diff --git a/05-cache/cache.tex b/05-cache/cache.tex
index c03d3c56a6b280e71c2b6fa938f9ebb6f60dd3cb..6999cd2cd946f887712c6ad3e301b851bbe7c3af 100644
--- a/05-cache/cache.tex
+++ b/05-cache/cache.tex
@@ -166,7 +166,7 @@ reach $\O(N/B \cdot \log N / \log(M/B) + 1)$. This is actually known to be optim
 --- surprisingly, there is a~matching lower bound (FIXME: reference) not only
 for sorting, but for only permuting items.
 
-In the cache-oblivious model, we have is no chance to pick the right number of ways.
+In the cache-oblivious model, we have no chance to pick the right number of ways.
 Still, there exists a~rather complicated algorithm called Funnelsort (FIXME: ref) which
 achieves the same I/O complexity, at least asymptotically. We refer readers to
 the original article.