

Table 24. Predictability of Market Advisory Program Performance by Winner and Loser Categories Between Pairs of Adjacent Crop Years, Corn, Soybeans and 50/50 Revenue, 19952000 Crop Years  
Corn  Soybeans  50/50 Revenue  
Twotail  Twotail  Twotail  
pvalue  pvalue  pvalue  
Year  Year  Winner  Loser  for Fisher's  Winner  Loser  for Fisher's  Winner  Loser  for Fisher's  
t  t+1  t+1  t+1  Exact Test  t+1  t+1  Exact Test  t+1  t+1  Exact Test  
number of programs  number of programs  number of programs  
1995  1996  Winner t  5  6  Winner t  6  5  Winner t  7  4  
Loser t  6  5  1.00  Loser t  5  6  1.00  Loser t  4  7  0.39  
1996  1997  Winner t  7  5  Winner t  6  5  Winner t  6  5  
Loser t  5  7  0.68  Loser t  5  6  1.00  Loser t  5  6  1.00  
1997  1998  Winner t  6  5  Winner t  6  4  Winner t  3  7  
Loser t  5  7  0.68  Loser t  4  7  0.39  Loser t  7  4  0.20  
1998  1999  Winner t  7  4  Winner t  7  3  Winner t  6  4  
Loser t  4  7  0.39  Loser t  3  8  0.09  Loser t  4  7  0.39  
1999  2000  Winner t  8  4  Winner t  8  4  Winner t  9  3 


Loser t  4  9  0.12  Loser t  4  8  0.22  Loser t  3  9  0.04  
19952000  Winner t  33  24  Winner t  33  21 

Winner t  31  23  
Total  Loser t  24  35  0.09  Loser t  21  35  0.02  Loser t  23  33  0.13  
Note: The selection strategy consists of ranking programs by net advisory price in the first year of the pair (e.g., t = 1995) and then forming two groups of programs: "winners" are those services in the top half of the rankings and "losers" are services in the bottom half. Next, the same programs are ranked by net advisory price for the second year of the pair (e.g., t+1 = 1996), and again divided into "winners" and "losers." For a given comparison, advisory programs must fall in one of the following categories: winner twinner t+1, winner tloser t+1, loser twinner t+1, loser tloser t+1. If advisory program performance is unpredictable, approximately the same counts will be found in each of the four combinations. Fisher’s Exact Test is the appropriate statistical test because both row and column totals are predetermined in the 2 x 2 contingency table formed on the basis of winner and loser counts. Two stars indicates significance at the one percent level and one star indicates significance at the five percent level.  