Tesseract missing quite obvious word

[Mark Zealey]

I've got a weird issue where tesseract is recognising 99% of all of my files fine but in this case it is dropping something that is obviously a word - in the first line here 'RAGMEN'. It looks to me like the segmentation is failing as running with scrollview shows the word in grey, the words around it are also usually shown with joined blobs - can anyone recommend a config option to fix this, I've tried messing with all manner of different config options but can't seem to make any difference. The file is here:

http://mark.zealey.org/tessmissingword.jpg

Thanks!

Mark