diff --git a/rat/EA-predictions/Analysis.md b/rat/EA-predictions/Analysis.md
index fec13ee..ff0be20 100644
--- a/rat/EA-predictions/Analysis.md
+++ b/rat/EA-predictions/Analysis.md
@@ -106,11 +106,16 @@ For the 35 people who took part in the original prediction making, their results
 
 The average accuracy is 55.12%, that is, the average participant got 13.22 out of 24 questions right. If it had been reached, a target credence of 80% would imply an average of 19.2 correct answers. In other words, in this limited domain, when these people say 80%, the thing happens 55% of the time. If they bet, they'd be replacing ~1:1 bets with 1:4 bets. 
 
+## Is this an spurious result because a small number of questions were really, really hard?
+No. See the following scatterplot:
+
+![](https://nunosempere.github.io/rat/EA-predictions/Scatterplot-questions.png)
+
 ## Further analysis
 
 Questions for further analysis:
-1. Should the two savants who got very near 80% be proud, or should we expect them to exist merely by chance? 
-2. Are the results an artifact of a small number of questions which were really hard (f.ex. the % of LessWrongers in EA)?
+1. Should the two savants who got very near 80% be proud, or should we expect them to exist merely by chance? Maybe.
+2. Are the results an artifact of a small number of questions which were really hard (f.ex. the % of LessWrongers in EA)? Solved: No, See above.
 
 I expect to answer those questions in the near future.