[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums

gnubg - Cubes: at what percentage does 'Doubtful' stop being triggered?

Posted By: sebalotek
Date: Sunday, 17 August 2014, at 10:10 a.m.

In Response To: gnubg - ‘Analyse -> Market Window’ function (Dmitriy Obukhov)

Thank you all for your feedback and links. Glad to hear that I might have had a valid question in terms of the usage / usefulness of the Market Window function and I also got the bonus of a lesson in cube handling is from great players. My understanding of cube theory is not great: probably at the level of Casual -> Intermediate.

I would like to ask one more question, hopefully valid:

Dmitriy mentioned in his post a “ "real" doubling window.” and I think this is really the crux of what I would like to be able to see on screen (if it can indeed be calculated).

After analysis, Gnubg classifies cube errors as either:

Very Bad

Bad

Doubtful

or

No comment – because the Cube was Good or better

We know that at that position, at that match score and match length, my 55.7% GWC was in the zone where it would be classified as a ‘Very Bad’ cube. If there were to be an increase in my GWC percentage and consideration of the other factors for that particular game position and score (volatility. market losers, gammons, cube efficiency, probability of redoubling in and out later, etc.), there should be a certain numeric percentage where a Cube offering / decision would stop being classified as ‘Doubtful’ and start being classified as ‘No comment – because the Cube was Good or better’.

Logically, it seems to me that, internally, gnubg must already know where these borders are in order to classify these cube errors with words like 'Bad' and 'Doubtful'.

I’ve added some notes to a screenshot to try and illustrate:

http://i.imgur.com/Tv0h3YY.png

If it were possible to display this ‘Cube Sweet Spot’ percentage where ‘Doubtful’ stops, then would it enable us to recognise positions where we had over or under-estimated our chances (and by how much), by comparing our actual percentage against this number. Is this the start of the "real" Market Window? Perhaps it could be a reference point to adjust and improve our cube handling and after some time, get a better feel for our GWC percentages?

Finally, looking back at the original position,

http://i.imgur.com/9mqOVuD.jpg

I’m must admit I still get a subversive twinge to cube here even though it’s already been proven to be scientifically and statistically very wrong. So, one of the lessons for me here is I don’t think I even considered the closeness of the pip-count and gave too much weight to the pleasant pair of Golden points / mini-prime I’d built so far.

Messages In This Thread

 

Post Response
Your Name:
Your E-Mail Address:
Subject:
Message:

If necessary, enter your password below:

Password:
 

 

[ View Thread ] [ Post Response ] [ Return to Index ] [ Read Prev Msg ] [ Read Next Msg ]

BGonline.org Forums is maintained by Stick with WebBBS 5.12.