Comments on: Rich tables for high resolution reporting http://www.bella-consults.com/waterfall-chart Musings of the office dog at Bissantz Thu, 03 Mar 2011 15:00:58 +0100 http://wordpress.org/?v=2.9.2 hourly 1 By: Robert D. Brown III http://www.bella-consults.com/waterfall-chart/comment-page-1#comment-108 Robert D. Brown III Tue, 12 Jun 2007 06:00:44 +0000 http://www.bella-consults.com/waterfall-chart#comment-108 Dr. Roland said to link to my modification of the waterfall (and tornado) charts here: http://home.comcast.net/~robert_d_brown3/images/Waterfall_tornado_In_cell.xls Best Regards, Rob ________________________________________________________________________ Robert D. Brown III Senior Consultant Decision Strategies, Inc. Decision Strategies - Confidence Through Clarity Dr. Roland said to link to my modification of the waterfall (and tornado) charts here: http://home.comcast.net/~robert_d_brown3/images/Waterfall_tornado_In_cell.xls

Best Regards,
Rob
________________________________________________________________________
Robert D. Brown III
Senior Consultant
Decision Strategies, Inc.
Decision Strategies – Confidence Through Clarity

]]> By: Robert D. Brown III http://www.bella-consults.com/waterfall-chart/comment-page-1#comment-105 Robert D. Brown III Wed, 06 Jun 2007 16:50:51 +0000 http://www.bella-consults.com/waterfall-chart#comment-105 Bella - What a great idea! I use waterfall charts extensively in my consulting practice, but not the particular graphic table implementation you have employed. I am frequently troubled by the lack of standardization in charts and the expanding memory requirements needed to develop them. Your method not only addresses the standardization issue, but it makes much smaller files. Not only do we need less chart junk, we also need less memory junk on our computers. I didn’t see, though, how net negative values might be displayed with this specific type of graphic table unless a second column was used for the net negative values. To help out a bit, I expanded the logic of your waterfall table to display net negative values and the cross-over bar needed to transition from positive to negative values. I sent an example of this to your friend, Dr. Roland. Maybe he will share that with other readers here. Another useful implementation of graphic tables may exist with something called tornado charts, which I also use frequently. If you are not familiar with these, they display the amount of variation an uncertain variable contributes to an objective function, like NPV. It is created by sequentially setting each uncertain variable to its p10 and p90 values while holding all the other variables at their mean (...there are various positions on this, but that is another topic) and recording the effect of the outcome on the objective function. Rank ordering the uncertain variables by absolute delta in the objective function produces the tornado shape. The real information from these charts comes from being able to observe which uncertainty would cause one to regret picking one strategy over another. This leads to value-of-information and value-of-control discussions, which lead to the development of hybrid strategies that employ the best features of all the strategies and mitigate the worst. This requires at least a pairwise comparison of tornados, and that would be difficult to implement properly in a graphic table. However, if a pairwise delta between tornados is considered, this would be easily implemented. Otherwise, the singly charted tornado only displays the relative variation imposed by uncertainties in a single strategy. The file that I sent to Dr. Roland also contains an example of a single tornado graphic table. Best Regards, Rob ________________________________________________________________________ Robert D. Brown III Senior Consultant Decision Strategies, Inc. Decision Strategies - Confidence Through Clarity Bella – What a great idea! I use waterfall charts extensively in my consulting practice, but not the particular graphic table implementation you have employed. I am frequently troubled by the lack of standardization in charts and the expanding memory requirements needed to develop them. Your method not only addresses the standardization issue, but it makes much smaller files. Not only do we need less chart junk, we also need less memory junk on our computers.

I didn’t see, though, how net negative values might be displayed with this specific type of graphic table unless a second column was used for the net negative values. To help out a bit, I expanded the logic of your waterfall table to display net negative values and the cross-over bar needed to transition from positive to negative values. I sent an example of this to your friend, Dr. Roland. Maybe he will share that with other readers here.

Another useful implementation of graphic tables may exist with something called tornado charts, which I also use frequently. If you are not familiar with these, they display the amount of variation an uncertain variable contributes to an objective function, like NPV. It is created by sequentially setting each uncertain variable to its p10 and p90 values while holding all the other variables at their mean (…there are various positions on this, but that is another topic) and recording the effect of the outcome on the objective function. Rank ordering the uncertain variables by absolute delta in the objective function produces the tornado shape. The real information from these charts comes from being able to observe which uncertainty would cause one to regret picking one strategy over another. This leads to value-of-information and value-of-control discussions, which lead to the development of hybrid strategies that employ the best features of all the strategies and mitigate the worst. This requires at least a pairwise comparison of tornados, and that would be difficult to implement properly in a graphic table. However, if a pairwise delta between tornados is considered, this would be easily implemented. Otherwise, the singly charted tornado only displays the relative variation imposed by uncertainties in a single strategy. The file that I sent to Dr. Roland also contains an example of a single tornado graphic table.

Best Regards,
Rob
________________________________________________________________________
Robert D. Brown III
Senior Consultant
Decision Strategies, Inc.
Decision Strategies – Confidence Through Clarity

]]>