My take is that when you use a slider - or the common 5-item Likert scales (strongly agree, agree, neutral, disagree, strongly disagree) - people will waffle over their answer. The fewer the options, the lesser cognitive overload of having decide is today a "great" day or just a "good" day.
I don't want answering the question "how is your day going?" to take more than the 2 seconds it takes to move your mouse to the link.
The problem with the slider is that you have you interface with it. So the workflow would be: get an email, click a link, move a slider (compared to get an email, click one of 3 links). I know I'd want to avoid the extra step as a user, so that's why I opted for that approach.
There is also an option for the user to enter an anonymous comment after they have logged their mood - so if you want to express why today is bad or add note that it is an especially excellent day you can do that.
At a higher level view, I think the value in this particular app is not a scientifically rigorous computation of a teams mood, but rather as a conversation starter ("why has there been a few 'bads' this week? let's talk about it") and a general trend indicator ("morale is down since we moved to the new office, what's up with that?").