FBP: A Tradition Like No Other

Hey, this ain't no fly-by-night operation! This instantiation of the Football Pool has been around for over two decades, with roots that disappear even deeper into the mists of time. Take a walk down memory lane, check out the official trophy for which our participants vie, and visit the hushed cloisters of the Hall of Fame where our past winners are enshrined.

Where It All Began (and How We Got Here from There)

The annals of the Football Pool go back to the 1970's at Purdue University. Graduate students in the Computer Science Department ran an annual pool that predated my on-campus arrival; somewhere around 1976, Janet and I jumped in. In those years, the user interface to FBP was considerably more primitive than the elegant design you see today; it consisted of a gigantic piece of paper taped to the wall of the coffee room in the CS Department. All the participants and games for the week were listed in a giant matrix; to make your picks, you simply checked off your chosen winners and filled in a Monday Night score. On Friday nights at quitting time, the FBP Commissioner (Bob Mead, as I recall) would take the paper home, tally up the scores by hand once games had been played, and then post a new slab of paper on Tuesday morning.

We learned about the power of Delphi polls a year or two later when Paula Perkins, the department head secretary, won the Weekend Competition. Paula waited until after most people had made their picks on late Friday afternoon, then made her choices by evaluating what other people had done. This enraged the graduate student population, primarily because they wished they'd thought of it first. The incident marked the end of open postings of the FBP database.

The Purdue computing environment at the time was primarily a batch environment; students punched programs on cards, submitted the card decks to the guardians of the dual CDC 6500's in the basement of the Math Science building, and eventually collected a printout when when the job finished. Bob wrote a program to tally FBP scores, posted a weekly list of games with A/B choices for each, and required participants to turn in a punched card with their picks each week. This marked the end of Paula's reign as FBP champ, but set the FBP firmly on its path to being the technology showcase it is today.

A year or so later, FBP made its first technology migration. While the Purdue computing environment was still based on batch job submissions, it did support a teletype interface. This wasn't really a timesharing interface; for the most part, you'd use a simple text editor to type up the same stuff you would have put on punchcards, then submit it as a batch job. But there was one simple timesharing environment supported: ALFIE (Algebraic Language for an Interactive Environment), which provided a superset of BASIC with FORTRAN-like formatting. FBP was rewritten to use ALFIE, and for the first time it was possible to make picks in an interactive environment, even if it was pretty dang minimal. The biggest issue in making the move to ALFIE was figuring out a protection scheme that prevented grad students from reading (or worse yet, altering) the database containing pick information. Eventually, an elaborate scheme involving passworded batch files that invoked other executable (but not readable) batch files was devised and apparently worked successfully (or at least no miscreants were apprehended).

Around 1978 or '79, the Computer Science Department made an important acquisition: A VAX 11/780 from Digital Equipment Company. This was the first computer that the department actually owned (previously all computing was done via the Purdue University Computing Center, which served the entire campus). And it had a real timesharing system, running under the VAX/VMS operating system. The opportunity was too good to pass up; FBP got rewritten in FORTRAN and ran on the VAX with an appropriately modified security scheme. I think it was somewhere around in here that the program's feature set also began to grow (e.g., the ability to see the past history of each team as you prepared to make a pick, and FBP's idea of which team is the favorite in each game).

Then along came Unix. In 1980, the CS faculty sensibly decided to run BSD Unix instead of VMS on the department VAX. There was only one problem: the football season was starting, and there was insufficient time to get the Pool moved to the new operating environment. Fortunately, my research was on performance analysis of paging in the VAX/VMS environment. Since my research was already under way when the move to Unix took place, I was allowed to "sign up" for the machine, shut down Unix, and reboot it using VMS so that I could run the necessary experiments and measurements. Nobody said that we couldn't run other software ... so throughout the season I'd bring up VMS at designated times, let people make picks for a few hours, work on my research, and then return the system to Unix. Another bullet dodged.

The Monday Night Scoring Formula also got an update around this time. While I don't remember the precise incident that precipated the investigation, controversy had erupted about the algorithm used in the Monday Night Competition. In response, a team of crack numerical analysts studied the problem for the better part of a week and devised the formula that is still used today. The formula is based on a distance metric, modified by the belief that getting the point spread right is more important than guessing the specific score of either individual team, and that high-scoring games are harder to predict accurately than low-scoring games. The amount of energy that went into devising the Monday Night Scoring Formula and the number of rewrites that had already been done on FBP software are possible indicators of why so many Computer Science PhD students were taking 5-6 years to graduate.

To the astonishment of many, I completed my thesis, left Purdue in the summer of 1981, and headed for Intel Corporation in Portland, Oregon taking a copy of the FBP source code with me. During that first fall in Portland, the source code remained on its 9-track tape reel as I adapted to, well, so many things. However, that didn't mean FBP participation ceased. My computing account at Purdue was still live, and Dave Capka was running the Pool (having now moved it to Unix). So every Friday afternoon in the fall of 1981 (this was before Thursday night games had intruded on the schedule), Kevin Kahn, Fred Pollack, and I would gather around our 10 cps teletype and use our 300 baud modem to call up Purdue and enter picks. To our dismay, we failed to win. Being the gracious losers that we are, we shrugged, mailed off our dollar to Capka to buy pizza for the winners, and went on building our whizzy new operating system. A couple of weeks later, we received a package in the Intel mail. Inside we found a clear tape case for a large 9-track tape. The case, which had been hermetically sealed with half a roll of scotch tape, contained several slices of pizza with a note: "Wanted to make sure you got your share of the pizza".

The Jan 1989 award ceremony was classy!

Michael Swieczkowski, one of the Siemens gang and 1985 season champ

Since it was clear that annual receipt of a package of potentially virulent foodstuffs would probably violate some Intel policy or other, I decided it was time to move the Pool to Portland. In the summer of 1982 I cleaned up the FORTRAN that I had brought from Purdue (it had accumulated a fair amount of cruft over the years from feature additions), moved it onto our Unix system, and we opened for business. The champions and stats in the FBP Hall of Fame date back to this year. Intel people took to the Pool just as eagerly as Purdue people had (heck, a bunch of them were Purdue people!). More interestingly, in those years I was working as part of a joint development project between Intel and Siemens Corporation, so we immediately signed up hordes of Germans and Austrians who were eager for the full American Experience even if they had no flaming idea how football (at least the American variety) was played. Apparently American prowess at pigskin prognostication did not prove a significant hurdle; Manfred Koch won the Weekend Competition in 1983 despite never having seen an American football game, and other Siemens folks appear in the Hall of Fame up through the point where the project ended in 1989. Throughout the years, we've had several instances in which software-based participants outpaced wetware-based participants to earn titles; in both 1985 and 1996, Home Teams (which just chooses the home team by a 17-16 score) trounced all comers to win the Monday Night Competition. (We threatened to mount a pizza on the VAX disk drive in 1985 to reward it.) To preserve the fragile egos of homo sapiens against this electronic onslaught, we also award a title to the highest-scoring flesh-and-blood participant in such years.

1993 champs and perennial foes Priscilla Richardson & Tom Dingwall

Over the next decade or so I continued to add improvements and features, moved the Pool to a couple more systems, and eventually rewrote it in C to make it more maintainable. In 1992, Janet earned her CPA and began working for Talbot, Korvola, and Warwick. She told them about FBP and they immediately wanted in. One small problem: there was no way for them to log into the Intel system where the Pool ran. Time for yet another rewrite, which gave us compatible FBP versions that ran on Unix and on DOS. For the next few years, Janet installed the FBP software on a PC at TKW every fall, then shuttled a floppy disk back and forth every week so that we could keep the databases merged. The awkwardness of this solution apparently didn't faze the accountants (especially Priscilla [right] and Pat Richardson), who immediately began to kick engineering butt and regularly win every title under contention.

The floppy-shuttling continued for the next few years until Janet left TKW. While the company was willing to let her take another job, they were not willing to drop out of the Football Pool and demanded that we find a way for them to participate. Trouble was also brewing at Intel, where I was getting increasingly nervous about maintaining an account on an engineering Unix system just to run FBP every fall (our department had moved to workstations and PCs by then). Fortunately, Tim Berners-Lee anticipated our difficulties and invented the World Wide Web to solve our problems. I did a major rewrite of the FBP software so that it generated HTML pages from the database each week and we went live online at CompuServe in August 1997. This not only kept the TKW accountants in the game but expanded our accounting base; Pat Richardson (who worked at Moss Adams, not TKW) not only stopped depending on Priscilla to enter his picks but also signed up new droves of Moss Adams people who have remained staunch FBP supporters. From time to time, we even regained old friends who had moved to other parts of the country, including a few of the Siemens folks from the early days of FBP. And now you're here too.

There are still improvements waiting to be made. I'll eventually do a rewrite that puts the live database on the web so you'll get immediate feedback about whether you typed your password correctly, can recheck your picks, and can see what other players have chosen once the games begin. And I annually threaten to resurrect the plotting software that shows you why you got such a lousy score for your Monday Night pick ... but those pleasures await another time. In the meantime, there's plenty of room for the names of new winners on the official FBP Trophy (we added a brand new engraving plate to the back in 2003). So enjoy the fierce but friendly competition, shudder deliciously at the frisson that runs up your back when a last-second field goal rescues a pick for you, and (if you win) bask in 20+ years of historic FBP glory. See ya on the field.

Official FBP Trophy

	Weekend Competition Champions
	Monday Night Champions
	Buster Dunsmore Memorial Awards

Weekend Competition Champions

Monday Night Champions

Buster Dunsmore Memorial Award

FBP Home Page	Make Picks	Standings	Getting Started	History