Evaluation of the Impact of Errors in the Sorting of Pigs for Market on Sort Loss and Optimal Market Weight at a Range of Marketing Ages - Yichen Que, from the 2017 Allen D. Leman Swine Conference, September 16-19, 2017, St. Paul, Minnesota, USA. More presentations at http://www.swinecast.com/2017-leman-swine-conference-material

1. Y. Que*, F. A. Cabezon and A. P. Schinckel
Department of Animal Sciences
Purdue University
Evaluation of the Impact of Errors in the
Sorting of Pigs for Market on Sort Loss at a
Range of Marketing Ages

2. INTRODUCTION
• Most commercial pork producers visually
evaluate the BW of each pig and try to
identify the heaviest pigs for marketing
on multiple marketing days to reduce CW
discounts, and target the optimal market
BW (Li et al, 2003; Boys et al., 2007; Flohr
et al., 2015).
• On large farms, pig sorting-marketing
crews target a specific number of heavy
pigs in each pen to be marketed each
marketing day (McBride and Key, 2003).
Table 1. Carcass weight discount rates for
different carcass weight classes.
Carcass weight, kg Discount, $/kg
< 68.5 0.441
68.5-73.0 0.286
73.0-75.3 0.176
75.3-77.6 0.121
77.6-82.1 0.077
82.1-107.0 0
107.0-109.3 0.0661
109.3-111.6 0.2425
111.6-113.9 0.2866
113.9-116.1 0.3307
> 116.1 0.3748
Indiana Packers Corporation (2015).

3. INTRODUCTION (CONTINUED)
• Errors in the visual assessment each pig’s BW result in marketing
errors (Alschwede and Jones, 1992). Thus, 2 types of pig marketing
errors exist: errors in the estimation of BW for the pigs that are
visually evaluated and the percentage of pigs that are not visually
evaluated (Cabezon et al., 2016).
• When the actual CW and sort loss data were evaluated for three large
wean-finish barns, a barn with the greatest mean CW, close to the
upper acceptable CW, had much greater mean sort loss than the
other 2 barns at the same approximate sorting accuracy (Que et al.,
2016).

4. OBJECTIVES
1) Evaluate the impact of sorting errors on sort loss at
different mean carcass weights (CW).
2) Demonstrate that the magnitude of sort loss due to
inaccurate sorting is affected by the pigs’ mean CW.

5. MATERIALS AND METHODS
• The BW growth curves for twenty-five 4,000-head wean to finish
barns were simulated. Pigs were assigned to 32 pens of 125 pigs in
each barn. Pigs were modeled to be born over a 5-d interval with 20%
of the pigs born each day to reflect the range in age in the larger
commercial production systems.
• A marketing strategy was simulated to represent that currently used
by pork producers. 25% of the pigs were targeted to be marketed at
169 d, 25% at 179 d, and the remaining pigs marketed at 193 d of age.

6. MATERIALS AND METHODS (CONTINUED)
• BWi, t = WT0 + {[(WF + wfi) – WT0](t/K)C}/[1 + (t/K)C] (Lopez et al.,
2000; Schinckel et al., 2009a, 2012a).
• CW = (1 + 0.02z2) × 0.721(BW) 1.0061 (Schinckel et al., 2012b).
• Pig-specific random effects (wfi) were generated as 30 times a value
sampled a standard normal distribution (z1 mean = 0 and SD = 1).
• Four BW estimation error (BWEE) rates were simulated to represent
zero, low, average, and high levels of visual assessment of BW
(Ahlschwede and Jones, 1992). Predicted BW = [1 + (proportional
error rate × z3) × actual BW], where the proportional error rates are
0, 0.04, 0.06, and 0.08 for each level of BWEE.

7. 0
20
40
60
80
100
120
140
160
180
0 50 100 150 200 250
BW,kg
Age, d

8. MATERIALS AND METHODS (CONTINUED)
• Each pig was randomly assigned to be evaluated for BW or not based
on values sampled from standard normal distributions.
• Marketing ages for the pigs were shifted in weekly intervals with
mean ages of 155.5, 162.5, 169.5, 176.5, 183.5, 190.5, 197.5, 204.5
and 211.5 d of age.
• Two variables: number of pigs with sort loss and mean sort loss per
pig in the barn were fitted to a model including the fixed effects of
level of marketing age (AGE), BWEE, PNVE, their interactions and
random effect of replicate barn using the MIXED procedure of SAS®.

9. RESULTS AND DISCUSSION
Table 2. Mean and SD of overall BW and CW at a range of marketing
ages with accurate sorting.
BW CW
Mean Age, d Mean, kg SD Mean, kg SD
155.5 104.67 6.67 75.49 5.07
162.5 111.01 7.07 80.09 5.38
169.5 117.20 7.47 84.58 5.68
176.5 123.21 7.85 88.95 5.97
183.5 129.03 8.22 93.17 6.26
190.5 134.65 8.58 97.26 6.53
197.5 140.08 8.92 101.21 6.80
204.5 145.30 9.26 105.00 7.05
211.5 150.32 9.58 108.65 7.30

10. RESULTS AND DISCUSSION
Table 3. The mean sort loss ($/pig), mean carcass weight (CW) and SD for carcass weight
for 4 levels of sorting accuracy.
BWEE, PNVE = 0,0 BWEE PNVE = 8,0 BWEE, PNVE = 0, 24 BWEE, PNVE = 8,24
CW CW CW CW
Mean
Age, d
Mean
Sort
Loss
Mean,
kg
SD
Mean
Sort
Loss
Mean,
kg
SD
Mean
Sort
Loss
Mean,
kg
SD
Mean
Sort
Loss
Mean,
kg
SD
155.5 11.42 75.49 5.07 11.97 75.55 6.00 11.3 75.85 5.83 11.67 75.98 6.57
162.5 6.08 80.09 5.38 6.56 80.16 6.37 6.12 80.48 6.18 6.48 80.61 6.97
169.5 2.55 84.58 5.68 3.25 84.66 6.73 2.78 84.99 6.53 3.34 85.13 7.36
176.5 1.18 88.95 5.97 1.42 89.02 7.07 1.44 89.37 6.87 1.72 89.53 7.74
183.5 0.79 93.17 6.26 1.13 93.26 7.41 1.38 93.62 7.19 1.79 93.78 8.11
190.5 1.12 97.26 6.53 2.43 97.35 7.74 2.45 97.73 7.51 3.59 97.9 8.47
197.5 2.96 101.21 6.80 5.88 101.29 8.05 5.13 101.69 7.81 7.33 101.87 8.81
204.5 10.13 105.00 7.05 12.37 105.09 8.35 11.48 105.51 8.11 13.67 105.69 9.14
211.5 19.47 108.65 7.30 19.19 108.75 8.64 19.55 109.18 8.39 20.11 109.36 9.46

11. RESULTS AND DISCUSSION
Figure 4. Relationship of mean sort loss of the first marketing cut to mean age with four
levels sorting accuracy with 2 levels of percentage of pigs not visually evaluated (PNVE = 0
and 24%) and 2 levels of BW estimation error (BWEE = 0 and 8%).
0.00
5.00
10.00
15.00
20.00
25.00
30.00
150 160 170 180 190 200 210
Meansortlossperpig,$
Mean age, d
BWEE, PNVE = 0, 0 BWEE, PNVE = 8, 0
BWEE, PNVE = 0, 24 BWEE, PNVE = 8, 24

12. RESULTS AND DISCUSSION
Figure 5. Relationship of mean sort loss of the second marketing cut to mean age with four
levels sorting accuracy with 2 levels of percentage of pigs not visually evaluated (PNVE = 0
and 24%) and 2 levels of BW estimation error (BWEE = 0 and 8%).
0.00
5.00
10.00
15.00
20.00
25.00
150 160 170 180 190 200 210
Meansortlossperpig,$
Mean age
BWEE, PNVE = 0, 0 BWEE, PNVE = 8, 0
BWEE, PNVE = 0, 24 BWEE, PNVE = 8, 24

13. RESULTS AND DISCUSSION
Figure 6. Relationship of mean sort loss of the third marketing cut to mean age with four
levels sorting accuracy with 2 levels of percentage of pigs not visually evaluated (PNVE = 0
and 24%) and 2 levels of BW estimation error (BWEE = 0 and 8%).
0.00
5.00
10.00
15.00
20.00
150 160 170 180 190 200 210
Meansortlossperpig,$
Mean age, d
BWEE, PNVE = 0, 0 BWEE, PNVE = 8, 0
BWEE, PNVE = 0, 24 BWEE, PNVE = 8, 24

14. RESULTS AND DISCUSSION
Figure 1. Relationship of mean sort loss to mean age with four levels of BW estimation
error (BWEE = 0, 4, 6 and 8%) and all pigs visually evaluated (PNVE = 0%).
0.00
5.00
10.00
15.00
20.00
150 160 170 180 190 200 210
Meansortlossperpig,$
Mean age, d
BWEE, PNVE = 0, 0 BWEE, PNVE = 4, 0
BWEE, PNVE = 6, 0 BWEE, PNVE = 8, 0

15. RESULTS AND DISCUSSION
Figure 2. Relationship of mean sort loss to mean age with four levels of percentage of pigs
not visually evaluated (PNVE = 0, 8, 16 and 24%) and no BW estimation error (BWEE = 0%).
0.00
5.00
10.00
15.00
20.00
150 160 170 180 190 200 210
Meansortlossperpig,$
Mean age, d
BWEE, PNVE = 0, 0 BWEE, PNVE = 0, 8
BWEE, PNVE = 0, 16 BWEE, PNVE = 0, 24

16. RESULTS AND DISCUSSION
Figure 3. Relationship of mean sort loss to mean age with four levels sorting accuracy with 2
levels of percentage of pigs not visually evaluated (PNVE = 0 and 24%) and 2 levels of BW
estimation error (BWEE = 0 and 8%).
0.00
5.00
10.00
15.00
20.00
150 160 170 180 190 200 210
Meansortlossperpig,$
Mean age, d
BWEE, PNVE = 0, 0 BWEE, PNVE = 8, 0
BWEE, PNVE = 0, 24 BWEE, PNVE = 8, 24

17. RESULTS AND DISCUSSION
• The main effects of AGE, BWEE, and PNVE, and AGE x PNVE, AGE x BWEE, and AGE x BWEE x
PNVE interactions impacted both variables (P < 0.001) The effects of BWEE and interaction of
BWEE x PNVE impacted (P < 0.001) both variables at all ages.
• The difference in sort loss/pig produced by the least accurate sorting (BWEE = 8% and PNVE =
24%) increased as the mean CW increased from $1.00 at 93 kg to $4.53 at 103 kg.
• Sort loss/pig increased more rapidly with increased CW at higher levels of BWEE and PNVE.
• The effect of inaccurate sorting to increase sort loss is minimized when the mean CW is close
to the middle of the pork processor’s acceptable CW range and increases as CW increases to
those approaching the upper acceptable CW range and is dependent on the marketing grid.

18. IMPLICATIONS
The impact of less accurate sorting of pigs for market on sort loss is
highly dependent on the mean CW of the pigs. There is no single value
that can be assigned to the impact of inaccurate sorting of market pigs,
it is highly dependent on the CW of the pigs being marketed. Current
marketing grids encourage pork producers with less accurate sorting of
pigs to market their pigs in the middle of the pork processor’s
acceptable CW range. The current marketing grids do not provide any
direct incentive to reduce the variation in CW that could result with
more accurate sorting of pigs for market.

19. ACKNOWLEDGEMENTS
• National Pork Board
• Swine Research and Education Experience (SREE) grant
• Purdue Univeristy
• Dr. Allan Schinckel
• Dr. Francisco Cabezon

20. Questions?