lets see if the mean is 25
our numbers would be +/- sqrt(13.5) from mean and +/- the chosen number 7 from mean
and 25 of course the mean for the 5th value
mean = ((25 + sqrt(13.5)) + (25-sqrt(13.5)) + (25 + 7) + (25 - 7) + 25)/5
we see that the added values cancel so we get
mean = (25 + 25 + 25 + 25 + 25)/5
so mean = 25
we got our mean lets confirm we got the sd too
variance = ((25 + sqrt(13.5) - 25)^2 + (25-sqrt(13.5)-25)^2 + (25 + 7-25)^2 + (25 - 7-25)^2 + (25-25)^2)
we see that the variance = (13.5 + 13.5 + 7^2 + 7^2)/5
5^2 = 125/5
so thats true
this worked out the trick was to know that when doing the final square root the value inside could be (25 - x2)^2 or (x2 - 25)^2 but since we were picking a number smaller than 25 for x2 the way for me to ensure this was to make the equation inside the square as shown.
(25 - x2)^2 that made the math pick a number lower than 25 not higher
unless it was a fluke it seems you can pick randomly one number for the second point and then plug it into the equation for the variance simplified and get another valid number that will make the standard deviation the value you wish ...
and since you are picking equal number of values on either side of the mean plus the mean for the 5 points the difference cancel and produce a mean automatically
if the number of points had been even you would have chosen ONLY the points on either side of the mean again the math would have worked out the rest and you would have your 4 points say ..
hope this is not too confusing I wanted to be VERY sure it worked out and It worked fantastically ...