Training Sequences in Fiction

I’ve recently read/watched some books/films that have contained extended training sequences during Act I, and in each case, I found it difficult to stick with the story. I think these sequences should be avoided in most situations, unless the writer takes one of the approaches I talk about below. To be clear, I’m only referring to long(ish) sequences near the beginning, not a Rocky-style training montage in the middle or near the end.

It’s all about stakes.

If your story is a spy thriller, and your main character begins the movie without much in the way of skills, you might figure she needs to go through some kind of training before she can go out and do all kinds of cool spy stuff in the real world. But here’s the problem with spending the first 25% or so of the story this way: the stakes remain too low for too long.

If the real conflict in your story is about averting a nuclear disaster, but I have to read about your character learning to use spy gadgets in safe, controlled field exercises for the first quarter of the book, I’m bored, and I’m not going to make it to the part of your book that’s actually good. Sure, you can make the training sequence a real challenge, throw in a nice try/fail circle or two, but if your book has epic stakes, make them epic from the start.

Now, there are definitely ways you can have an interesting training sequence in Act I, or even throughout the whole book. You just need to up the stakes enough.

One good example of this is Brian Staveley’s The Emperor’s Blades. Valyn, one of the Emperor’s sons, is nearing the end of his training to become a Kettral, the most elite soldiers in the world. He’s approaching Hull’s Trial, the final test for graduation, one that many cadets fail. But even before he gets to the trial, he gets word from a dying guardsman that someone might be looking to assassinate him. Then he learns that his father has been murdered. Further attempts on his life make things even more difficult as he tries to discovering who’s after him while he prepares for the trial.

This works because the training sequence is really a complication to the real conflict, which is the threat of assassination. Of course, it’s possible to make the training the main obstacle, as long as the stakes are high enough. Maybe your main character is a criminal who volunteered to join an elite fighting force to stay out of prison. If failure to pass the training results in a life sentence, this is enough to make the reader care if MC succeeds.

I’d say this principle applies to any bridging conflict (a conflict that takes you from your opening to the introduction of the main conflict of the plot). The stakes of the bridging conflict have to be lower than the main conflict, but if they’re too low, you’ll never get the reader to the end of that bridge.

Thoughts on the MUBB Italy Trip

The Marquette men’s basketball team just finished its tour of Italy, which included four games against European teams. Because of the level of competition, it’s impossible to draw any strong conclusions about the state of our team at this point, but I did see a few things in the games I watched that I found encouraging.

Here are my impressions from what I saw:

Traci Carter impressed me as much as anyone on the team. He’s quick and active, producing a number of steals, and he consistently pushed the ball up the court, showing good speed. Given his inexperience, I’m sure he’ll struggle at times as our starting PG early on, but I believe we have a real keeper here. I think the PG position will be in good hands for many years.

Henry Ellenson looks just as advertised, which is to say awesome. His opponents in the Big East will be larger and more athletic than what he saw in Italy, but I expect him to dominate at times even so.

Jajuan Johnson’s jump shot form is much improved. I’m not going to try to make any predictions on what kind of percentage he will shoot this season, but he isn’t going to be the ultra-weak outside shooter he was last year.

Duane Wilson is ready to be one of the top players in the Big East.

Sandy Cohen looks more confident, and I am expecting him to start during the season like he did in Italy. I think he’ll be a major contributor this season.

My main concern came from game 1 against Hauker, where we struggled up until the last seven minutes or so. The score didn’t mean much, but we continued to struggle with some of the same problems as last year, especially defensive rebounding, defensive rotations, and shooting. We cleaned those up somewhat in the second half and more so in the final three games, but it’s hard to say how much of that was our progress and how much was our opponents failing to compete.

All in all, I’m very happy with how the trip went – including no injuries – and I’m really glad that the team put so much effort into making the games and analysis available to the fans. Great work by everyone involved!

Comments on P-values

Noah Smith, popular economics blogger, recently posted a rebuttal to the criticism on the use of p-values in hypothesis testing. While he makes a few good points on why p-values and significance testing have value, I think that his post fails to address a couple of major issues.

First, he states that good science involves replication of results. This is absolutely true, and is, in my opinion, the best antidote for many of the issues related to significance testing. But from my experience in academia (I was an engineering grad student from 2003-2008), the real problem isn’t the lack of good scientists, it’s the poor system of incentives. This extends from the management of journals to the peer review process to the tenure system itself.

Because journals are reluctant to publish negative (non-significant) results, if dozens of independent groups perform similar studies, but only one of these shows significance, this single positive may be the only one published. In fact, the groups that found non-significance will probably not even attempt to publish their work, and no one will have any reason to believe that the lone positive result is false. In this case, no researcher has to do anything wrong in order to produce a bad conclusion by the field.

Also, the tenure system requires that professors continually publish papers in respected journals, which requires doing original work and finding interesting, previously unknown effects. Replicating studies that others have already accepted as legitimate (whether your own or not) gets you no closer to tenure.

The other major problem with p-values is the way they’re interpreted. The common perception is that a p-value of 0.05 means there’s a 95%  chance the effect is real (non-random). But the p-value actually represents p(x|h0), where x is the data and h0 is the null hypothesis. What the researcher wants to know is p(h0|x). The first value (what you have) tells you the probability of observing the data you found, assuming that the null hypothesis is true. But you want to know the probability of the null hypothesis, given the data.

Bayes’ theorem could be used to convert from the term you have to the one you want if you knew p(x), the prior probability of the data. Unfortunately, there’s no way to find this value. However, this paper does a nice job of setting bounds on the value of p(h0|x), depending on the form of the distribution on the data. An interesting result from this work is that for many types of tests, simply subtracting 1 from the t-stat will give you a decent approximation.