Introduction

WASDE reports get published around the 10th of every month and is a date that every commodity trader looks forward to. This is the time of the month when the USDA gives the market a glipse of their take on the current balance sheets of the main agricultural commodities of the major producing and consuming nations. A classic number that many commodity traders look at is the stock-to-usage ratio. This is a handy shorthand that summarised the entiry balance sheet of a commodity within a particular country. The greater the stock-to-usage number

• the higher the number of stocks,
• the cheaper the commodity and
• the more contango the futures curve.

The main idea of this write-up is to express the values of the different calendar spread under certain stock-to-usage conditions. The plot below shows an example of the Corn Sep-Dec calendar spread since 1999. The x-axis shows the decile buckes of the stock-to-usage calculated for every WASDE report. The values of the spreads between two consecutive WASDE reports are associated with the first. This is because we argue that the fundamental truth is given by the USDA report and the subsequent price moves should reflect this new fundamental truth. Within each of the decile buckes we determine the 5th, 25th, 50th, 75th and 95th percentile of the spread. In the plot below the darker shaded region represents the spread lying in the 25th to 75th percentile, while the lighter shade represents those spreads lying in the 5th to 95th percentile respectivley. For the remainder of this write-up we will regard calendar spreads to be undervalued (overvalued) when the spread, at the close of a WASDE date, is below (above) the 25th (75th) percentile.

We have to be careful to avoid look ahead bias. To this end, on each WASDE publucation date, we use the following methodology:

• Only consider the publications from the ten years prior
• Divide those stock-to-usage numbers into deciles
• Calculate the percentile of the current spread within the decile it belongs to

An event is triggered when the calculated percentile is greater than 75 or less than 25. After these overvalued and undervalued events have been triggered we determine the statistics of the spreads on successive days before and after the trigger. What we hope to see is that overvalued spreads decrease and that undervalued spreads increase.

In the following we consider the spread evolutions of

• Corn (C)
• Chicago Wehat (W)
• Kansas City Wheat (KW)
• Minneapolis Wheat (MW)
• Rough Rice (RR)
• Soybeans (S)
• Soybean Meal (SM)
• Soybean Oil (BO)

under the events mentioned. The results will be in the form of event plots where we show the day, as indicated by bin, on the x-axis and the statistics withing each bin on the y-axis. The y-axis is expressed in the same units as the underlying spread. In most of the cases we consider this is USd/bu. Throughout the shaded region shows the 25th to 75th percentiles. The vertical red line highlights the zero bin, this corresponds to the day of the event, i.e. WASDE publication dates. For the reader who prefers a table we show the median change in the spread from bin zero to every 5th bin in a table below each plot.

We find that the results are not uniform but differ by commodity and calendar spread considered.

Corn

By eye it might be a little tricky to see which events performed better than the others. The table below shows the same results as the image above. From the image and the table you can see that the moves are sliglty more pronounced in the percentile > 75 case, i.e. it seems like spreads that are high with respect to current stock levels correct more quickly compared the spread that are suppressed. For the percentile < 25 case we see that the ZN calendar spreads gave the best median moves after the event. The NZ spread gave the largest magnitude move in the percentile >75 case.

Table 1: USd change from event date.
group calRef 5 10 15 20
percentile < 25 HK 0.000 0.000 -0.250 0.125
percentile < 25 HN 0.000 0.000 -0.500 0.250
percentile < 25 KN -0.125 0.000 -0.250 0.000
percentile < 25 NU 0.250 0.250 0.250 0.375
percentile < 25 NZ 0.500 0.500 0.375 1.250
percentile < 25 UH 0.000 -0.500 -0.250 0.500
percentile < 25 UZ 0.250 0.000 -0.125 0.000
percentile < 25 ZH 0.250 -0.250 0.000 -0.250
percentile < 25 ZN 0.250 -0.125 -0.250 -0.500
percentile > 75 HK 0.000 -0.500 -0.750 -0.750
percentile > 75 HN 0.250 -0.500 -1.125 -1.000
percentile > 75 KN 0.000 -0.250 -0.500 -0.500
percentile > 75 NU 0.250 -0.250 -0.750 -1.750
percentile > 75 NZ 0.125 -0.125 -0.750 -2.750
percentile > 75 UH 0.250 -0.500 -0.500 -0.500
percentile > 75 UZ 0.000 -0.375 0.000 0.000
percentile > 75 ZH 0.000 -0.250 -0.125 -0.375
percentile > 75 ZN 0.250 0.250 0.000 -1.000

The table below shows the total events triggered for each of the Corn calendar spreads in the percentile < 25 and percentile >75 groups. It is interesting to note the imbalance in the data. There are many more instances of spreads trading at a premium with respect to current stock-to-usage numbers compared to those trading at a discount.

Table 2: Total number of events by calRef.
calRef percentile < 25 percentile > 75
HK 60 51
HN 59 42
KN 57 42
NU 30 41
NZ 29 36
UH 43 43
UZ 45 44
ZH 56 43
ZN 59 45

The table below gives a summary of the fraction of calendar spreads that ended up in the money after the indicated number of trading days after the WASDE report date. In the case of percentile < 25 a spread is defined to be in the money when the change (delta) is a positive number, i.e. the spread moved in a backwardated direction. Conversily, in the percentile > 75 case the spread end in the money when the change is negative, i.e. the spread moved more contango.

Table 3: Fraction of events ending in the money.
group calRef 5 10 15 20
percentile < 25 HK 0.45 0.45 0.37 0.35
percentile < 25 HN 0.39 0.44 0.36 0.41
percentile < 25 KN 0.33 0.37 0.21 0.32
percentile < 25 NU 0.50 0.60 0.47 0.37
percentile < 25 NZ 0.52 0.62 0.52 0.41
percentile < 25 UH 0.44 0.37 0.35 0.35
percentile < 25 UZ 0.51 0.47 0.36 0.31
percentile < 25 ZH 0.55 0.34 0.39 0.32
percentile < 25 ZN 0.59 0.42 0.36 0.36
percentile > 75 HK 0.39 0.55 0.61 0.67
percentile > 75 HN 0.38 0.64 0.64 0.64
percentile > 75 KN 0.40 0.60 0.57 0.50
percentile > 75 NU 0.44 0.51 0.56 0.54
percentile > 75 NZ 0.44 0.50 0.53 0.56
percentile > 75 UH 0.44 0.56 0.51 0.47
percentile > 75 UZ 0.45 0.55 0.45 0.41
percentile > 75 ZH 0.49 0.56 0.47 0.42
percentile > 75 ZN 0.38 0.44 0.44 0.44

From the different outcomes for each of the percentile threshold groups we can determine the expected value of the spread at differnet bins. Mathematically we can write

$E[X] = \sum^{k}_{i=1} x_i p_i$ where $$X$$ is a random variable with a finite number of finite outcomes $$x_1, \dots, x_k$$ occurring with probabilities $$p_1, \dots, p_k$$, respectively. The expected value for each of the spreads at all the bins are shown in the plot below. This point of view is interesting because it combines the magnitude of the move with the probability of that move into a single image. From here is become easier to see which of the spreads reacts the best.

In the following sections we omit explanations as the interpretation follows from the Corn expample.

Chicago Wheat

Table 4: USd change from event date.
group calRef 5 10 15 20
percentile < 25 HK 0.000 0.500 0.250 0.000
percentile < 25 HN 0.125 0.375 -0.250 0.250
percentile < 25 KN 0.250 0.250 -0.125 -0.250
percentile < 25 NU 0.375 -0.125 0.000 -0.250
percentile < 25 NZ -0.125 -0.125 -0.750 -0.750
percentile < 25 UH 0.000 -1.000 -0.500 -0.250
percentile < 25 UZ 0.000 0.000 -0.500 -0.750
percentile < 25 ZH 0.250 0.250 0.125 0.250
percentile < 25 ZN 1.000 2.000 1.125 -0.750
percentile > 75 HK -0.500 -1.625 -1.750 -1.625
percentile > 75 HN -1.750 -2.000 -2.875 -2.875
percentile > 75 KN -1.000 -1.750 -2.750 -2.000
percentile > 75 NU 0.000 -0.625 -0.875 -0.500
percentile > 75 NZ 0.250 -1.000 -2.250 -0.750
percentile > 75 UH -0.125 -1.625 -1.750 0.875
percentile > 75 UZ 0.250 -0.375 -1.375 -1.250
percentile > 75 ZH 0.250 -0.500 -1.250 -1.500
percentile > 75 ZN 0.000 -2.250 -3.500 -11.500
Table 5: Total number of events by calRef.
calRef percentile < 25 percentile > 75
HK 37 34
HN 31 39
KN 30 47
NU 43 30
NZ 37 29
UH 28 26
UZ 40 26
ZH 34 35
ZN 29 33
Table 6: Fraction of events ending in the money.
group calRef 5 10 15 20
percentile < 25 HK 0.46 0.59 0.46 0.35
percentile < 25 HN 0.48 0.48 0.35 0.35
percentile < 25 KN 0.53 0.57 0.33 0.30
percentile < 25 NU 0.58 0.44 0.47 0.33
percentile < 25 NZ 0.41 0.43 0.46 0.38
percentile < 25 UH 0.46 0.39 0.32 0.36
percentile < 25 UZ 0.42 0.40 0.28 0.22
percentile < 25 ZH 0.53 0.53 0.41 0.41
percentile < 25 ZN 0.66 0.66 0.45 0.31
percentile > 75 HK 0.53 0.62 0.59 0.53
percentile > 75 HN 0.67 0.59 0.62 0.59
percentile > 75 KN 0.55 0.60 0.68 0.57
percentile > 75 NU 0.47 0.57 0.60 0.50
percentile > 75 NZ 0.38 0.59 0.62 0.48
percentile > 75 UH 0.50 0.65 0.54 0.35
percentile > 75 UZ 0.46 0.58 0.58 0.38
percentile > 75 ZH 0.46 0.57 0.63 0.49
percentile > 75 ZN 0.48 0.58 0.61 0.33

Kansas City Wheat

Table 7: USd change from event date.
group calRef 5 10 15 20
percentile < 25 HK 0.250 -0.250 0.000 0.000
percentile < 25 HN 0.000 -0.250 -0.125 0.125
percentile < 25 KN -0.250 -0.250 -0.250 -0.500
percentile < 25 NU 0.250 0.250 0.250 0.000
percentile < 25 NZ 0.750 0.000 0.500 -0.250
percentile < 25 UH 0.500 0.625 0.375 0.250
percentile < 25 UZ 0.000 0.000 0.500 -0.250
percentile < 25 ZH 0.250 0.000 0.250 0.250
percentile < 25 ZN -0.250 -0.750 -0.750 -1.250
percentile > 75 HK -0.625 -1.000 -0.500 -1.500
percentile > 75 HN -3.875 -2.125 -3.250 -7.000
percentile > 75 KN -0.625 -1.125 -0.750 -3.500
percentile > 75 NU -1.000 -0.750 -1.250 -3.000
percentile > 75 NZ -1.625 -2.750 -3.375 -6.250
percentile > 75 UH -2.500 -3.000 -3.375 -3.750
percentile > 75 UZ -1.125 -0.875 -1.500 -4.500
percentile > 75 ZH -0.375 -0.875 -2.250 -2.250
percentile > 75 ZN -4.000 -3.500 -4.500 -14.125
Table 8: Total number of events by calRef.
calRef percentile < 25 percentile > 75
HK 44 16
HN 34 22
KN 42 34
NU 70 11
NZ 64 12
UH 36 12
UZ 56 12
ZH 45 24
ZN 25 27
Table 9: Fraction of events ending in the money.
group calRef 5 10 15 20
percentile < 25 HK 0.52 0.36 0.30 0.30
percentile < 25 HN 0.44 0.32 0.35 0.29
percentile < 25 KN 0.38 0.31 0.26 0.17
percentile < 25 NU 0.56 0.56 0.53 0.40
percentile < 25 NZ 0.61 0.48 0.53 0.39
percentile < 25 UH 0.56 0.61 0.53 0.42
percentile < 25 UZ 0.46 0.46 0.48 0.32
percentile < 25 ZH 0.53 0.49 0.47 0.47
percentile < 25 ZN 0.28 0.32 0.28 0.24
percentile > 75 HK 0.62 0.81 0.50 0.62
percentile > 75 HN 0.73 0.64 0.55 0.68
percentile > 75 KN 0.62 0.62 0.59 0.59
percentile > 75 NU 0.64 0.55 0.73 0.45
percentile > 75 NZ 0.58 0.75 0.67 0.58
percentile > 75 UH 0.75 0.75 0.83 0.58
percentile > 75 UZ 0.75 0.58 0.50 0.50
percentile > 75 ZH 0.62 0.58 0.58 0.42
percentile > 75 ZN 0.63 0.59 0.67 0.59

Minneapolis Wheat

Table 10: USd change from event date.
group calRef 5 10 15 20
percentile < 25 HK 0.75 0.25 0.750 1.250
percentile < 25 KN 0.50 0.50 0.500 0.875
percentile < 25 NU 0.75 0.50 0.125 0.500
percentile < 25 UZ 0.75 0.50 1.250 1.625
percentile < 25 ZH 0.50 0.00 -0.125 0.750
percentile > 75 HK -0.75 -2.25 -2.250 -1.500
percentile > 75 KN -0.50 -1.00 -1.125 -1.250
percentile > 75 NU 0.00 -0.25 -0.750 -1.500
percentile > 75 UZ -0.50 -1.25 -1.750 -1.250
percentile > 75 ZH -0.50 -0.75 -2.250 -3.500
Table 11: Total number of events by calRef.
calRef percentile < 25 percentile > 75
HK 29 37
KN 29 43
NU 29 40
UZ 31 39
ZH 32 31
Table 12: Fraction of events ending in the money.
group calRef 5 10 15 20
percentile < 25 HK 0.59 0.52 0.48 0.45
percentile < 25 KN 0.59 0.62 0.62 0.45
percentile < 25 NU 0.69 0.55 0.48 0.52
percentile < 25 UZ 0.58 0.52 0.55 0.48
percentile < 25 ZH 0.62 0.47 0.38 0.53
percentile > 75 HK 0.57 0.68 0.49 0.38
percentile > 75 KN 0.56 0.67 0.53 0.47
percentile > 75 NU 0.48 0.50 0.50 0.50
percentile > 75 UZ 0.56 0.64 0.59 0.38
percentile > 75 ZH 0.52 0.58 0.52 0.45

Rough Rice

Table 13: USd change from event date.
group calRef 5 10 15 20
percentile < 25 FH 0.0000 -0.0025 0.0000 -0.0050
percentile < 25 HK 0.0050 -0.0025 0.0000 -0.0025
percentile < 25 KN 0.0000 0.0000 0.0000 0.0000
percentile < 25 NU 0.0000 0.0100 -0.0250 -0.0250
percentile < 25 UX 0.0000 -0.0050 -0.0050 -0.0050
percentile < 25 XF 0.0000 0.0000 -0.0075 -0.0050
percentile > 75 FH 0.0000 0.0000 -0.0175 -0.0350
percentile > 75 HK 0.0000 -0.0050 -0.0100 -0.0100
percentile > 75 KN 0.0000 0.0000 0.0000 -0.0150
percentile > 75 NU -0.0275 -0.0250 -0.0100 -0.0725
percentile > 75 UX 0.0000 0.0000 -0.0100 -0.0250
percentile > 75 XF 0.0000 -0.0150 -0.0200 -0.0475
Table 14: Total number of events by calRef.
calRef percentile < 25 percentile > 75
FH 53 39
HK 47 40
KN 46 45
NU 11 46
UX 49 41
XF 55 48
Table 15: Fraction of events ending in the money.
group calRef 5 10 15 20
percentile < 25 FH 0.45 0.42 0.34 0.32
percentile < 25 HK 0.53 0.43 0.34 0.23
percentile < 25 KN 0.41 0.43 0.24 0.30
percentile < 25 NU 0.45 0.55 0.36 0.36
percentile < 25 UX 0.45 0.35 0.37 0.20
percentile < 25 XF 0.31 0.33 0.31 0.29
percentile > 75 FH 0.33 0.46 0.49 0.49
percentile > 75 HK 0.38 0.52 0.57 0.57
percentile > 75 KN 0.42 0.47 0.42 0.49
percentile > 75 NU 0.57 0.54 0.50 0.50
percentile > 75 UX 0.44 0.46 0.51 0.46
percentile > 75 XF 0.42 0.62 0.65 0.56

Soybeans

Table 16: USd change from event date.
group calRef 5 10 15 20
percentile < 25 FH 0.125 0.000 -0.250 -0.500
percentile < 25 HK 0.250 0.000 -0.250 0.000
percentile < 25 KN 0.250 0.250 -0.375 -0.250
percentile < 25 NQ 0.250 0.250 0.000 0.250
percentile < 25 NX 0.000 -0.250 -1.500 -1.250
percentile < 25 QU 0.750 0.625 -0.125 -0.500
percentile < 25 UX 0.500 0.250 0.000 0.000
percentile < 25 XF 0.000 -0.250 -0.500 -0.500
percentile < 25 XN 0.250 0.000 -0.750 -1.750
percentile > 75 FH -0.250 0.000 -1.375 -2.250
percentile > 75 HK -0.750 -1.750 -2.500 -3.750
percentile > 75 KN 0.000 -0.250 -0.750 -1.500
percentile > 75 NQ 0.625 1.625 2.500 1.625
percentile > 75 NX 2.500 6.000 2.500 -0.500
percentile > 75 QU 1.250 0.750 -0.250 1.250
percentile > 75 UX -1.000 -1.500 -0.375 -2.500
percentile > 75 XF 0.000 -0.250 -0.750 0.000
percentile > 75 XN -1.250 -1.500 -2.625 -6.500
Table 17: Total number of events by calRef.
calRef percentile < 25 percentile > 75
FH 60 48
HK 62 45
KN 69 39
NQ 47 30
NX 46 37
QU 32 35
UX 55 40
XF 52 45
XN 53 43
Table 18: Fraction of events ending in the money.
group calRef 5 10 15 20
percentile < 25 FH 0.50 0.45 0.37 0.32
percentile < 25 HK 0.50 0.47 0.32 0.32
percentile < 25 KN 0.54 0.52 0.28 0.26
percentile < 25 NQ 0.53 0.53 0.43 0.38
percentile < 25 NX 0.48 0.43 0.33 0.28
percentile < 25 QU 0.56 0.53 0.41 0.31
percentile < 25 UX 0.60 0.53 0.42 0.36
percentile < 25 XF 0.44 0.37 0.31 0.25
percentile < 25 XN 0.51 0.47 0.36 0.28
percentile > 75 FH 0.50 0.46 0.52 0.54
percentile > 75 HK 0.60 0.69 0.64 0.62
percentile > 75 KN 0.49 0.54 0.54 0.59
percentile > 75 NQ 0.40 0.33 0.30 0.30
percentile > 75 NX 0.41 0.46 0.49 0.43
percentile > 75 QU 0.34 0.43 0.51 0.34
percentile > 75 UX 0.60 0.52 0.48 0.52
percentile > 75 XF 0.44 0.53 0.49 0.38
percentile > 75 XN 0.58 0.53 0.56 0.44

Soybean Meal

Table 19: USd change from event date.
group calRef 5 10 15 20
percentile < 25 FH 0.00 -0.20 0.00 0.00
percentile < 25 HK 0.10 -0.20 -0.20 0.05
percentile < 25 KN 0.00 0.10 0.00 -0.20
percentile < 25 NK 0.15 0.35 0.05 -1.65
percentile < 25 NQ 0.10 0.10 0.05 0.00
percentile < 25 NZ -0.10 0.20 -0.30 -0.30
percentile < 25 QU 0.10 0.00 0.05 -0.10
percentile < 25 UK -0.40 -0.10 0.00 -1.40
percentile < 25 UV -0.05 -0.10 0.05 0.00
percentile < 25 UZ -0.20 -0.10 -0.10 -0.15
percentile < 25 VZ 0.05 -0.10 0.10 -0.10
percentile < 25 ZF 0.00 0.00 0.00 0.10
percentile < 25 ZN 0.00 0.40 0.30 -0.10
percentile > 75 FH -0.60 0.00 0.10 -0.30
percentile > 75 HK -0.55 -0.35 -0.50 -0.75
percentile > 75 KN 0.10 0.40 -0.25 -0.45
percentile > 75 NK 0.40 1.70 1.25 3.35
percentile > 75 NQ 0.05 0.05 -0.05 -0.10
percentile > 75 NZ -0.20 0.00 -0.35 0.50
percentile > 75 QU 0.40 -0.25 0.20 -0.10
percentile > 75 UK 0.00 0.30 1.60 2.65
percentile > 75 UV -0.35 -0.20 -0.45 -0.10
percentile > 75 UZ -0.90 -0.40 -0.25 -0.05
percentile > 75 VZ -0.20 -0.45 0.10 -0.30
percentile > 75 ZF -0.15 -0.45 0.20 0.00
percentile > 75 ZN -2.20 -1.00 0.35 -2.15
Table 20: Total number of events by calRef.
calRef percentile < 25 percentile > 75
FH 60 35
HK 69 40
KN 59 34
NK 24 22
NQ 40 44
NZ 41 47
QU 38 46
UK 11 14
UV 14 41
UZ 15 39
VZ 52 36
ZF 64 40
ZN 37 37
Table 21: Fraction of events ending in the money.
group calRef 5 10 15 20
percentile < 25 FH 0.47 0.42 0.37 0.37
percentile < 25 HK 0.51 0.43 0.38 0.36
percentile < 25 KN 0.46 0.49 0.41 0.29
percentile < 25 NK 0.50 0.54 0.50 0.33
percentile < 25 NQ 0.52 0.52 0.50 0.35
percentile < 25 NZ 0.49 0.51 0.46 0.29
percentile < 25 QU 0.55 0.45 0.50 0.37
percentile < 25 UK 0.45 0.45 0.36 0.18
percentile < 25 UV 0.43 0.36 0.50 0.36
percentile < 25 UZ 0.40 0.47 0.47 0.40
percentile < 25 VZ 0.50 0.44 0.52 0.35
percentile < 25 ZF 0.48 0.47 0.38 0.38
percentile < 25 ZN 0.49 0.54 0.43 0.32
percentile > 75 FH 0.63 0.49 0.43 0.46
percentile > 75 HK 0.60 0.60 0.50 0.48
percentile > 75 KN 0.44 0.44 0.50 0.47
percentile > 75 NK 0.41 0.45 0.45 0.32
percentile > 75 NQ 0.45 0.43 0.50 0.41
percentile > 75 NZ 0.51 0.45 0.47 0.36
percentile > 75 QU 0.43 0.52 0.48 0.43
percentile > 75 UK 0.50 0.50 0.43 0.36
percentile > 75 UV 0.56 0.49 0.54 0.39
percentile > 75 UZ 0.54 0.54 0.51 0.38
percentile > 75 VZ 0.58 0.61 0.44 0.47
percentile > 75 ZF 0.55 0.60 0.42 0.40
percentile > 75 ZN 0.65 0.57 0.43 0.51

Soybean Oil

Table 22: USd change from event date.
group calRef 5 10 15 20
percentile < 25 FH 0.010 0.015 -0.010 0.000
percentile < 25 HK 0.010 0.000 -0.010 0.010
percentile < 25 HN 0.010 0.010 -0.025 -0.005
percentile < 25 KN 0.010 -0.010 -0.005 0.000
percentile < 25 NQ 0.000 0.010 0.000 0.010
percentile < 25 NZ 0.010 0.020 -0.030 0.020
percentile < 25 QU 0.000 0.010 0.000 0.010
percentile < 25 UH -0.015 -0.015 -0.045 -0.010
percentile < 25 UV 0.000 -0.010 -0.010 0.005
percentile < 25 UZ -0.020 -0.020 -0.060 0.000
percentile < 25 VZ 0.000 0.000 0.000 0.010
percentile < 25 ZF 0.010 0.010 0.010 0.010
percentile < 25 ZN 0.055 0.030 0.020 0.010
percentile > 75 FH -0.010 -0.020 -0.020 -0.030
percentile > 75 HK -0.015 -0.040 -0.045 -0.070
percentile > 75 HN -0.020 -0.030 -0.100 -0.100
percentile > 75 KN -0.020 -0.010 -0.025 -0.050
percentile > 75 NQ -0.030 -0.050 -0.060 -0.060
percentile > 75 NZ 0.000 0.000 -0.055 -0.090
percentile > 75 QU 0.000 -0.055 -0.055 -0.060
percentile > 75 UH 0.000 -0.010 -0.010 -0.070
percentile > 75 UV 0.010 -0.030 -0.040 -0.060
percentile > 75 UZ -0.005 -0.040 -0.065 -0.055
percentile > 75 VZ -0.020 -0.030 -0.030 -0.060
percentile > 75 ZF 0.000 -0.015 -0.030 -0.070
percentile > 75 ZN -0.035 -0.040 -0.085 -0.060
Table 23: Total number of events by calRef.
calRef percentile < 25 percentile > 75
FH 49 36
HK 51 36
HN 49 33
KN 57 30
NQ 58 25
NZ 48 28
QU 44 24
UH 18 15
UV 27 23
UZ 21 22
VZ 68 40
ZF 54 30
ZN 44 37
Table 24: Fraction of events ending in the money.
group calRef 5 10 15 20
percentile < 25 FH 0.59 0.57 0.35 0.29
percentile < 25 HK 0.61 0.47 0.25 0.33
percentile < 25 HN 0.55 0.53 0.27 0.29
percentile < 25 KN 0.51 0.40 0.42 0.33
percentile < 25 NQ 0.47 0.55 0.41 0.40
percentile < 25 NZ 0.52 0.52 0.42 0.44
percentile < 25 QU 0.45 0.55 0.43 0.39
percentile < 25 UH 0.44 0.33 0.39 0.33
percentile < 25 UV 0.33 0.30 0.33 0.37
percentile < 25 UZ 0.24 0.29 0.48 0.29
percentile < 25 VZ 0.49 0.46 0.41 0.43
percentile < 25 ZF 0.56 0.56 0.44 0.41
percentile < 25 ZN 0.68 0.64 0.43 0.43
percentile > 75 FH 0.53 0.67 0.56 0.56
percentile > 75 HK 0.67 0.67 0.69 0.75
percentile > 75 HN 0.64 0.64 0.67 0.70
percentile > 75 KN 0.53 0.53 0.63 0.80
percentile > 75 NQ 0.76 0.72 0.84 0.68
percentile > 75 NZ 0.46 0.50 0.54 0.57
percentile > 75 QU 0.38 0.58 0.75 0.50
percentile > 75 UH 0.47 0.53 0.53 0.47
percentile > 75 UV 0.43 0.61 0.61 0.43
percentile > 75 UZ 0.50 0.64 0.55 0.41
percentile > 75 VZ 0.62 0.60 0.62 0.55
percentile > 75 ZF 0.47 0.53 0.60 0.43
percentile > 75 ZN 0.54 0.57 0.59 0.46

Remarks

• Corn, the wheats and soybeans performs the best
• Some spreads are better than others
• Definite opportunity to systematise this approach
• Extending the results to relatives will also be interesting
Mauritz van den Worm
Portfolio Manager and Quantitative Researcher

My research interests include the use of artificial intelligence in managing commodity portfolios