Start by reading this article about how to avoid choking under pressure:
One of the aspects of doing well under pressure is to practice in pressure situations. In particular, setting up fake high pressure situations so that when the real event occurs, everyone both can handle it and knows he can handle it.
One way to create high pressure situations at practice is to ask the team to focus and play hard. This can work well if everyone is sufficiently motivated, but doesn't work as well as the practice wears on or even through out the season. Putting aside internal motivation (which should be an incredibly high motivator), how else can high pressure situations be created during practices? How can we simulate the pressure from Nationals without playing against another team from Nationals who is just as fired up as we are?
Well, how about keeping statistics and reward those statistics?
A college that had a dominant women's soccer team would track all its players' speed by running sprints at the end of each practice. Every one lined up on one line according to speed, the fastest on one end, the slowest on the other. At the end of each sprint, the order would adjust, with the faster person still at the one end, but the middle runners adjusted according to who crossed the line first. This particular way of running sprints made it easy to see who was faster and slower, as each person was next to others who were close in speed to her. At the end of all of the sprints, the order would be recorded and posted.
The result of the sprints tracking was that the slowest person was incredibly motivated to become faster. The article I read went on to point out that the slowest freshman one year became the fastest sprinter by her senior year, because of the motivation from the sprints.
We could have a similar setup at practice for sprints, sure. It would help motivate those in the middle, and keep those at the fast end honest in moving!
Speed is only one aspect of the game, however. Scrimmages at practice could also be tracked, as the teams are fairly stable after they're declared at practice (and fairly stable through the season as offense teams and defense teams are selected). At the end of practice, keep the team divisions, but make a note of which team won how many scrimmages, maybe even how many points. Keep track of those values and rank the players on how they did, either by points or by scrimmages won.
I don't know that I'd recommend keeping stats on scrimmages the way that game stats are kept at tournaments. That requires a lot more commitment from a non-player.
The trick in tracking statistics, however, is to make sure every player continues to grow and expand upon his skill set. If you're tracking how many turnovers a player made at practice, she's going to stop trying to throw those throws that are *just* beyond her reach. Yet, practice is when you want her trying those throws so that she *can* make them in a game: you want growth at practice, not withdrawal.
Possibly having a non-tracked practice for people to try new positions and throws could also be beneficial.
For this reason, I would strongly argue against tracking "how many turnovers I had at practice." The skills and drills parts of practices don't lend themselves particularly well to statistics, and are opportunities for growth that shouldn't be wasted.
Of course, the true source of pressure in sports comes from actual competition. Heading out to a tournament and experiencing the pressure is a better source than the artificial pressure of tracking scrimmage stats. Just make sure the tournament's level is high enough, and that the team learns at the event, as even a loss is a chance to learn.