From a4c61f3356e210cd32d6834dba3022920bdd2161 Mon Sep 17 00:00:00 2001
From: Ronan McGovern <78278410+RonanKMcGovern@users.noreply.github.com>
Date: Mon, 16 Oct 2023 22:28:56 +0100
Subject: [PATCH] Update 2023-10-04 Five Star Reviews are Overrated.md

---
 blog/2023-10-04 Five Star Reviews are Overrated.md | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/blog/2023-10-04 Five Star Reviews are Overrated.md b/blog/2023-10-04 Five Star Reviews are Overrated.md
index 26ef865..b0b9dde 100644
--- a/blog/2023-10-04 Five Star Reviews are Overrated.md	
+++ b/blog/2023-10-04 Five Star Reviews are Overrated.md	
@@ -28,6 +28,8 @@ Summary
 * We can define the cdf for X<=x_observed
 * Set p such that the observed result is a median for the CDF
 
+Comment: setting p so that the observed result is a median is not the same as a max entropy approach? CDF and entropy of different sets of states are different tings? I added a script to calc max entropy for coins and then stars. I haven't yet confirmed not equivalent to taking CDF approach but I don't think CDF factors in the concavity needed to land on the max entropy solution.
+
 ##### Binomial Solved as Multinomial
 See 2023-10-04 Wolfram Notebook
 
@@ -91,4 +93,4 @@ Finally
 * And relate all this to the binomial problem solved in Nassim's blog
 
 ### Application Notes
-* This is mainly useful in the small sample space, where you have a handful of reviews (with less you can't do much, with more than 30 or so the binomial starts to converge to it's true generating distribution (law of large numbers))
\ No newline at end of file
+* This is mainly useful in the small sample space, where you have a handful of reviews (with less you can't do much, with more than 30 or so the binomial starts to converge to it's true generating distribution (law of large numbers))