This is one of those many things where the true answer is "it depends," mostly on your teams budget and design/build process. My experience follows, along with the reasons; your mileage may vary.
The limiting factor for how much stock to pre-order is set by budget ($$) and storage facilities. Inventory ties up your money and space in things you can't easily exchange for other things you might need, but some of those items may save a lot of down time waiting for parts. Dollars in the bank are more versatile, but take time to become parts, especially if you have to work through a school or other bureaucracy on top of shipping time. Possibly that was obvious to you, but probably not to at least one reader of this thread.
A team which is heavily based in CAD (especially if they outsource machining to a sponsor or contractor) might go into a season with practically no inventory, whereas a team which builds a prototype and iteratively tinkers with it until the clock runs out will need to have stock and basic parts in hand in order to do their best. (We're about halfway down that spectrum, working our way slowly towards CAD.)
When selecting WHAT to pre-order, we prioritize based on likelihood that we will use the parts in any given year, then on the lead time to get an order, and thirdly on the amount of money it will tie up. Unless you have a lot of cash, don't bother stocking up on things you can consistently obtain locally if aren't pretty sure you'll need them. (Again, put it in context of your team!) -- Don't be afraid to let your local hardware store or home center hold your inventory.
In case it isn't obvious from what I've already said,
get to know your suppliers. At a bare minimum, go to the AndyMark and Vex and your other online vendors' web pages and figure out how many days it takes them to ship you something, and become familiar with your local vendors' stock and hours. My example: Slidell is 2 day ground shipping from both AM and VEX, and has four hardware stores (one of which is really just a garden and feed store), three home centers, and a metal supplier. None of them stock quite the same things, none are the same distance away (OK, two of them are pretty darn close), and they have different store hours. When the team figures out they need a part, being able to glance at my watch and know who (if anyone) is going to be open when I get there and would likely have that part saves a lot of wasted miles. The last two build seasons, I have spent at least a dozen hours between Christmas and kickoff scouting the local suppliers. I plan to do it again this year, because it clearly paid off, even before you count the relative weight of "off season" vs "build season" time; factor that in, and it was a grand slam.
As for pre-building parts of the "second robot" (suggested by other posters), we do NOT do this, for two reasons.
- We do not decide what needs to be on our control board until after we've designed the robot, which is after we've worked out our high-level strategy, which is after we've given the rules a couple of good reads.
- Second, but even more importantly, we do not build a practice robot and a competition robot, but two essentially identical robots. We do not know a priori which robot we will bag and which we will keep for driver practice and iteration. In 2015, we literally changed our decision on which robot to bag on Tuesday afternoon when "chassis 2" (Peabody) suddenly came together, at least to the appearance of those of us who had been focusing on chassis 1 (Atlas) for most of the preceding week.