User`s Manual 3.x
Contents
1. Price Prediction Reliability 20 275 Estimate of Nutrient Unit Costs Nutrient name Estimate RUP 0 03119 eNDF 0 03958 neNDF 0 03638 ic RDP 0 01506 U Methionine 10 36242 ne U Lysine 1 276768 za NEI 0 ne Section 2 Program Overview 2 18 4 2 2000 2 20 18 PM A blank means that the nutrient unit cost is likely equal to zero means that the nutrient unit cost may be close to zero means that the nutrient unit cost is unlikely to be equal to zero means that the nutrient unit cost is most likely not equal to zero Name Bakery waste dehy 92 000 orn grain ground Hominy feed Molasses sugarcane Brewers grains dehy Canola meal solv extd Corn gluten feed Corn gluten meal 60 Cottonseed meal 41 Cottonseed Whole Distillers dried grains Feather meal Fish meal Menhaden Meat and bone meal Poultry byproduct meal Soybean meal solv 44 Soybean meal solv 48 114 13 110 00 110 00 86 00 141 00 149 00 96 00 306 00 175 00 174 00 115 00 225 00 415 00 195 00 gt Q pee Ss v silsilesi ssis slisse Si Calibration set 107 95 113 52 99 18 13 92 Soybean Hulls 82 00 66 83 Tallow 350 00 344 07 Wheat Middlings 66 00 89 85 Blood meal 430 00 431 86 122 54 144 46 125 16 302 52 148 17 176 17 132 44 222 80 403 44 209 42 vU 0 2 O ae D a RSN e RR SlSSlal Si Ni Sins S 4 Se 229 999 262 235 253 392 271 079 186 000 168 769 157 582. To create a new price list click the Database button followed by the Create new price list option Create new price list Ppce De Normand St Pierre The Oleo State Univerzity By price kstz Br fects Price kat 1s hoa SONOS amp Price lists 12000 11000 Midwest V39 3 Scutheost 2 00 5 Wert Coast 4 000 34 00 no 80 200 20 seem To edit a price list select the appropriate list from the list in the left window Then select the price box for a feedstuff in the Price column on the right and make the necessary corrections Select the appropriate Edit current prices price list of feedstuffs Price Di Numuri S4 Picee The Ohiv Stute Urnivuiriy By 2 ce katy Bafian Pw kae 11 Dotoboz a A Prica lixta east Const iea fel Midusest 7 99 3 Sovthcost 2 00 5 liect Coocet 4 Hiv tsr arre H am ry hbi I clara ANS Save your changes by clicking the green button Section 2 Program Overview 3 Create copy a problem These functions are accessed by clicking the Solver button in the main menu Solver To create a new problem click on the Database button and select Cre ate new problem To copy a problem select the problem to be copied from the choices listed in the left window Then click on the Database button and select Copy problem Select problem to be copied Suha Create new problem or Copy problem
3. cccccccccscesescccccceecessesseeeeeceseeeaaeeeens Tutorial Problem 1 ou toeni ni e a Tutorial Problem 2 ccccccccccsssseesccccceeeccsessseecccceseeeaaeneees Tutorial Problem 3 ccccccccessssseccccceeccesessescccceseeeauaenees Tutorial Problem 4 cccccccccssssecccccceeecesseesesececeseeeaaeesees Tutorial Problem Sra nsreci enaa a a Answers to Tutorial Questions cccccccseeeeeecceceeeeeeaeeeees Section 5 Advanced Topics The Short Solver Report jsjsssscsncpviscergiassetenes ices bessdeaseoeconas s The Full Solver Report sssssssssseeeseessssssserereessssssseeereeesss Details of Methodology Used 000 0000 eee eeeeesseceeeeeeeeeeeeenneees Estimation Problems 25 5 sce sesye Better cihs he eal eet etal Section 6 Help HED Rora te NCR eee A eet rene Seen 6 1 Section 1 Getting Started Getting Started Installation 1 Insert the SESAME CD into your CD ROM drive 2 If your operating system supports AutoPlay installation will proceed automatically If not use Add Remove Programs from the Start gt Set tings gt Control Panel menu 3 Click Install and follow the prompts to install SESAME 4 Reboot your computer after a successful installation 5 Start SESAME using the Start menu and clicking SESAME the green clover icon The following window will appear depending on the type of software license that you have Usor nome Uzor nome Enorme Poee
4. Copyright 2000 All educational programs and activities conducted by The Ohio State University Extension are available to all potential clientele on a nondiscrimina tory basis without regard to race color creed religion sexual orientation national origin gender age disability or Vietnam era veteran status Index Section 1 Getting Started Installation enine ipp p aa Si a canis Changing Your Password sseseeeeeseeseesreeesssrressserersssrreesssree System Requirements sssseseeeseeeesssssserereessssssserrreeesssssee OUR PASS WOLG erdee o A E E E E Section 2 Program Overview What It Is amp What It Does sssseesseseesessserrssrressssreressererssse What It Isn t amp What It Doesn t Do oe eeeeeeeeeeeeeee Menu eS EN oss so soles e tet tee eee oe Nutrient Composition The Feedstuffs Menu Nutrient Definition The Configuration Menu Prices of Feedstuffs The Price Lists Menu Finding Nutrient Costs and Break even Prices The Solver Meni n us cane ele Setting Up a Problems s ds qssccsecescsssaspicsadiiescesecsdandiotsecaness Quick Overview of the Principles Behind SESAME Assumptions Used 5 10 35 sassncscoadorsdaspeseadaeqdacdanesecsanndacsdaeness A Simple Example e ie ere tec eee o a i a nee Section 3 Helpful Hints Help ital Hint i243 a ssatsscasieinan ana aia Section 4 Program Tutorial About the Tutorial
5. FAX 614 292 1515 email st pierre 8 osu edu 3 Describe the problem in as much detail as possible We will try to answer your problem as quickly as possible
6. Break even prices are calculated based on the nutritional composition and the estimated costs of nutrients Because they are statistically derived break even prices are reported with a confidence interval Thus feedstuffs belong to one of three categories under priced neutrally priced and over priced SESAME provides answers to a variety of marketing production and nutri tional questions For example A Company just invented a new bypass lysine product What is its break even price Is it the same across all major dairy producing regions A commodity broker sells wet brewers grains What is its break even price A feed manufacturer has 12 ingredient bins at one of its manufacturing feed plant What ingredients should be inventoried u A dairy producer is being offered three types of almond hulls each with a different nutrient profile Which one should he buy In early October a dairy producer can lock in a price for March deliv ered whole cottonseed Should he buy 2 1 Section 2 Program Overview 2 2 What It Isn t amp What It Doesn t Do SESAME does not use indicator feeds to calculate unit costs of nutrients For example the cost of energy and protein are not determined from the price of corn and soybean meal SESAME is not a ration balancer It does not provide amounts to be fed SESAME does not select the ingredients to make a nutritionally optimal diet SESAME does not value feedstuffs
7. Di Nuneund 54 Phare The Cihiv Stute U viy 7 Splue problem il La Fag eye di n leap E Dennehey ey rProblems Examples Dairy 1 dew East Canet Ferds QExamples Others 2 MUzer s Problems 25 Fass wren amp Pride tists E P Poe Lenaers Lette nen ezpa Problama a m Print Help A End aa ere ete ae Loy Arde y clams ha 7 he ae asf m Ley lianas AP one Cun Peay in Padune DMa ir Rhei Hl xe Select a problem type For your convenience certain nutrients are already grouped for typical problems In SESAME V1 1 types are as follow Type 10 Type 11 Type 12 Type 13 Type 14 Type 18 Dairy Dairy Dairy Dairy Dairy Dairy High Performance Medium Performance RDP RUP and NE CRONE Heifers High Rare Inert NE Type 20 Type 30 Type 40 Type 50 Type 60 Type 70 Beef Cattle Poultry Swine Horse Rabbit Fish 2 11 Section 2 Program Overview 2 12 The dairy types automatically group the following nutrients Dairy Types Nutrients 10 11 12 13 14 18 Crude Protein x Non effective NDF Effective NDF Net Energy Gain Net Energy Lactation Rumen Degradable Protein Rumen Undegradable Protein RUP Methionine RUP Lysine Rare and Rumen Intert NE K MK KM OK aK MK KM KM xx KM KM xX x xX 4 Select the active Price list This is done in the Solver window The active pric
8. The normal sequence for setting up a problem is as follow 1 Create edit feedstuffs not already in a library If all feedstuffs that you intend to use are already in one of the libraries default or user s then you can bypass this step However if one or more feedstuffs are not already in a library or if you need to modify the nutrient composi tion of at least one feedstuff then you must execute this step Feedstuffs and their libraries are accessed via the Feedstuffs button on the main menu Feedstuffs 2 8 Section 2 Program Overview You cannot edit a feedstuff belonging to the Commercial CPM or NRC libraries If you want to modify the composition of a default feedstuffs you must first copy it to a user s library To copy a feedstuff select the feedstuffs by left clicking on its name Then click the Database button and select the Copy Feedstuffs option Copy feedstuffs option Fendstulls database amp Composition Price lidts Nit a e Be g SF cedstulls Commer cicl Library 1014 CPM Library 12 SNAC 2989 Library 1000 BAI cilia Howo gen corly v The Ohio State Unewernty 13H60 10 1018 1 00063 1B 4 58 40423 143 208 100 8 10 er 40r 10 7 SAlfolia Hoy sen late veg 4 ifolia Hoc aorty bloom 6 BAI ollo Hod cen midbioom Allella Hoye sun full bloom 10 Alfolia Sitoge o corby veg 58 Bhiltclia Siloge b
9. VITAMINS 1005 Laad m Print i WES Fide ate BM FR Plane styl A AU Se SDL Double click in the box to add remove a nutri ent from the problem A blue background indi cates that the nutrient is a default selection for the type of problem 7 Solve the problem This is done by clicking on the Solve problem button in the Solver menu Solve Problem Suha Di Hunmund 5t Pie The Ohiv State Uniwaviy lols Z Solve problem amp Price lista J Optobpze f Laad m Print Help A Ena Iakes Feah iten coza BEES E m Khaz waf AD Tag Fi baler e JEH ir gt Pr a Jwr a ba T a A Ale 3H 3 nm T4 Corn groin cracked 1 OF A Ta babes ban JH nm li Corn groin Hoked 10 Sy 4 Fa te ae ie am 1a torn groin ground nae Fuvdy Hage Peslurmurnx C uwr Es ea om Ta Carn groin high moist ans ee z i BCern Silaga law gars as onst fesada kuj aCern Silage well eare Wbs nbh ER ty Ly iaw uw T1 Catton hulls 28 Renionmanes es am E Cottonseed meol 41 Pag L ua ied DCottan coad Whole 15 m ih acct ain mo iS aOlstillars dried grains Gs sa r hn 24 Feather meal 156 fe cen eS BFescuc Hoy lote veg Se TENE ala BFizh meol Anchovy 1 rn ghae Uin Fish meol Menhaden spoken gths Seimin rrends Wominy feed 116 A Maai and bang meal 1 s 5922 Maacn tre od 15 Tim ba Aw TH aMalasers maererag Shoei Hic r at ME Coli tcc yer es z N k FLA sarhil nays ZEIDEN ko re D w
10. attributes that are not nutrient related For example it will not provide the proper valuation of a mold inhibitor Menu Structure The main window of SESAME contains the main menu for operation of the program Section 2 Program Overview 2 3 The following is a list of the main menu items and their function Feedstuffs is used for viewing and editing the nutritional compo sition of feedstuffs Solver is used to select feedstuffs nutrients and prices and to ob tain estimates of nutrient costs and break even prices of feedstuffs Price lists allows the user to set different price series to reflect price differences do to things such as regional differences Backup is used to produce a backup copy of the database or to recover a previously saved copy of the database Configuration provides access to the underlying structure of the nutrient definition table Help opens a context sensitive help system About prints a disclaimer and provides information on how to contact the program authors End terminates the program z Fe Feedstuffs T Solver Solver Price lists Backup Configuration Help About Section 2 Program Overview 2 4 Nutrient Composition The Feedstuffs Menu z a Feedstuffs The Feedstuffs form is used for viewing and editing the nutritional composition of feedstuffs It is possible to Insert Edit Copy and Delete all the
11. by the estimated break even price Table 4 Appraisal Set 99 302 5170 E The last table reports the Calculation log It contains useful information whenever one of the two Wizzards was used The table reports the action taken by the Wizzard followed by the reason for such action Table 5 Calculation Log Feedstuffs Corn gluten feed removed The Full Solver Report The full report prints all the statistics from the solution The first table is labeled Main results Table 1 Main Results oer mosa serai7 6387 Section 5 Advanced Topics 5 4 The Main Table reports e The R square which represents the proportion of the total variance of prices of feedstuffs explained by the selected set of nutrients and their cost estimates e The Adjusted R Square which is the R square adjusted penalized for the number of nutrients in the model e The Root Mean Square Error which is an estimate of o the error term in the model The RMSE should be less than 30 for results to be reliable e The Mean of the Dependent Variable labeled Dep Mean This is the mean price of all the feedstuffs selected in the calibration set e The C V which is the coefficient of variation of the error The second table is labeled Analysis of Variance SESAME uses regression techniques to estimate the cost of nutrients Thus the results can be summa rized in an ANOVA format Table 2 Analysis of Variance 668261 344 95465 9
12. feedstuffs within a library as well as to Copy and Key to Feedstuffs Delete whole libraries By default SESAME contains three pro menu task bar tected feedstuffs libraries and two unprotected user libraries Pro tected databases are Commercial library CPM Library and NRC First row 1989 Library Feedstuffs in these protected libraries can be used in setting up problems but they cannot be edited The user can cus Previous row tomize the nutritional composition of a feedstuff by first copying it h Nearer to a user library The copied feed can be edited once it resides in a J Last row user library amp Mark position return Composition Price lists Database Find Print Help End peas Delete row Edit row amp Composition amp Price lists 2 Database Find amp Print Help End E EE Post change SP ESAS I 0 Commercial Libeary 0 CPM L rary fio 0 NRC 1989 Library 0 User Libranes Cancel change Refresh data Commercial Library 1014 CPM Library 1012 ONRC 1989 Library 1000 User Libraries 1017 aa ah o View the available feedstuffs in each library in this window Select the appropriate feedstuffs library Menu items in Feedstuffs are Composition opens a window used for viewing and editing the nutrient composition of a feedstuff The name of the selected feedstuff is displayed in the caption appearing in the upper left corner of the win
13. lt OK gt button to return to the main menu Section 4 Program Tutorial 4 2 Once you have completed the tutorials you can recover your original work ing files using the backup file that you created in step 1 To recover follow the steps described in step 2 above locating this time the backup file gener ated in step 1 Tutorial Problem 1 An old friend of yours in the Southeast wants to compare feedstuffs solely on the basis of their energy NE and crude protein content As far as he knows feedstuffs traded in the area have pretty much the same nutritional composition as those reported by NRC 1989 Therefore you can use the NRC 1989 library to select the feedstuffs from Market prices in the Southeast are as follows ID Group Feedstuff ton 22 1 Citrus pulp dehydrated 80 00 28 1 Cotton hulls 100 00 100 4 Bakery waste dehydrated 90 00 102 4 Barley grain 137 50 104 4 Beet pulp dehydrated 147 00 112 4 Corn grain ground 93 50 116 4 Hominy feed 110 00 118 4 Molasses sugarcane 63 50 128 4 Soybean hulls 75 00 130 4 Tallow 200 00 134 4 Wheat middlings 70 00 132 4 Wheat grain ground 98 20 136 5 Blood meal 410 00 138 5 Brewers grains dehydrated 107 00 146 5 Corn gluten feed 75 00 148 5 Corn gluten meal 60 250 00 150 5 Cottonseed meal 41 125 00 152 5 Cottonseed whole 132 00 154 5 Distillers dried grains 117 00 156 5 Feather meal 183 00 160 5 Fish meal Menhaden 370 00 162 5 Meat and bone meal 190 00 164 5
14. on the variance of the estimate of the break even price for feedstuffs i is clearly seen In SESAME feedstuffs with leverage values h gt 2 p 1 n where p number of parameters and n number of feedstuffs are deemed leveraged and their inclusion in the data set should be questioned Feedstuffs with leveraged values h gt 3 p 1 n are automatically removed from the dataset by the wizzard Collinearity Collinearities are approximate linear dependencies among the col umns of predictor variable values Collinearities can be due to model con straints population characteristics that restrict variate values sampling defi ciencies or overdefined models Collinearities produce large coefficient estimates and inflate the variance of parameter estimates In short collinearity induces ill conditioning of the X X matrix that is the determinant of X X is close to zero This condition can be thought of in two ways either the model is overspecified we are trying to estimate the cost of more nutrients than are needed to express the break even prices of the feedstuffs selected or the data are not adequate to estimate all of the selected nutrients Collinearities can be measured with variance inflation factors VIF which are the diagonal elements of X X when X X is scaled to correla tion form In general estimability is acceptable when all VIF lt 10 by defi nition VIF 1 in the absence of collinearity In instance
15. the authors For security reasons we rec ommend that you change your User name and Password immediately User License User licensees can change their own User name and password To do a User Password change 1 Click on the Password change button at the bottom of the Password check window The Password change window will appear Enter current User name and password Bass won Ch ag Ges User nome Uzor Old posaword HAHA New password New poeeword repooted vok Enter new pass word twice for verification X Concel Section 1 Getting Started 1 3 2 Enter your current User name and password followed by the new pass word entered twice for verification 3 Click OK to save the change Administrator License Administrator licensees can change their own User name and password and have the ability to add unlimited Users and Administrators The Password change button on the Password check window allows Admin istrators or Users to change their own passwords only refer to User Li cense The Administration button allows Administrators to add edit or delete an unlimited number of other Administrators and Users and their pass words through the Password desktop window To use the Password desk top 1 Click on the Administration button in the Password check window Uzor nome ehorite Poeeword EEE YOR Conceal Click on the Administration Button Adminietrotio
16. what kind of hay they should buy They are offered two options 1 20 alfalfa hay NRC feed 6 and 2 15 alfalfa hay NRC feed 10 Your friend wants to know how much these two hays are worth Step 1 Create new problem Tutorial Problem 3 Hint copy from Tutorial Problem 2 Section 4 Program Tutorial Step 2 Select a problem type Hint forages are used as sources of effective fiber Problem Type 12 would make sense Step 3 Move NRC feedstuffs 6 and 10 from the Feedstuffs window to the Appraisal window Hint drag and drop Step 4 Solve Hint use Wizzard Additional Exploration in Tutorial Problem 3 Q1 What are the break even prices of the two hays Q2 Are there any feedstuffs that were considered over priced in Prob lem 2 that are no longer over priced 4 11 Tutorial Problem 4 Section 4 Program Tutorial Your sister in law calls you from the Midwest with the same question as the one addressed in Problem 3 except that the availability and prices of feedstuffs in the Midwest are very different Use the following Midwest prices to an swer your sister in law s question regarding the break even prices of the two hays ID 220 28 100 102 104 112 116 118 128 130 134 136 138 142 146 148 150 152 154 156 160 162 166 168 170 Group NaOAnranrnrnrnnranrnnrarnrnrnrsb HHH HHH HR Feedstuff AlfPellet18Cp46Ndfl7LNDF Cotton hulls Bakery waste dehydrated Barl
17. 0 00 144 00 177 00 108 00 185 00 410 00 185 00 279 00 173 00 182 00 4 14 Section 4 Program Tutorial Step 1 Create two new feedstuffs Click on the Feedstuffs button in the main menu Once the Feedstuffs database is open click on NRC 1989 Library in the left window Lo cate and click on record 12 Almond hulls in the right window Click the Database button on the menu then select Copy feedstuffs Type Al mond Hulls Hi CP in the feedstuffs name field and click the OK button Repeat this process to copy 134 Wheat Middlings to a new feed named Wheat Midds Lo CP When done double click on User Libraries in the left window By default SESAME copies feedstuffs to User Libraries To keep your data organization clear you should move the two newly created feedstuffs to Library 1 Click on Almond Hulls Hi CP in the right window then click on Composition in the menu The Chemical composition window opens Click on Basic chemical analysis in the left window In the value field for nutrient 2 crude protein type 5 0 fol lowed by the lt down arrow gt key on your keyboard this automatically post your change to the database alternatively you can type 5 0 fol lowed by the lt ENTER gt key but then you will need to post your changes by clicking on the green check mark on the task bar Click on the End button in the menu to return to the Feedstuffs database Repeat this process for editing the crude protein of Wheat Midds Lo CP Ste
18. 06 834 820 0 0000 E ea O j a0 This table reports e The DF which stands for the degrees of freedom for each source e The Sum of squares for each source e The Mean square for the model and the error term e The F statistics used for testing the significance of the model e A Probability estimate of the significance of the model Section 5 Advanced Topics 5 5 The third table is labeled Inference of coefficients Table 3 Inference of Coefficients 0 064080 0 015360 4172 0 0016 8 0 001060 0 014938 0 071 0 3447 9 6136 0 024937 0 008757 2848 0 0159 27166 5 7154 0 037804 oossst 2 364 0 0375 7432 10 787557 0 666875 18 176 4 086 raaa oreas eors wooo smsa 0 056681 0 005202 10 896 0 0000 9 7487 For each nutrient it reports e An estimate which is simply the estimated cost per unit of nutrient i e per Ib or kg For U S users the value is in dollars per unit e A standard error of the estimated cost of the nutrient e A t statistic to test whether the estimated cost of the nutrient is signifi cantly different from zero e A probability that the estimated cost is significantly different from zero e The Variance inflation factor VIF associated with each nutrient VIF measures the degree of collinearity due to the association between the estimates of nutrient costs If the nutrients were independent orthogo nal VIF would all be equal to one Large VIF gt 15 20 indicate that the mo
19. 29 412 is less than 30 which is accept able Thus we are relatively confident that nutrients costs and break even prices of feedstuffs are estimated accurately Look at a graph of the results by clicking on the Graph command on the menu Feedstuffs above the horizontal blue line are relatively under priced Those under the blue line are relatively over priced Blue line w b a e Rk Hw HH WH BB He He SE i Atasi coat Ton 4 7 Section 4 Program Tutorial 4 8 Additional Exploration in Tutorial Problem 1 Use the Help system to improve your understanding of the output Help index Short solver report You should be able to answer the following ques tions Q1 Do you have reliable estimates Q2 What is the estimated cost of 1 pound of crude protein Q3 What is the estimated cost of 1 Mcal of NE Q4 List 3 under priced feedstuffs Q5 List 3 over priced feedstuffs Tutorial Problem 2 Your Southeast friend calls you back and points out that most people now use rumen degradable RDP and rumen undegradable protein RUP to char acterize feedstuffs fed to ruminants He wants to know if the relative rank ing of feedstuffs remains the same using these two nutrients plus NE You have gained familiarity with SESAME You know how to access and search the Help system Therefore we are no longer providing you with the detailed steps for each task within SESAME in this tutorial Step 1 Create t
20. 4 179 954 167 464 195 007 Name Alfalfa Hay a sun early ve Beet pulp dehydrated sal set pral 0 000 Actual T Predicted T 128 051 99 859 Action Calculation log Feedstuffs Beet pulp dehydrated removed Reason Outlier SESAME Regression results Dr Normand St Pierre page 1 Section 2 Program Overview The numbers in the column labeled Estimate in the Estimate of Nutrient Unit Costs table are the calculated values per unit per pound for all nutri ents except NE which is per Mcal of nutrient Thus a pound of RUP had an estimated worth of 0 031 This estimate would appear low except that it represents the value of one pound of RUP which is free of both methionine and lysine because the value of those nutrients were already factored in the evaluation In fact markets were valuing U Methionine at 10 36 Ib and U Lysine at 1 27 lb Effective NDF was valued at close to 0 04 Ib whereas neNDF was valued at 0 036 indicating that the markets were discounting feedstuffs for their neNDF content Likewise markets were giving little value to RDP 0 015 Ib but were valuing NE 0 065 Mcal The break even prices of feedstuffs Predicted values are reported in the two tables labeled Calibration Set and Appraisal Set Additionally the Calibration set contains the lower and upper 75 confidence limits of break even prices Using the
21. CP and NE from the composition and market prices of multiple feedstuffs As with the Petersen method we will use corn and soybean meal but we will also add market information from hominy and canola meal In the real world there would be many more feedstuffs avail able but two additional feedstuffs are sufficient to illustrate the procedure In SESAME we first set the following equations Corn 90 176 0 CP 1566 NE E Soybean meal 145 991 8 CP 1638 NE E Hominy 94 207 0 CP 1698 NE E Canola meal 117 738 9 CP 1292 NE E where E are the errors associated with the prices of feedstuffs i e the dif ference between the calculated value of a feedstuffs and its price SESAME finds estimates of CP and NE that minimize the sum of E squared Addi tionally because SESAME verifies that certain statistical conditions on the E are met these estimates of nutrient costs have maximum likelihood prop erties i e they are best at explaining differences in feedstuff prices The method used by SESAME is easily extended to a large set of feedstuffs and multiple nutrients 2 16 Section 2 Program Overview 2 17 Assumptions Used 1 Over time the classical economic mechanisms of supply and demand are the primary determinants of feed prices 2 Buyers and sellers of commodity act rationally A buyer would not keep buying an overpriced feed commodity and a seller would not keep sell ing commoditie
22. H iFa orn Cu Daip ip Patuvar s JEx Midwest Coast Feeds High lhe ae a he er nas ir Rhee BE Ex llest Coosi Feeds High Per aE fest Seat EGS Cie ele ort m DEx Wsat Cooet Vory High Pert amp Examplies Others 2 Problems MUzer s Problems 25 tabs Feedstuffs tabs and Chem ition composition s 7 tabs Calibration zei _ aenep e yx Jo sarhil naya Iwlti jina amaaa AA Calibration Set Appraisal Set 2 7 t Solver Solver Key to Solver menu task bar First row Previous row Next row Last row Mark position return Insert row Delete row Edit row Post change Cancel change Refresh data Section 2 Program Overview Menu items in Solver are Solve problem solves the problem with the maximum likelihood proce dure Results are displayed in both table and graphic formats Price list is a short cut to the price list editing section of the program The Database button opens a menu of possible actions to be done on a library copy or delete or a problem create copy or delete The Load button opens or saves a document Print produces three reports pertinent to a problem All three open the report printing utility Reports are first displayed on the screen be fore they are sent to the printer All three reports show the structure of the problem and not its solution They are useful in the diagnostic of spurious results Setting Up a Problem
23. Peanut meal solvent 98 00 166 5 Poultry byproduct meal 200 00 168 5 Soybean meal solv 44 179 00 170 5 Soybean meal solv 48 189 00 Section 4 Program Tutorial Step 1 Create a price list for the Southeast a Click Price lists on the main menu b Click the Database button on the menu c Select Create new price list d Use the lt TAB gt key to move from the JD field to the Price list name field Enter Southeast 2 00 then click the OK button e Position the cursor on the new Southeast 2 00 library in the left win dow f Add prices of feedstuffs in the right window There are two ways to do this 1 Use the Database gt Add new feedstuffs command A window will open Select the desired feedstuffs either by entering its number at the prompt or by selecting it from the scroll down window Press the lt TAB gt key on your keyboard to move to the price field Enter the price and press lt ENTER gt 2 Use the drag and drop feature In the Available feedstuffs win dow scroll down until you locate the desired feedstuffs To speed up this process you can enter the first few letters in the search box located in the top left corner of the Available feedstuffs win dow Click and hold down the left mouse button Then drag and drop the feedstuffs into the Southeast price list window top right window Once you release the left mouse button the cur sor will be automatically positioned in the Price field Type in the pric
24. ation set automatically e Studentized Residuals Feedstuffs with an absolute value of Studentized residual exceeding 2 are possible outliers It is advisable to remove them from the Calibration before the problem is resolved Both wizzards remove outliers automatically e The Lower 75 confidence limit and the upper 75 confidence limit on the estimated break even price The sixth table reports the results for the Appraisal set Table 6 Appraisal Set O o eeana seat 4 0 0 000 81 708 81 708 97 000 130 114 33 11 Section 5 Advanced Topics 5 8 The Appraisal Set lists e The actual price of each feedstuff e The Predicted estimated break even price of each feedstuff e The residual value errors labeled Pred Act The last table reports the Calculation log It contains useful information if one of the two Wizzards was used The table reports the action taken by the Wizzard followed by the reason for such action Details of Methodology Used In the case where two nutrients are being valuated the set of equations is as follows 2 Y X Bte 1 J where Y price per ton of feedstuffs i X amount of nutrient j in a ton of feedstuffs i B cost per unit of nutrient j and e error term This set of equations can be easily expanded to multiple nutrients using matrix notation Y XBre 2 where Y is ann x 1 vector of prices for n feedstuffs X is ann x m matrix of coefficients of m nu
25. d hulls Hi CP 81 59 Wheat Midds Lo CP 78 79 Q2 10 98 per pound 4 17 Section 5 Advanced Topics Section 5 Advanced Topics The Short Solver Report The short report prints only the most important statistics In the first section a single number is reported the Price Prediction Reli ability For those with a background in statistics this is the square root of the error mean square This value should generally be less than 30 to ensure reliable results Table 1 Price Prediction Reliability Price Prediction Reliability 10 694 The second table lists the Estimates of Nutrient Unit Costs For each nutri ent the column labeled Estimate represents the estimated cost per unit of nutrient i e per lb or kg For U S users the value is in dollars per unit The next column reports the significance of the estimate i e whether the estimate is significantly different from zero 5 1 Section 5 Advanced Topics 5 2 Table 2 Estimate of Nutrient Unit Cost The third table reports various statistics for the feedstuffs in the Calibration set For each feedstuff the table reports the actual price the predicted value break even price plus the lower and upper limit for the predicted value the 75 confidence range Table 3 Calibration Set Section 5 Advanced Topics 5 3 The fourth table reports the results for the Appraisal set For each feedstuff the table reports the actual price followed
26. del is over parameterized too many nutrients for the level of information that can be extracted from the set of feedstuffs in the Cali bration set If one or more VIF are large the problem should be re solved after removing the nutrient with the largest VIF from the prob lem This action is handled automatically by both Wizzards Section 5 Advanced Topics 5 6 The fourth table prints the Correlation matrix of the estimated costs Table 4 Correlation Matrix For example the number in the third column and first row represents the correlation between the estimated cost of the third nutrient and the estimated cost of the first nutrient The fifth table reports the results for the Calibration set Table 5 Calibration Set Section 5 Advanced Topics 5 7 The Calibration Set lists e The actual price of each feedstuff e The Predicted estimated break even price of each feedstuff e The residual value errors labeled Pred Act e The Leverage value The leverage value measures the weight of each feedtuffs in determining the estimates of nutrient costs High leverage values are undesirable Feedstuffs with a high leverage value flagged with a single should be questioned Feedstuffs with a very high lever age value flagged with a double should almost invariably be re moved from the Calibration set before the problem is resolved Both Wizzards test for high leverage feedstuffs and remove them from the Calibr
27. dow Price lists is a short cut to the price list editing section of the program The Database button allows the User to copy or delete a library or to create copy or delete a feedstuff The Find button opens a dialog box where you can enter search criteria for locating or parsing a range of feedstuff records Section 2 Program Overview 2 5 Print opens the report printing utility The report is first displayed on the screen before it is sent to the printer Help opens a context sensitive help system End closes the feedstuffs database and returns the User to the main menu Nutrient Definition x The Configuration Menu Kord Configuration The Configuration menu gives access to the underlying structure of the nu trient definition table Using this form the User can insert or delete nutrients or edit ID names or equations used in the automatic calculation of some nutrients Over 140 nutrients are defined in SESAME to cover applications across a multitude of species Nutrients are arranged in logical groups to facilitate data entry Nutrients can be defined as direct entries e g crude protein or as calculated nutrients e g non fiber carbohydrates NFC Calculated nu trients are defined using equations inserted in the Formulae section of the program Most SESAME users will never have to use the Configuration form Ad vanced users are referred to the Help system for additional information Section 2 Pr
28. e Southeast but under priced in the Midwest Tutorial Problem 5 Section 4 Program Tutorial Go west my son Your son Billy Paul followed your advice He is now in California and works with the top herds in the County He is an Ohio State graduate Therefore he is a top nutritionist and he considers amino acids especially Methionine and Lysine when balancing diets for lactating cows Local West Coast prices of feedstuffs are reported in the following table In addition your son is being offered Almond hulls with higher crude protein 5 CP and lower protein Wheat Midds 13 CP He wants to know the break even prices of these two bargains ID 12 22 28 102 104 110 112 116 118 128 130 134 136 142 146 148 150 152 154 156 158 162 166 168 170 Group AMAAAAnAanAnnnrAnnpR RRR RR RR RRR Feedstuff Almond Hulls Citrus pulp dehydrated Cotton hulls Barley grain Beet pulp dehydrated Corn grain flaked Corn grain ground Hominy feed Molasses sugarcane Soybean hulls Tallow Wheat middlings Blood meal Canola meal solv extd Corn gluten feed Corn gluten meal 60 Cottonseed meal 41 Cottonseed whole Distillers dried grains Feather meal Fish meal Anchovy Meat and bone meal Poultry byproduct meal Soybean meal solv 44 Soybean meal solv 48 ton 90 00 56 00 90 00 105 00 114 00 115 00 105 00 112 00 62 50 74 00 225 00 73 00 370 00 136 00 97 00 29
29. e and press either the lt ENTER gt key of the lt down ar row gt key on your keyboard 4 3 Section 4 Program Tutorial 4 4 Your price list should look as follows when done Ice Normand StF Phr B pia ht DEEN ead Halp End Me iihrart rast Coast izy Wo aT CHEII it SEREA CELLE 10 Han a en DT N DEBRA SA P Fa CEUCLE a Return to the main menu by pressing the END button not the X button in the top right corner Step 2 Create the problem a From the main menu select Solver b Click the Database button on the menu and select Create new problem c Use the lt TAB gt key on your keyboard to move to the Problem name field Type Tutorial Problem 1 Press the lt TAB gt key to move to the Type field Enter 14 to select Dairy CP and NE d In the tree window left window the new problem was created as a child of the Problem folder You should move the Tutorial Problem 1 to the User s Problems folder To do this position the cursor pointing hand and while holding the left mouse button down move the arrow until it points to the User s Problems folder Release the mouse button Now double left click on the User s Problems folder The Tutorial Prob lem I will appear under its new folder Double click on Tutorial Prob lem 1 to select it Section 4 Program Tutorial 4 5 Step 3 Select the price list for the problem a You should still be in the Solver window In the top rig
30. e list label is located in the top right corner To its right there is a button with three red dots Click this button and select the active price list from the choices in the Active price lists window Click OK to save the price list amp Print Help End Mwa lt r gt hr eo teeHa ayv Xe Twe sid ice Cows Dairy Medium Performance mance Cows Dairy High Performance nce Cows Dairy High Performance se Cows Dairy High A Inert NEI Section 2 Program Overview 2 13 Select the feedstuffs that are part of the problem Click the Feedstuffs tab in the top left corner Select a library optional Then drag and drop feedstuffs from the right window to the Calibration or Appraisal sets as appropriate Feedstuffs Sulva 0 Hume StPicic The Ohiv tac Unive sity oT Salve prabiem amp Price tists Ustebesn cy Print Help A End mhea Eb buly kes ewur se po qa 4 A G ao Fader an Corn groin crocked 1 Ltorn groin loked NE miCorn groin ground 1 CGiCorn groin high maisi FiCorn Silage few eors Corn Silage wall core J gka EjCotion hulls 28 Oe dipa teen PiDistillers dried groins z DFoother maal 156 ee ae z Drag feedstuffs lh ta SCardarsel stewie from the library CiFish meol Menhoden eA a nh E and drop into the Hominy feed 116 x Dhool ond bona meol Appraisal or nlecemr comnrennon z DMa H Calibration sets Collbratlon zot 13 Approlcol cat FF maaeprermp
31. e to the bulk of the remaining abbreviations The presence of one or more outli ers invalidates the assumption of normal distribution of errors in 2 result ing in an estimate of b that is no longer a maximum likelihood estimate Many different rules have been proposed for identifying and rejecting outli ers However the automatic rejection of outliers is not always a very wise procedure At times the outlier is giving information not provided by other observations In SESAME we chose to include an automatic rejection scheme of possible outliers The user is provided with studentized residual mea sures and observations with studentized residuals that exceed a default 2 0 standard deviations are automatically rejected by the Wizzard and should be looked at closely to ensure that the nutritional content of the feedstuffs is realistic and that the market price used is real 5 10 Section 5 Advanced Topics Leveraged Observations The detection of leveraged observations is a topic that receives little coverage although leveraged observations can result in substantial prob lems in parameter estimation A leverage point is an outlier among the pre dictor variables Leverage points can be identified statistically from the le verage values h which are the diagonal elements of the projection matrix H X X X X 8 Thus we can rewrite 7 as follows A Var Y h0 9 Expressed this way the effect of the leverage value
32. e unit costs of nutrients using market prices and composition of feedstuffs if the residuals are independent have common variance and are all distributed according to a normal distribution with a mean of zero and a variance of o Consequently it is important in the appli cation of our ML method to ensure that these assumptions are realistic A direct benefit of using a ML approach is that various statistics can be calculated using conventional well known statistical method In particu lar an estimate of the variance 0 is o Y Y Y X X X X Y n m 5 Y Y Y Xb n m Y Y b X Y n m The variance of estimated unit costs of nutrients b is Var b X X 0 6 5 9 Section 5 Advanced Topics Lastly the variance of the predicted break even price of a feedstuffs is given by Var Y X XXX o 7 Estimation Problems The Intercept Question The inclusion of an intercept term in model 2 allows the estimation of an overall average residual i e a number equal to the average value of feedstuffs after the value of the nutrients have been accounted for In gen eral this term should be close to zero Failure to include an intercept term when indeed one is present in the population of feedstuffs being studied leads to biased estimates of the other elements of b In SESAME the default model does not include an intercept Outliers The term outlier refers to an observation which is extreme relativ
33. ey grain Beet pulp dehydrated Corn grain ground Hominy feed Molasses sugarcane Soybean hulls Tallow Wheat middlings Blood meal Brewers grains dehydrated Canola meal solv extd Corn gluten feed Corn gluten meal 60 Cottonseed meal 41 Cottonseed whole Distillers dried grains Feather meal Fish meal Menhaden Meat and bone meal Poultry byproduct meal Soybean meal solv 44 Soybean meal solv 48 ton 100 00 120 00 85 00 92 50 135 00 78 57 72 00 66 25 55 00 220 00 48 00 410 00 80 00 110 00 58 00 250 00 155 00 170 00 90 00 205 00 375 00 170 00 240 00 158 00 167 00 4 12 Section 4 Program Tutorial 4 13 Hints 1 Create a Midwest 2 00 price list 2 Go to SOLVER and change the active price list to Midwest 2 00 3 Copy Problem 3 as a base to create Tutorial Problem 4 Remove double left click from the Calibration Set feedstuffs that are not traded in the Midwest easily identified by a 0 00 price in the Calibration Set Add feedstuffs to the Calibration set as appropriate these are easily located in the feedstuffs window because they have a price in the price column but are shown on a gray background Remove 22 Citrus pulp 132 Wheat grain 164 Peanut meal Add 220 Alfalfa pellets 142 Canola meal 4 Solve Additional Exploration in Tutorial Problem 4 Q1 Is effective fiber worth more in the Midwest than the Southeast Q2 Are there feedstuffs that were over priced in th
34. g point operation Section 4 Program Tutorial 4 1 Section 4 Program Version 2 See tutorial in file Tutorial pif About the Tutorial The tutorial is not designed to teach you everything about SESAME The tutorial is designed to be progressively more complicated with fewer in structions as you move through the sequence of problems provided in the tutorial Your answers should match those provided in the section Answers to Tutorial Questions If not you can load all the data required for a given tutorial from the CD ROM directory TUTORIALS To load the tutorial database 1 First Backup your current database This is important because the tuto rial files will override your current working files To do a backup a Press the lt Backup gt button on the main screen Select Backup c Choose the directory where you want the backup file to reside Ac cept the suggested file name and press lt Save gt 2 Do a backup recovery of the tutorial files a Press the lt Backup gt button on the main menu b Press Recover c Locate the appropriate tutorial backup file on the CD ROM TUTORIALS directory These are labeled Tutorial 1 Tuto rial 2 Tutorial 5 to correspond to each of the five tutorial exer cises Press lt Enter gt d SESAME warns you that all current working databases will be re placed Press lt yes gt to continue SESAME will decompress the da tabase files When done press the
35. he problem in Solver a Copy Tutorial Problem 1 to Tutorial Problem 2 Step 2 Change the problem type Section 4 Program Tutorial 4 9 a Highlight Tutorial Problem 2 Click in the field under the Type column then press the lt Enter gt key on your keyboard to open the option win b Select Type 13 and post your changes by clicking on the green dow SESAME comes with pre defined sets of nutrients Nutrients 10 Crude Protein Non effective NDF Effective NDF Net Energy Gain Net Energy Lactation Rumen Degradable Protein Rumen Undegradable Protein RUP Methionine RUP Lysine Rare and Rumen Intert NE x gt lt xm K KX XX taskbar Step 3 Solve the problem a Use the Wizzard with the default values 11 x gt lt x KK XX Types 12 13 X X X X X X X X 14 18 x lt lt in the Section 4 Program Tutorial 4 10 Additional Exploration in Tutorial Problem 2 Q1 What are the estimated costs of 1 Ib of RUP 1 lb of RDP and 1 Mcal of NE Q2 What feedstuffs did the Wizzard remove from the calibration set during the solving procedure Q3 Are there feedstuffs that were considered over priced using crude protein and NE Problem 1 that are no longer over priced when protein degradability is factored in Tutorial Problem 3 Your Southeast friend calls you back two weeks later He is now dealing with a group of progressive dairy farmers They want to know jointly
36. ht corner locate the price list indicator the small box located just under the Help and End button in the menu bar Open the price list box by pressing the button on its right red dots The Active price list window opens Select Southeast 2 00 then click lt OK gt Prices of feedstuffs in the Calibration set are updated Step 4 Select the feedstuffs in the problem a You should still be in the Solver window with Tutorial Problem 1 high lighted b There are three tabs sitting on top of the tree window Click on the Feedstuffs tab The feed libraries should appear in the tree window Double click the NRC 1989 library All NRC feedstuffs should appear in the feedstuffs window on the right gray background c One by one move the feedstuffs that are also in your Southeast price list from the feedstuffs window to the Calibration Set To do this first select feed 22 Citrus pulp dry by left clicking on it in the feedstuffs window While holding the left mouse button down drag and drop Citrus pulp in the Calibration window Repeat this step for all the feedstuffs in the Southeast Locating the relevant feedstuffs is easy because only those with a number in the cost column should be put in the Calibration Set Notice the change in the background color of feedstuffs in the feedstuffs window once they are moved to the Calibration Set d Select the Problems tab Section 4 Program Tutorial 4 6 Your screen should now look like th
37. ion 3 Helpful Hints 2 The problem is properly specified but the error occurs during an itera tion of the Wizzard This can happens if for example the Wizzard re moves from the Calibration set because they are considered outliers or highly leveraged all the ingredients containing a certain nutrient The problem is diagnosed using the Stepwise Wizzard You can remedy this problem by solving the problem using the Standard Solver as opposed to any of the two Wizzards Ex West Coast Feeds High Performance Cows Use the standard Solver to get a solution Wizzard tm type C Standard Solver Solve using the Wizzard tm C Solve using the Stepwise Wizzard tm Use this Wizzard to diag Variance inflation threshold fi 5 nose the problem Student res threshold 2 M Include log with report Jv OK X Cancel Invalid floating point operation This error occurs whenever you are trying to Solve for a nutrient which is an exact linear combination of one or more other nutrients also part of the problem For example if NE is calculated from TDN then you cannot put both TDN and NE in the problem Similarly you cannot put Crude protein CP Degradable pro tein RDP and Undegradable protein RUP in the same problem be cause CP RDP RUP The same would apply for NDF effective NDF and non effective NDF In each instance the solution is to re move one nutrient from the problem Invalid floatin
38. is gt Solve prablem amp Price ists 3 Dotobase ties Jiwe sey ange Prebiens Lond Print Help End lt s bP hw Fe ee oe KO MIZZ s Vrrotiama a Larsia Jil tiv Vag FFI Exomplee Joiry 1 Ho Toad dbl mart oat Exomples Jihers 2 PUser s Problems 29 Exonple of User s Prob az i Ed Colltro lonset 3 BE Appralza set z D s 4 E y KD n yene yx e Step 5 Solve the problem a You should still be in the Solver window b Click the Solve problem command on the menu An option window opens up Do not click the With intercept option in general you don t want an intercept Keep the wizard type Solve using the Wizzard The Variance inflation threshold should be 15 and the Student res thresh old at 2 These are options useful to experienced users and control the actions of the Wizzard For now you should use the default values c Press lt OK gt d SESAME goes through a series of iterations The following Short report is automatically displayed Section 4 Program Tutorial ii Graph 4Opan WA Sove 2Daleto morked Pr rtsotup Print MConcal Ds i KPRI IDL A d haxhi iki 9e O TEATE ETORT OLE Re reyd s EEEE AAE E a GEERT EE E 8 eae OTET Pea ee v Seieereadtse dtm ecbesh wr ix wag acon Cok Meu eeu elite Sere i Fibra Ta Sear sad ed bak there oh wow The Price Prediction Reliability
39. iy we BOK ay JcSonerd es 14 4 lonad cts Amii a A rons ae AAAA AR Section 2 Program Overview Quick Overview of the Principles Behind SESAME To understand how SESAME works we must go back to the Petersen method first published in 1932 in which the energy and protein composition of corn grain and soybean meal are equated to their respective prices setting a set of two equations with two unknowns Using NRC 1989 as a reference for nutritional composition the following set of equations is derived with corn and soybean meal priced at 90 and 145 per ton respectively Corn 90 176 0 CP 1566 NE Soybean meal 145 991 8 CP 1638 NE where CP represents the cost of one pound of crude protein CP and NE the cost of one Mcal of Net Energy lactation NE The first equation says this in one ton of corn there are 176 lbs of crude protein and 1566 Mcals of NE The cost of one ton of corn 90 is equal to its CP value 176 lbs multiplied by the unit cost of one lb of CP plus it s NE value 1566 Mcals multiplied by the unit cost of one Mcal From elementary algebra we know that we can solve a system of two equations in two unknowns That is we can find the values of CP and NE that satisfy the set of equations In our example the solution is sound when CP 0 063 and NE 0 050 These two values would change whenever the price of either corn or soybean meal would change Knowing that a ton of hominy contai
40. late veg fsz Ail ollo Siogec early bloom 54 mair at QAG 1 Angle porsace chy 16 1 Berrnschagrars Hap 18 1 Biene Hapa te veg a 1 Bione Hap bh ite biam a a a4 1 Cime Ha Ladino Select the feedstuff to copy by left mouse Composition clicking on its name To edit a feedstuff select the feedstuff by clicking on its name and then click the Composition button in the Feedstuffs database window In the composition window select the appropriate group of nutrients Basic Chemical Analysis groups nutrients used across species Values for spe cific feedstuffs can be edited in the Value column Select the appropriate nutrient group Edit values Ihan canprctwn amey iaar Di Naman Seinen Iha lihin ttaba IHmivarzty 2 F Becalculatian JA Dvorviow B Print Ina Shem composition dA MIND ACIDS 1001 Flues atoe BASIC CHEMICAL ANALYSIS 1000 Tahe due BMINERALS 1002 e ALUES FOR FISH 1021 ON ALUES FOR HOASES 1913 AW ALUES FOR POULTRY 1010 VALUES FOA RABBITS i020 YV ALUES FOR RUMINANTS 1014 AW ALUES FOR SWINE 1006 QW ITAMINS 1005 W478 9990740774 Save your changes by clicking the green button 2 9 Section 2 Program Overview 2 10 2 Create edit a price list Feed prices are entered in the Price list win dow It is accessed via the Price lists button on the main menu or via a shortcut in the Solver window Price lists
41. n S Poeceword chongo 2 The Password desktop window is displayed Crystal Soft Administration Center Priority Level Administrator Charlie HHH Billy Paul Hie SESAME SESAME Priority Level User name Password field field Section 1 Getting Started 1 4 3 To add a new User or Administrator click the button on the task bar This creates a new row in the table Click on the User field and enter a User name Press lt tab gt to move to the Pass word field Enter a password of your choice Finally click the appropriate circle to assign that individual an Administrator or User priority level To edit the User name or password for a particular User or Administrator that you have entered highlight the appropriate User name Click the button on the task bar You may then edit the User name password or priority level for a particular individual To delete a User or Administrator highlight the appropriate User name in the User field Click the button on the task bar 4 Click on the green check mark in the task bar The additions edits or deletions that you have made are now saved Button Password desktop Crystal Soft Administration Center m croneet x o M username Password application i Application SESAME Bity Paul SESAME Close 5 Press the Close button in the lower right to finish the process Key to Password de
42. ns 207 lbs of CP and 1638 Mcals of NE we could easily calculate its economic value as follow Hominy Value ton 207 x 0 063 1638 x 0 050 94 94 Thus hominy is considered under priced if its cost is less than 94 94 ton We have studied this method extensively over a 15 year period using com modity prices in three U S markets west coast mid west and east coast and found that it is fundamentally flawed It assumes inefficient markets in commodity trading and implies economically irrational and incoherent be havioral pattern by buyers and sellers of commodities According to this method corn and soybean meal should never be purchased because there are always other commodities that are better priced 2 15 Section 2 Program Overview The FEEDVAL program from Wisconsin suffers from the same flaws Worse yet it assumes that other referee feeds e g blood meal urea are also per fectly priced according to their nutritional contents In SESAME we assume that feedstuffs are sold and purchased for their nu tritional contents but we do not rely on referee feeds That is we do not assume that the price of any feedstuffs is exactly equal to its nutritional con tent As a result corn and soybean meal are sometimes over priced and sometimes under priced depending on the market condition and the class of livestock being considered To understand how SESAME works let us consider the problem of estimat ing the value of
43. ogram Overview 2 6 Prices of Feedstuffs The Price Lists Menu a Price lists Various price lists can be set to reflect different prices across space mar kets or time Feedstuffs can be added to a price list using a convenient es drag and drop feature There are no limits to the number of price lists Key to Price lists that can be generated menu task bar Database Print Help End First row Previous row Price Di Heweand St Pime The Ohiu Stuer Universi eres tn ani F zx Next row D EN P Pias lists TF al 7 e na emMidwest vas 1a E amp Mark position Southeosi 2M0 5 i y Weat Coari 4 3 Mr return T z Last row Cr o Ta Insert row Select the price list by clicking Delete row on the Price Edit row Lists root direc Post change tory and select ing the desired Cancel change price list Refresh data educa car mika POE PEETERS PITTEN elda as gt cbein Individual feedstuffs can be dragged from the Available feedstuffs window and dropped into the price list Menu items in Price lists are The Database button opens a menu of possible actions to be done on a price list library copy or delete a price list create or delete or a feedstuffs add or delete Print opens the report printing utility The report is first displayed on the screen before it is sent to the printer Help opens a context sensitive help system End closes the p
44. p 2 Create new price list West Coast 2 00 Step 3 Create new problem Tutorial Problem 5 Hint select problem Type 11 Do not forget to change the active price list and post your changes And add the two unknown feedstuffs Almond Hulls Hi CP Wheat Midds Lo CP to the Appraisal set Step 4 Solve Additional Exploration in Tutorial Problem 5 Ql Q2 What are the break even prices of the two feedstuffs What is a 1 Ib of undegradable Methionine worth 4 15 Section 4 Program Tutorial 4 16 Answers to Tutorial Questions Tutorial 1 Q1 Yes because the Prediction Price Reliability is equal to 29 4 which is less than the maximum recommended level of 30 Q2 0 094 per pound of CP Q3 0 052 per Mcal of NE Q4 Underpriced feedstuffs Bakery waste Peanut meal Corn gluten feed Soyhulls Cottonseed meal Wheat grain Distillers dried grains Wheat Middlings Feather meal Q5 Overpriced feedstuffs Barley grain Cotton hulls Beet pulp Fish meal Blood meal Meat and Bone meal Corn gluten meal Soybean meal 44 Tutorial 2 Q1 0 214 per pound of RUP 0 013 per pound of RDP 0 054 per Mcal of NE Q2 Tallow Blood meal Fish meal Menhaden and Feather meal Q3 Corn gluten meal Section 4 Program Tutorial Tutorial 3 Q1 Early bloom alfalfa hay 113 50 Full bloom alfalfa hay 93 88 Q2 No Tutorial 4 Q1 Midwest 0 14 Ib Southeast 0 045 Ib Q2 Meat and Bone meal Tutorial 5 Q1 Almon
45. rewv x amp egay Ke codduty nave irae Price afan keta rome dQ la canes ieved 1xXX VF 3 Alc od lig 1 Qusiy AA Zor ddorised PX 15 Almere ds Love Osiy ar Jawn el 0 y XE nhanh alex w81 Ed oiread oroz wh Eq n ke cledgers xx AARAA K Calibration Set Appraisal Set Select the nutrients that are part of the problem optional Nutri ents were selected according to a type i e Type 18 Dairy Heifers when you created the problem You can edit the nutrients selected by clicking on the Chem Composition tab in the top left corner Select the appro priate group of nutrients in the left window A list of nutrients appear in the right window Double click in a nutrient box to either add or remove a nutrient Make sure that you click the green in the taskbar just below the active price list label to save your changes Section 2 Program Overview 2 14 Chem Composition Click the to save changes Suha Di Nuneund 54 Pane The Ohiv Stube Univerty lt 7 Solve problem amp Price tists Detebeee Terolars Leste ote Iran compre Wa T a 4 Chem camposiilon AMINO ACIDS 1001 EASIC CHEMICAL ANALYSIS 1000 MINERALS 1002 VALUES FOR FISH 1021 VALUES FOR HORSES 1013 M VALUES FOR POULTRY 100 VALUES FOR AABBITS 120 VALUES FOR RUMINANTS 1014 SELES es uH ENERGY VALUES 1018 SJHDECAADABLE AXING ACIDS 4 OM 1011 DJHDEGAADABLE AMINO ACIDS 3 UIP 1016 VALUES FOA SHIHE 1006
46. rice list database and returns the User to the main menu Section 2 Program Overview Finding Nutrient Costs and Break even Prices The Solver Menu The core engine resides within the Solver menu To create a prob lem the user must indicate what feedstuffs nutrients and prices are part of the problem refer to Program Tutorial Section 5 The Cali bration set contains all feedstuffs traded in a given market Feedstuffs are added or deleted from this set through a simple drag and drop function Alternatively the user can identify those feedstuffs for which prices are not available but for which break even prices are desired in the Appraisal set The Solver form has three tabs that are used to perform different functions The Problems tab is used for setting viewing and chang ing problems the Feedstuffs tab for selecting feedstuffs within a problem and the Chemical composition tab for selecting the chemi cal elements nutrients for which economic values are to be esti mated Solve problem Price lists Database Load Print Help End Suha Di Nuneund 54 Pawe The Ohiv Stat Univavay fl Solve problem amp Price tists Dalobase Laad m Plint Help A Eha Ianis Leet Uren cre x m m 4 er o wt awzaywxe e Problems 4 ee a a ae Problemas i 2 Po ke Teed bons Teer ca Tae boy Aa y Aiae Pe samples Dairy 1 EFs y a3 5a gt Fie A te ar af et Ly gleams aEx Eost Cone Feeds High Par TE Sea Dent Fok
47. s at discounted prices 3 The value of a feed ingredient is equal to the sum of the values of its nutrients Feedstuffs are used exclusively as sources of nutrients Feedstuffs with valuable characteristics other than nutrient content e g free flow agents will not be evaluated properly by SESAME 4 Markets are in a constant state of price adjustments Therefore at any point in time the price of a feedstuffs is equal to the value of its nutrients plus or minus a deviation called error We assume that all errors are independent and follow a Normal bell shape distribution A Simple Example We used February 2000 Ohio prices and standard nutritional composition of 22 feedstuffs to estimate their break even prices when used with high producing dairy cows Thus we selected the following nutrients for our evaluation rumen undegradable protein RUP effective NDF eNDF non effective NDF neNDF rumen degradable protein RDP rumen undegradable methionine U Methionine rumen undegradable lysine U Lys and net energy for lactation NE Prices used were those reported by Feedstuffs magazine for the week of February 14 2000 for the Buffalo NY market with an additional 20 ton added to cover handling margins and transportation charges In a few instances Chicago IL prices had to be used for which a 30 ton gross margin was added Results as reported by SESAME are shown in the following output screen oee 9 20000 rnc
48. s where VIF gt 20 action must be taken Additional feedstuffs with different nutritional char acteristics can be included in the data and this may result in reduced col 5 11 Section 5 Advanced Topics 5 12 linearities Alternatively the nutrient associated with the largest VIF can be removed from the model The remaining parameters must then be re esti mated using the reduced X matrix In SESAME the Wizzard automati cally removes nutrients with a VIF exceeding the default value of 15 Section 6 Help 6 1 Section 6 Help When the Help System Fails When the help system fails 1 Backup your data You may have to send us your backup if we cannot diagnose your problem on the fly 2 Contact us Dr Normand St Pierre Department of Animal Sciences The Ohio State University 221 A Animal Science Building 2029 Fyffe Road Columbus Ohio 43210 614 292 6507 FAX 614 292 1515 email st pierre 8 osu edu 3 Describe the problem in as much detail as possible We will try to answer your problem as quickly as possible Section 6 Help 6 2 When the Help System Fails When the help system fails 1 Backup your data You may have to send us your backup if we cannot diagnose your problem on the fly 2 Contact us Dr Normand St Pierre Department of Animal Sciences The Ohio State University 221 A Animal Science Building 2029 Fyffe Road Columbus Ohio 43210 614 292 6507
49. se results for example a user would determine that under these market conditions brewers dried grains are relatively over priced whereas wheat middlings are relatively under priced Results are also re ported graphically In these graphic representations 1 e refer to page 4 7 feedstuffs appearing above the middle horizontal line are relatively well priced whereas those appearing under the line are relatively over priced During the solution process SESAME concluded that dehydrated beet pulp was so over priced that it appeared to be an outlier Automatically SESAME moved this feedstuffs from the Calibration set to the Appraisal set Out of curiosity we added the NRC 1989 alfalfa hay sun cured early vegetative to the Appraisal set SESAME calculated its breakeven price at 128 ton a reasonable figure given the market conditions 2 19 Section 3 Helpful Hints Section 3 Helpful Hints SESAME is not a ration balancer SESAME cannot tell which ingredients to purchase and in what amounts they should be fed SESAME identifies properly the relative value of feedstuffs Thus you can identify feed ingredients that are economically more attractive than others Estimated break even costs of feedstuffs are based on the specific class of livestock considered Blood meal for example does not carry the same inherent value in low vs high producing herds SESAME reflects this conditional valuation Backup your work regularly
50. sktop task bar First row H Previous page Previous row Next row Last row Last row I Add new row Mark position return Insert row Delete row Edit row Post change Cancel change Refresh data If you want to change a User name and or password that has been functional in the past it is recommended that you follow the procedure used to add a new User or Administrator When you are sure that your new User name and password are functional you can re enter the Password desktop win dow and delete the old User name and Password by clicking the minus but ton on the task bar Section 1 Getting Started 1 6 System Requirements SESAME operates off of a Windows 95 98 system The program requires approximately 30 MB of hard drive space to install Once SESAME is in stalled it will occupy approximately 12 MB of hard drive memory Your Password The User name and password for SESAME as shipped is User name sesame Password open We strongly recommend for security purposes that you change the User name and password immediately Section 2 Program Overview Section 2 Program Overview What It Is amp What It Does SESAME estimates unit costs of nutrients These costs are extracted from market prices of all feedstuffs traded in a given market SESAME estimates break even prices of feedstuffs The value of a feedstuff is calculated as the sum of the values of its nutrients
51. st be fore you start drag and dropping feedstuffs to the Calibration Set Once you have selected the desired price list all relevant feedstuffs in the feedstuffs window have a price attached to them in the Price column This makes it easy to locate the feedstuffs to drag and drop Use the Print gt C comp of problem in SOLVER to see the structure of the problem and the composition of the selected ingredients on an as fed basis Error A inversion This error occurs whenever you try to estimate the cost of a nutrient which has a zero value for all the feedstuffs in the Calibration set This can occur one of two ways Error A inversion 3 2 Section 3 Helpful Hints 3 3 1 The problem is initially misspecified This means that the initial Cali bration set does not contain any feedstuffs containing this nutrient You can diagnose this condition using the Print gt C comp of problem option 3 Ex Et Coen Fond High Pedowsowe Com Problema Exomplee Osiry 1 3 En Mido Cone Food High Pedowscrwce Com Bix Eort Conert Feader R 7 3 En What Cost Voy H gh Podarworoo Cone mbuser s Prestens Print the structure of the problem One column of the printed table would then contain zeros only as in the following example Column contains only Zeros If this is the problem then you must either remove the nutrient from the problem or add a few ingredients that contain the nutrient in question Sect
52. trients for n feedstuffs B is an m x 1 vector of unit costs of m nutrients and is ann x 1 vector of errors Section 5 Advanced Topics The system of equations 2 has an infinite number of solutions because it contains more unknowns than the number of equations n unknown e and m unknown B in a system of n equations The system however has a unique solution if we set for objective to minimize the sum of squares of deviations i e mimimize U e In matrix notation this implies the minimization of qf F E Y Y 2 B X Y B X XB 3 Differentiating 3 with respect to B and setting the resulting matrix equa tion equal to a vector of zeros the least squares estimate of B is the vector b b X X X Y 4 This solution b has a unique set of properties First b is an estimate of B that minimizes the error sum of squares E E regardless of any distribu tion properties of the errors Second the elements of b are linear functions of Y Y Y and provide unbiased estimates of B These estimates have the minimum variances among any estimates that are linear functions of the Y s irrespective of distribution properties of the errors Last and impor tantly if the errors are independent and identically distributed from a normal distribution with a mean of 0 and a variance of 0 i e N O Io then b is the maximum likelihood ML estimate of B Therefore equation 4 produces ML estimates of th
53. using the Backup feature on the main menu Often it is easier to copy a database and then make modifications to the copy than to create a database from scratch Thus if you need a new price list it can be easier to find copy and edit an existing price list with a similar list of feedstuffs than to create one from scratch If you don t know what to do consult the Help system This system is context sensitive It will provide you with numerous tips and ideas Section 3 Helpful Hints In general it is better to use all feedstuffs traded in a given market as op posed to just a small subset For example you shouldn t use just forage extenders in the calibration set even if your sole interest is in comparing the relative value of forage extenders In SOLVER the Calibration set must contain at least one more feedstuffs than the total number of nutrients whose market values are being esti mated The Appraisal set is optional and may contain as many feedstuffs as you wish In general feedstuffs with known market price and compo sition should be part of the Calibration set The Appraisal set should contain only those feedstuffs with unknown market prices Double right clicking in a window with the exception of the Calibration and Appraisal windows expands it to a full screen format Double right click again to return the window to its original size When creating a new problem in SOLVER select the desired price li
54. word eee OK Concel Poeeword OK Concol Adminietrotion fd miner tien S Poceword chongo S Poceword chango Administrator User License License 6 Enter your Username lt TAB gt then Password Click OK You must enter a valid User name and password to access the main menu THE PASSWORD IS CASE SENSITIVE 1 1 Section 1 Getting Started 1 2 7 Ifyou have a User license SESAME will begin immediately after click ing OK If you have an Administrator license you will gain access to the Password desktop after clicking OK please refer to the Changing Your Password section If you do not need to make any additions edits or deletions to the User name or password click OK again to start SESAME Individuals with an Usor nome pns F Adminstrator license will TITT gain access to the Password Jok ren desktop after clicking OK To start SESAME click OK one more time Adminietrotion SS Poeeword change Changing Your Password As shipped SESAME has two valid User names each associated with a unique password The two User names provide different access levels to the program The first access level is called User If you have a User license there will be a few functions within the program that you cannot access but most of the program functions will be available to you An Administrator license gives you access to all program functions with the exception of two functions that are reserved for
Download Pdf Manuals
Related Search