01Points per hour is the wrong ratio
The obvious calculation is points available divided by hours required, largest first. Run it and the paper wins every time, because papers carry the most points. Then the night ends with a mediocre paper and three zeros, which is usually a worse result than the one you could have had.
The correction is one subtraction. Ask what each item scores if you do nothing at all, and subtract that from what it scores if you do the work. What is left is the marginal value, and it is the only figure worth dividing by hours. A long paper you can only get halfway through may gain very little over the half you already had. A short graded post that goes from nothing to most of its marks in twenty minutes may gain almost all of its value in that twenty minutes.
02The subtraction, run on four items
Write every outstanding item on one sheet with three columns: marks still available, marks you would get by submitting what already exists, and honest hours to a submission you would not be ashamed of. Subtract column two from column one, divide by column three, and sort. The list almost always reorders itself, and the reorder is the whole point of the exercise.
Two things fall out of that sheet immediately. Items already partly written are worth less than they look, because much of their value is banked. Items at zero with a low ceiling are worth more than they look, because all of their value is still on the table and the hours are small.
- Marks still available on this item, taken from the syllabus rather than from memory
- Marks the current state of it would already earn if submitted unchanged
- The difference between those two numbers, which is what the hours actually buy
- Honest hours to a credible submission, counted at the speed you work at one in the morning
- The quotient, and then the order, which is simply that column sorted downwards
03Where the ratio gets overruled
Late policies come in two shapes and they behave completely differently. A slope deducts a percentage per day, which means an item keeps most of its marginal value into tomorrow and can safely go last. A cliff sets the value to nothing after a stated hour, which means that item has to be ranked by its hour rather than by its quotient. Find out which shape each of your courses uses before you sort anything.
Two other things beat the arithmetic. A gate is an item you must pass irrespective of what it contributes to the average, and a gate outranks everything. A dependency is an item another item is built on, such as a proposal a later assignment extends, and doing it out of order costs you the second one as well as the first.
04The stop rule, and choosing what you drop
There is an hour after which starting a new item costs more than finishing the open one, because the half-written thing on your screen still carries most of its marginal value and the new thing carries all of its setup cost. Find that hour before the night starts and write it down. When it arrives, you finish what is open, submit it, and stop.
Then choose the drop out loud. Something is not getting done, and an undecided drop is expensive in a way people underestimate: it sits in the back of your head all night and slows everything you are actually working on. Name the item, write down that it is abandoned, and give the attention back to the list.
If the same night keeps arriving every term, the problem is a schedule rather than a deadline, and it is worth handing to urgent triage earlier than feels necessary.
FAQQuestions to control
Does a partial submission ever beat submitting nothing?
Usually yes, and the sheet tells you by how much. A partial submission collects whatever marks the completed sections earn, and it also changes what a late conversation is about, because there is now something on record. The exception is a course whose policy scores an incomplete submission at zero regardless of content, which is worth knowing before the night rather than after it.
How do I estimate hours honestly at one in the morning?
Take your calm estimate and add half of it again. Tired work is slower and needs more correction, and nearly every triage plan fails because the hours were counted at daytime speed. If adding half pushes an item past your stop hour, that item was never in tonight's list, and knowing that early is what makes the rest of the night usable.
Is it worth writing to the instructor tonight?
Yes, and send it before you start rather than afterwards. A message that names the item, offers a date you will genuinely meet and asks for nothing else takes two minutes and occasionally changes the arithmetic entirely. Send it, then work rather than wait for a reply. The reply arrives on their schedule, and the items on your sheet do not pause while it is coming.
What if two items are gates?
Then the arithmetic is finished and the order is set by their hours. Gates cannot be traded against each other, because failing either produces the same outcome as failing both. Do the one that closes first, submit it, then do the other. If the two windows genuinely overlap and neither can be moved, that is the moment to write to the programme rather than to write faster.
