Roger
I think , the way you have put forward is not correct
If you say, my results are 50% ie 2 hits out of 4 in a month, it is mandatory for you to get 3 hits out of 4 in the months. Therefore the number of persons getting three or more hit in months should be taken
Your results are
A) September
nil hit.......0
1 hit...6
2hit.....12
3 hit ....10
4 hits ...2
B) October...
nil....6
1.....17
2.....7
3.....nil
4....nil
C) November
nil..0
1...0
2...1
3 hit...12
4 hit...17
D) December
nil hit...2
1....14
2....8
3hit..6
4....nil
E) January 2015
nil hit...3
1....20.
2hit...7
3hit...nil
4 hit...nil
Thus ,you have 3 or 4 hits total 47
This is out of 30 persons for five months ie 150
Now 47/150 =31.3 %
For ,you proving better than me ,you should only take in account three or four hits by a person ( total 30 fictitious persons)
Amit
I think , the way you have put forward is not correct
If you say, my results are 50% ie 2 hits out of 4 in a month, it is mandatory for you to get 3 hits out of 4 in the months. Therefore the number of persons getting three or more hit in months should be taken
Your results are
A) September
nil hit.......0
1 hit...6
2hit.....12
3 hit ....10
4 hits ...2
B) October...
nil....6
1.....17
2.....7
3.....nil
4....nil
C) November
nil..0
1...0
2...1
3 hit...12
4 hit...17
D) December
nil hit...2
1....14
2....8
3hit..6
4....nil
E) January 2015
nil hit...3
1....20.
2hit...7
3hit...nil
4 hit...nil
Thus ,you have 3 or 4 hits total 47
This is out of 30 persons for five months ie 150
Now 47/150 =31.3 %
For ,you proving better than me ,you should only take in account three or four hits by a person ( total 30 fictitious persons)
Amit
14 comments:
Roger
If you want to be more specific,you can check how many persons ,out of 30 have better score than me.
Amit
Roger
To the best of my knowledge the persons at sr no 2,10,17,18,19 and 20 have better score than me ie number of hits are
10,13,10,11,9 and 10 respectively. These are out of 19 predictions
Thus 6 out of 30 persons ie 20%random dates are better than my predictions
If my predictions are below odds,at least 75% persons should score better than me
Amit
Amit;
Let's look at this month by month.
Sep 2014, we made 4 predictions
You got 2/4 hits, I got 25/30 hits
Oct 2014, we made 3 predictions
You got 1/3 hits, I got 25/30 hits
Nov 2014, we made 4 predictions
You got 3/4 hits, I got 30/30 hits
Dec 2014, we made 4 predictions.
You got 2/4 hits, I got 25/30 hits
Jan 2-15 we made 3 predictions
You got 1/3 hits, I got 16/30 hits
Feb 2015 we made 1 prediction.
You got 1/1 (maybe), I got 20/30 hits
I'm defining a hit as your score or better so if you got 2/4 a hit for me is 2, 3, or 4
Do you see a month where you did better than me?
Roger
Amit;
It's not necessary for me to beat you by a wide margin. All I have to do is show that my chance scores are as good as your scores.
That proves your results are just by chance and not due to some skill or valid method.
Check with a mathematician.
Roger.
Roger
If you a statement that
"my predictions are below odds" at least two third should be able to score more than my score.
we have following persons scoring more than me in each month (out of 30)
Sept.....12
Oct........7
Nov.......17
Dec......6
Jan 15....7
........................
total 49 out of 150 ,scored more than me
ie roughly one third.
Amit
http://quakeearthquakepredictions.blogspot.com/
Amit;
I don't know where you're getting this 2/3 better business but it's wrong.
If I can equal or better your results by chance that's proof enough that YOUR results are just due to chance.
Roger
Even if we take 48 hours window for my predictions ( ie the date of prediction + 12 Hrs on either side, I get 9 hits
Amit
Amit;
The odds on a hit for a 6.1+ quake in 2 days is 0.412
9/19 is 0.474
The difference is not significant.
Roger
Amit;
5.7 today.
You missed it.
Roger
Amit;
Correction; 6.7
Roger
Amit;
Wow, 3 qualifying quakes today.
Too bad you missed them.
Roger
Roger
The 6.8 Japan quake is on the threshold of window ( 43 minutes earlier than predicted)
Amit
Amit;
A 6.5 today gives you a hit.
It doesn't help. You lost the competition, doing no better than my random predictions.
Your method does not work.
Roger
Post a Comment