You should ask yourself, if you describe the problem that you try to solve with your database, whether the items you are putting in this database typically have 10 doubles in them, or is the number 10 "just a management decision", and could they have selected 9, or 11 if they had been in a different mood?
Furthermore: are your 10 doubles really just 10 doubles, or does x[0] have a different meaning than x[1]. If you would have several strings in your name / address, would you call them AddressLine1, AddressLine2, City, PostCode, or would you call them AddressData[4]? Think whether your doubles have a special meaning, or whether they are interchangeable.
Can it be, that in a few years, when you have zillions of records, they suddenly decide they want to save 11 doubles instead of 10 doubles? Or maybe worse: some records have 10 values, some have 11 values, and some may even have a variable number of values.
If you really think, that the number 10 is not just a decision, but is intrinsic to the problem you are solving, as solid as the number of squares on a chess table, then you should add all 10 doubles as separate values to your tables, because that would be the fastest for database actions
If on the other hand, you might think that in future people might want different number of doubles, I'd say go for a one-to-many relation.
This would improve readability of your code: a "Student with this grades" is easier to read as "student with his grade for arithmetic and his grade for biology, and his grade for ...", even if you are absolutely sure that your School will only have 10 courses.
For the cost of slowing down some of your database accesses if you put your values in a one-to-many relationship, you'll have the benefits of improved reusability / maintainability (in future some items might have 11 values, or zero values), lower learning curves because it is more common to implement it as a one-to-many.