Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

$$1-(1-1.40*10^{-36})^x \ge 1.09*10^{-9}$$

I want to estimate $x$ such that the probability on the left becomes larger than the probability on the right.

A solution must exist because $1-(1-1.40*10^{-36})^0=0$, $\lim\limits_{n \to \infty} 1-(1-1.40*10^{-36})^n=1$ and the function is continuous.

However Wolfram|Alpha seems to return a wrong result.

Is my reasoning correct? Is there any (practical) way to estimate the result? For my problem would be sufficient to find the order of magnitude of $x$.

share|cite|improve this question
up vote 7 down vote accepted

You can solve the inequality by logarithming :

We get $1-1.09*10^{-9}\ge (1-1.4*10^{-36})^x$

Taking logarithm gives $$\frac {\log(1-1.09*10^{-9})}{\log(1-1.4*10^{-36})}\le x$$

Note that we divided by a negative number!

GP/PARI gives

$$\frac{log(1-1.09*10^{-9})}{log(1-1.4*10^{-36})}\ =$$ $$778571428995750000308340238.34730623955575580298358$$

So, $x$ must be very large

An excellent approximation of the given value is $\frac {1.09*10^{-9}}{1.4*10^{-36}}$ because for $x \approx 0$ we have $log(1+x)\approx x$. And this is very easy to calculate.

share|cite|improve this answer
+1 Great step by step explanation: thanks! Here's the correct result following your approach – collimarco Aug 30 '14 at 10:10
+1, for the approximation ;-) – Jean-Claude Arbaut Aug 30 '14 at 11:20

Mathematica input:

Solve[1 - (1 - 7/(5 10^36))^x == 109/(100 10^9), x, Reals]


$$x\to \frac{11 \log (2)-2 \log (3)+11 \log (5)-\log (11)-\log (19)-\log (4441)-\log (11971)}{36 \log (2)+37 \log (5)-\log (170207)-\log (29375995111834413390753611778599)} \\ \approx 7.7857142899575*10^{26}$$

share|cite|improve this answer
Is there any reason why you have used integers? – collimarco Aug 30 '14 at 9:40

Could you try replacing $1.4$ by $14/10$ and $1.09$ by $109/100$ ? This would make a significant difference (I hope).

Doing the same as user111187, the solution of the equation is $$x=\frac{\log \left(\frac{100000000000}{99999999891}\right)}{\log \left(\frac{5000000000000000000000000000000000000}{4999999999999999999999999999999 999993}\right)}$$ which is $$x=\frac{\log(1-109 \times 10^{-11})}{\log(1-14 \times 10^{-37})}\simeq \frac{109\times 10^{-11}} {14 \times 10^{-37}}=\frac{109}{14}10^{26}$$

Added later to this answer

The problem you posted was "simple" because it was possible to isolate the solution of the equation; it was a very nice illustration of the power of Taylor series.

But, let me take a more complex problem. What is $x$ such that $$e^{x^4}-\cos \left(x^3\right) \gt \epsilon$$ You basically could not answer this question since there is no way to find the solution of the equation. Then use Taylor to find that $$e^{x^4}-\cos \left(x^3\right) \simeq x^4+\frac{x^6}{2}+\frac{x^8}{2}+O\left(x^9\right)$$ So $x\gt \sqrt[4]\epsilon$ is the solution

share|cite|improve this answer
Is there any reason for using integers? Or is it only a trick with Wolfram|Alpha due to its implementation? – collimarco Aug 30 '14 at 9:39
When non integers, CAS do not use infinite precision. Did that make any change ? – Claude Leibovici Aug 30 '14 at 9:45
+1 Thanks! The issue was that when using real numbers (float precision) instead of integers I should have used more decimal digits! Here's the correct result – collimarco Aug 30 '14 at 10:05

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.