Understanding Percentile Bands in Analysis
I've read the help text in Analysis, as well as the section in the manual, and I've googled percentile ranking to try to work it out, but still not sure that I'm correctly interpreting the Savings percentile bands in the Analysis graph.
As an example, let's say the graph shows the 10-20th Percentile fails, with savings falling below 0 before the plan end, but that all other percentiles succeed by ending above 0.
Does that mean that 10% of the analysis results failed, but the remaining 90% succeeded?
Thank you for any light you can shed. 🙂
Yes, that's an accurate description (actually, that 10-20% of the test cases failed and that 80-90% succeeded).
Thank you for the clarification. For some reason, I thought it wasn't that straightforward, but as usual, it seems I was overcomplicating things. ?
I'd like to probe one step further on this...if my red fixed rate savings line terminates at the end of the planning period at the top edge of the 30-40% band (bottom edge on 40-50% band, what is that telling me? Are the horizontal bands the same as a bell curve, just turned sideways (meaning the median would be the tiny light blue line between 40-50 and 50-60%?) If so, is the red line telling me that I will likely end up better than 40% of those with similar numbers, but worse than the other 60%?
@nc-cpl The tiny blue line separates the 40-50% band from the 50-60% band so, yes, that means that half the results are better and half are worse than that line. If your red line terminates near the top edge of the 30-40% band, that suggests that your fixed rate analysis falls near the 40th percentile of Monte Carlo analysis results. In other words, 40% of Monte Carlo result were worse and 60% were better.
and if I were to adjust our NSDS spend down, would that line move into a higher band?
Not necessarily, but the dollar value would definitely move up.
Would it be accurate (and possibly easier for some to understand) to describe the blue bands as representing the most likely final outcome (dark blue) to least likely final outcome (lightest blue)? Said another way...."I have a 10 to 20% chance of ending up with $XX,XXX (lower light blue band) or $Y,YYY,YYY (highest light blue band), but, I'm MOST likely to end up somewhere between $XX,XXX and $YYY,YYY (the 40 to 60th percentile dark blue area)?
With the red fixed rate savings line representing approximately where WITHIN that dark blue area your specific data suggests you'll be?
No, that's not accurate. Ignoring the fact that the dark blue band is actually adjacent two bands shown in the same color, the probability of ending up in any given band is equal to that of any other band. In other words, you're just as likely to end up in the 50-60th percentile band as you are the 80-90th percentile band or the 10-20th percentile band. And all of this is using "your specific data"; it's just that the mean and standard deviation values you specified on the Financial Assets > Asset Classes page is used in the Monte Carlo analysis to simulate market volatility which, in turn, produces 500 different sets of outputs and those outputs are distributed across the blue bands with 10% of the results falling into each band.