The algorithm accepts the two lists (Words and Word IDs of words learnt in the previous lesson) and creates a word list that has the updated values of numberOfTimesLearned for the Words that were recently learned.
Then it edits this updated list by removing words that have never been learnt.
This list is then sorted in Descending order by numberOfWordsLearned and split into two lists of Words: sent and unsent where 
- sent contains all words that have hasAlreadyCollected as True.
- unsent contains all words that have hasAlreadyCollected as False.
The algorithm then returns 3 words from these 2 lists, where priority is given to the 'unsent' list as the word stickers from that class haven't been awarded yet.

It logically appealed to me: As someone playing the game, i'd like the opportunity to eventually earn all stickers of the words i've learned and not just any 3 random stickers from the huge list of words i've learned
throughout. By prioritising words i've never got stickers for before, that opportunity is given.
Also, it makes sense to earn stickers for a word i've learned many more times than a freshly learned word.
By combining these two thoughts, the current algorithm was created.

It's honestly my first time programming in C#, so I focussed mainly on getting the functionality in sync with my intended solution.
If given more time, I'd like to explore the language a lot more to understand and apply more efficient ways of writing code in C#.
With respect to the algorithm, since the guideline was open-ended in terms of the implementation, I applied the solution that seemed most appropriate to me, but there might have been something i missed. I'd like to spend a bit more time on understanding the problem completely so that i can find the best solution.
Also, I'm unaware of what kind of inputs will be actually provided to the function, so there was a limited amount of testing.