Teaching DOE with the DOE-Golfer
An interview with Dr. Leonard Lye

Why did you invent the DOE-Golfer?

The inspiration came from the example that was used in Douglas Montgomery's book where the game of golf was used as an example for an experiment. Why not invent a toy that can be used to teach DOE via golf? Golf is a universal game and many people know about the game and some are really fanatical about it. I feel that it has wider appeal than table hockey or the catapult.

How do you compare the DOE-Golfer with the catapult and table hockey?

I have used the table hockey experiment described in "DOE Simplified" by Anderson and Whitcomb (2000) and I have also used the catapult toy described in "Understanding Industrial Designed Experiments" by Schmidt and Launsby (1994). I find that the catapult appeals more to the younger students (first year) but not to "older" students. Table hockey is great as well, but students from overseas and southern climes do not find it as fascinating. The DOE-Golfer, however, seems to appeal to all students both "young" and "old". Everybody wants to be as good as Tiger Woods, I guess.

How many factors can be varied in the DOE-Golfer?

I can think of at least six: 1) length of club, 2) angle of swing, 3) weight of club, 4) type of ball, 5) type of green or carpet, and 6) direction. The first three are the major ones if you use only one type of ball, play on the same carpet, and if the floor and carpet are flat and uniform in all directions.

How about competing with the DOE-Golfer?

For bonus marks in my graduate DOE course, I organized a golf tournament to be played in our carpeted faculty lounge. During the Fall term, there were 24 students divided into 6 groups of 4. They were given two weeks to "train" for the tournament. During this period, the groups had to conduct the appropriate experiments and arm themselves with the equation(s) they would need for the tournament. The tournament consisted of two components: 1) Lines—masking tape was stuck on the carpet in three places along the North-South and East-West directions. The number of strokes to get to within 0.5 inches of each of the three lines was then totaled. 2) Holes—Three-hole golf challenge. Plastic golf holes were put at three places in the lounge and the students had to get the ball in the holes. The number of strokes to get the ball in all three holes was totaled. The group with the least number of strokes overall was the winner.

What designs were used and did the groups use any software?

I left that up to the groups as to what designs they would use. Because it was a competition and the winning group got bonus marks, most groups were very careful and secretive about their experiments. Most started with a simple two-level experiment with three main factors, then checked for curvature, and finally augmented the two-level design to a response surface design. Some chose a face-centered CCD, and some used Box-Behnken. Some used replications and some lazy groups used only one. All groups used Design-Expert® from Stat-Ease, Inc. The one-month free use was a great help.

Were there any surprises?

I guess that the students were surprised at the results. The group that won the tournament clearly had the best design and was the most meticulous. They noticed that direction was important. Going North to South is different from going South to North. Also, the East-West direction is different from the North-South direction. They took this into account and that made quite a difference. They even got a hole-in-one for one hole and the maximum number of strokes they needed for any hole was two. It was amazing really. Some groups needed more than 5 or 6 strokes for one hole. The moral here is not to assume that the floor is flat or the carpet is uniform in all directions.

Was it a valuable learning experience?

It was an overwhelming success I must say. The students really enjoyed it and they have learned much about both DOE and golf. After the tournament, the winners were asked to share their secrets of their success and every group was asked: If they had a chance to compete again, what would they do differently? Almost everybody agreed that they should have done more replications and should have considered combinations that would have given short distances. Their first stroke usually got them fairly close to the line or hole (maybe 2 inches away), and then they needed a combination for 2 inches. Some groups completely forgot about this possibility. Also, if they overshot in the first stroke, the next stroke was in the opposite direction. If they did not consider direction as a factor, then again their prediction was not accurate. The winning group considered all these possibilities - hence they won.

Do you have any pictures from the tournament?

Yes. I have included some taken during the tournament (see below). In the future, I will videotape the tournament for the sports channel!

Where can one get a DOE-Golfer?

The DOE-Golfer is beautifully made of oak and can also be made of maple or some other hard wood. It sells for US $220 (including shipping). Contact me if you want one. I would suggest getting at least two for class use. Send me a money order and allow about 4 weeks for delivery.

Dr. Leonard Lye
Professor of Civil Engineering
Faculty of Engineering and Applied Science
Memorial University
St. John's, Newfoundland, A1B 3X5, Canada
llye@mun.ca

References:

Anderson M. J. and P. J. Whitcomb (2000): DOE Simplified: Practical Tools for Effective Experimentation, Productivity Inc.

Schmidt, S.R. and R.G. Launsby (1994): Understanding Industrial Designed Experiments, 4th Edition, Air Academy Press, Colorado.