels at the ends suggest a smaller number means thicker to me (thick and safe are similar, whilst shingoi is sort of opposite of thick).

P.P.S Here is Zen19A's results, strangely saying it is a 1k not 8d!:
Territoriality Moyo Territoriality: 2.5 ± 1.4 Territory
Novelty Classic Novelty: 5.3 ± 1.6 Novel
Aggresivity Calm Aggresivity: 6.8 ± 1.4 Fighting
Thickness Safe Thickness: 2.6 ± 1.5 Shinogi

And Sweetrip 9d (Andy Liu), known for being territorial/shingo style, so it at least gets that right:
Territoriality Moyo Territoriality: 7.4 ± 1.4 Territory
Novelty Classic Novelty: 6.6 ± 1.6 Novel
Aggresivity Calm Aggresivity: 7.4 ± 1.4 Fighting
Thickness Safe Thickness: 5.7 ± 1.5 Shinogi

Author:	nadoss [ Sun Jun 19, 2016 12:11 pm ]
Post subject:	Psychological Survey of Myers-Briggs Type Indicators
I'm conducting a survey of go players, in an effort to compare a person's Myers-Briggs Type Indicators to their go style. For this survey, I would like to use the test found at https://www.16personalities.com along with the GoStyle web app found at http://gostyle.j2m.cz The web app works very easily if you play on KGS, as you can just type in your user name. Otherwise, you can still participate if you can create a collection of sgf files to be used for the app. Details on how to do so are on their site. I feel like, for accuracy's sake when it comes to go styles, it would be ideal to have a large number of SDK and Dan players. There are likely small flaws in both the Myers-Briggs style of test, as well as the GoStyle app, but this is meant to be a survey designed to explore the idea of a correlation between personality and go style. I will begin this survey by sharing my own results.

Life In 19x19 http://www.lifein19x19.com/

Psychological Survey of Myers-Briggs Type Indicators http://www.lifein19x19.com/viewtopic.php?f=10&t=13301	Page 1 of 1

Author:	daal [ Sun Jun 19, 2016 1:16 pm ]
Post subject:	Re: Psychological Survey of Myers-Briggs Type Indicators
I'm not sure what form you want your answers, but I will orient mine on your example, though I don't know how you made the attachment: My personality type is: Mediator (INFP-t) Style Territoriality Moyo Territoriality: 4.410741806030273 ± 1.43 Territory Novelty Classic Novelty: 6.483994960784912 ± 1.63 Novel Aggresivity Calm Aggresivity: 7.514821529388428 ± 1.42 Fighting Thickness Safe Thickness: 4.649231433868408 ± 1.48 Shinogi

Author:	emeraldemon [ Sun Jun 19, 2016 1:19 pm ]
Post subject:	Re: Psychological Survey of Myers-Briggs Type Indicators
Well I got INTP this time. I have recieved both INTP and INTJ in the past, maybe it depends on my mood. Here's the gostyle link: http://gostyle.j2m.cz/WAPP_RES/22d0cdb9 ... fa885309c/ My completely unsolicited critiques, which you are free to ignore: Myers-Briggs is not a very good personality test. The concept of personality test is itself a bit dubious, but something like the five factor model has a stronger statistical basis and more empirical support (in my amateur opinion). As for the go styles, I think you are throwing away the most interesting part of the style analysis: the specific shape moves and pattern analysis. The "moyo/fighting/shinogi" style markers are made indirectly. First the programmers asked strong players to manually annotate pro games with "moyo" etc. ratings. Then the classifier attempts to associate specific shapes with specific styles based on these markers. Then the classifier applies those learned shape classifications to your games and gives you a score. So even if these styles exist (do they?), the machine is guessing which you have based on guesses it made about underlying patterns in styles of pro games, which are themselves guesses made by "experts" (Alexander Dinerchtein 3-pro, Motoki Noguchi 7-dan, Vladim´ır Danˇek 5-dan and V´ıt Brunner 4-dan) about what style a pro player has. By comparison, the pattern results are saying something much more direct: players stronger than you play this particular shape more often than you do.

Author:	nadoss [ Mon Jun 20, 2016 2:11 pm ]
Post subject:	Re: Psychological Survey of Myers-Briggs Type Indicators
Your critiques are noted and well thought out. The Myers-Briggs was selected because of its ease of use, and easy comprehension after the fact. The purpose of this, also, is just to see what kind of correlation there is, if any, between particular factors in personality and in go, both of which, though not perfectly represented by any stretch in either method, are not completely inaccurately represented either. Therefore some correlation might still be found and still be of interest to those viewing it.

Author:	sybob [ Mon Jun 20, 2016 2:37 pm ]
Post subject:	Re: Psychological Survey of Myers-Briggs Type Indicators
Interesting topic. Over the years, I have done many psychological tests, including Myers-Brigg tests. I have always found psychology and such tests interesting, but I was classified differently in different tests. Probably, time, mood, circumstances all have a part. So, I think emeraldemon has a point. emeraldemon wrote: Myers-Briggs is not a very good personality test. One should take results with caution. I also made this go style test once. Just curious, as a kyu. It could not detect any style of game preference Well, perhaps that's the similarity?

Author:	Anzu [ Wed Jun 22, 2016 5:58 pm ]
Post subject:	Re: Psychological Survey of Myers-Briggs Type Indicators
My personality type is Entertainer and I like big moyos and playing in the center of the board as much as possible.

Author:	daal [ Thu Jun 23, 2016 9:11 am ]
Post subject:	Re: Psychological Survey of Myers-Briggs Type Indicators
Is it possible that this go test has a bias? So far, it seems that everyone has a style that favors territory, novelty, fighting and shinogi.

Author:	Knotwilg [ Thu Jun 23, 2016 9:29 am ]
Post subject:	Re: Psychological Survey of Myers-Briggs Type Indicators
You're misreading it ... The words at the end represent that end of the scale, not results, so it's - moyo vs territory - classic vs novel - calm vs fighting - thickness vs shinogi You need to look at the red bar to find out where you are. Respondent Madanalogy for example has moyo-classic-calm-thickness

Author:	daal [ Thu Jun 23, 2016 1:47 pm ]
Post subject:	Re: Psychological Survey of Myers-Briggs Type Indicators
Ah, ok.

Page 1 of 1	All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/