Is the A/B Test Result Significant?

A Marlow & Finch checkout test finishes and the whole month collapses into one number: a z-statistic. That is the gap between the two arms measured in units of the wobble you would expect from luck alone. Someone now has to turn it into a sentence, and the same call has to be made for every test in the programme.

The convention the team uses is the standard one. Start from the null hypothesis — the assumption that the new button made no difference whatsoever. Then ask whether June's result is too strange to have come from that world. If it is, you reject the null and call the result statistically significant, which means the gap is bigger than luck comfortably explains. If it is not, you fail to reject, which is a deliberately weak phrase: it means the evidence was not strong enough, not that the button was proved useless.

The conventions, all of which change the answer:

• Two-tailed. The team would have wanted to know about a drop as well as a rise, so a z far from zero counts in either direction. Only the distance from zero matters, not the sign.
• The cut-off comes in with the test, because not every test is run at the same significance level. At 5% it is 1.96.
• The boundary is not significant. A z sitting exactly on the cut-off fails to reject; it has to beat the cut-off, not merely reach it.

Task: Print exactly reject if the result is statistically significant, and exactly fail to reject if it is not. No quotation marks, nothing else on the line.

Input

The first line holds the test's z-statistic, a decimal number that may be negative. The second line holds the critical value, a positive decimal number.

Output

One line holding one of the two phrases above, in lower case, and nothing else.

Example:

Input:
1.44
1.96

Output:
fail to reject