Mac Musings

March Madness and a Macintosh Performa

Daniel Knight - 2008.03.24

I ran across an interesting tidbit last week: Pete Tiernan of Bracketscience.com uses a 1995 Performa with System 7.5 and a FileMaker Pro database as tools for analyzing the teams in the NCAA playoffs.

It just goes to show that you don't need 8-core 2.8 GHz power to do statistical analysis. The Performa 630 that Tiernan uses has a 33 MHz 68LC040 CPU that officially supports up to 68 MB of RAM. Today's Mac Pro must have 1,000 times the power, but the old Performa has all the power needed for this task.

Tiernan took a break from the games last Friday to talk with us.

Dan: Hi Pete, thanks for agreeing to this interview. I'm sure my readers will enjoy reading about your low-end setup.

How long have you been a Mac user? What was your first Mac? What drew you to the platform?

Pete: The first Mac I owned was the Performa. I used Macs in my advertising day job . . . loved the simplicity.

Dan: You do all of your analysis on a Performa with FileMaker. Which Performa do you have, and what versions of FileMaker and the Mac OS are you running?

Pete: Performa 630CD, Mac System 7.5 Update 2.0, FileMaker Pro 2.0.

Dan: We're dealing with a decidedly old solution here. Are there areas where you feel hobbled by your hardware, software, or operating system?

Pete: I have the database duplicated in a SQL database for my website. We're still building the interface - and there are a number of queries I depend on FileMaker to do, particularly omits and multi-variant queries. My biggest problem, believe it or not, is printing. My Mac printer is a tad buggy, and I haven't networked the Performa to my new high-end iMac.

Dan: What inspired you to collect and analyze all of this data about team performance in the playoffs?

Pete: Way back in 1990 I was just getting tired of losing in the office pool to the know-nothing secretary . . . and figured there had to be a better way. Since this was really before the Internet boom, much of my data gathering was from old NCAA programs that my dad owned. The older data has become my differentiator, since no one else can get it on the Internet.

Dan: I'm sure you spent a lot of time analyzing trends and patterns. Was there one particular thing that surprised you?

Pete: A number of things have surprised me. Most experts think you need strong guard play to make a deep tourney run; it's the opposite. You need a strong frontcourt. Most people think pre-tourney momentum matters; it doesn't. Most think that senior-laden starting units will overachieve; they actually underachieve against seed expectations.

Dan: How long have you been doing this analysis? What have you learned along the way?

Pete: 18 years. The biggest thing I learned: No matter how much data you have or how valid your statistical analysis, there are limits to the predictive accuracy you can achieve. Numbers will help you avoid embarrassing yourself in your office pool, but they don't necessarily lead to winning the pool.

Dan: Do you use a more modern Mac as your main computer? Tell us a bit about it.

Pete: I just bought an iMac for Christmas . . . the highest end 24-inch one. Love it . . . and plan an upgrade to the newest version of FileMaker. The question mark is getting the data from one to the other. I'll have to use a floppy. Remember those?

Dan: Isn't it nice that you can still buy floppy drives for modern Macs? I picked up a USB floppy drive least year to transfer my wife's recipes from DOS floppies to the eMac she was using at the time, but now she uses a flash drive to store her work.

As a statistician, what's your analysis of the argument that Macs don't have viruses or other malware only because they're such a small target relative to Windows?

Pete: I've heard that . . . and I can't really speak to the validity of the argument. I think the Apple people would tell you that they've got more built-in safeguards. I'm not sure if that's true.

Dan: Thanks for your time, Pete.