fixes a grammatical error in readme

readme.markdown

@@ -411,9 +411,9 @@ If a word in a passphrase is very uncommon, it may even be auto-corrected by som
 
 ### The similar-words problem
 
-Some of these lists contains very similar words, such as "tabs" and "tab" and/or "suggested", "suggesting", and "suggestions". I don't think this effects entropy-per-word calculations. 
+Some of these lists contains very similar words, such as "tabs" and "tab" and/or "suggested", "suggesting", and "suggestions". I don't think this affects entropy-per-word calculations. 
 
-But I could see how it would be awkward to get two similar words in the same randomly generated passphrase. However, I _think_ the chances of this are pretty small. If we assumed a hypothetical 18,000 word list that was just 9,000 words and their plurals, I think the odds of getting at least one "awkward double" in a 4-word passphrase is `1/9000 + 2/9000 + 3/9000 - 11/9000**2 + 6/9000**3`, which is a really small number. (Thanks to [bwbug](https://github.com/bwbug) for [checking my math](https://github.com/sts10/generated-wordlists/issues/1)!)
+But I could see how it would be awkward to get two similar words in the same randomly generated passphrase. However, I _think_ the chances of this are pretty small. If we assumed a hypothetical 18,000 word list that was 9,000 words and their plurals, I think the odds of getting at least one "awkward double" in a 4-word passphrase is `1/9000 + 2/9000 + 3/9000 - 11/9000**2 + 6/9000**3`, which is a really small number. (Thanks to [bwbug](https://github.com/bwbug) for [checking my math](https://github.com/sts10/generated-wordlists/issues/1)!)
 
 I'll also note that this issue is similar to the risk of getting the same exact word in a passphrase, which we can't help without changing how we generate the passphrases in a way that lowers the entropy of the resulting passphrase.