为什么R和python的游程检验结果不一致？

Question

R语言游程检验

library('tseries')
a=c(1, 1, 1, -1, -1, 1, -1, 1, -1, 1, 1, 1, -1, 1, -1, 1)
runtest1=runs.test(factor(a))
runtest1

Runs Test

data: factor(a)
Standard Normal = 1.3868, p-value = 0.1655
alternative hypothesis: two.sided

Python游程检验
from statsmodels.sandbox.stats.runs import runstest_1samp
x=[1, 1, 1, -1, -1, 1, -1, 1, -1, 1, 1, 1, -1, 1, -1, 1]
rst=runstest_1samp(x)
print(rst)
(1.1094003924504583, 0.26725749315438785)

两个pvalue相差好大，为什么呢？