Plans and Procedures for the Experimental Testing of the FCC Package Bidding System for the 700 MHz Band Auction Submitted by Cybernomics Inc. under Subcontract FCC-00-003 Plan and Procedures for the Experimental Testing of the FCC Package Bidding System for the 700 MHz Band Auction I. Overview Cybernomics Inc., in consultation with Computech, has designed a test environment and protocol to verify the algorithms and rules of the FCC Package Auction for the 700 MHz Band Auction. The test environment uses human bidders to submit strategic bids into the auction system designed by the FCC and implemented by Computech. The bidders have an incentive to discover and exploit any weaknesses in the system , since they will be paid based on the success of the bids they make in the auction vis-à-vis other participants. In addition, they will be given a monetary bounty for any errors they find in the software implementation of the auction rules. Thus, real people with real incentives will be providing Cybernomics and Computech with feedback on the system operation. In addition to the human subject checking procedure, Cybernomics is building a backend system to check the solution and eligibility constraints obtained by the Computech system. The backend system will use a different solver than that used by Computech. Our plan is to have subjects trained in the rules of the auction and then have them participate in a series of auctions. At the end of each day Cybernomics will supply Computech with a summary of the results of the experiments and comments by subjects. Any discrepancies between what the system provides and what the backend or subjects' calculate will be the first items in the report. II. Backend Checking Procedure Cybernomics is developing a backend system that will take the round bids in a standard format and calculate the optimal set of winning bids (bids that maximize revenue and allow each license to be assigned to at most one bidder). The solver used by Cybernomics is based on an off-the-shelf package XA (a professional linear/integer programming system). In addition to XA, Cybernomics has written code to check for minimum bids and eligibility. We will be using a dual Pentium III 500 Mhz processor with 500MB of memory to run the backend. We will post our calculation results on the website http://linus.econlab.arizona.edu/Test after each round. III. Experimental Procedure The Computech system will be tested in the Economic Science Laboratory (ESL) at the University of Arizona using methods in Experimental Economics. The auctions that will be conducted at ESL will use human subjects who will participate in a series of auctions. The subjects will be motivated to participate by being paid based on the success of the bids they make in the auction. Each subject will be paid the difference between their private value for any license they obtain, minus the cost of winning that license in the auction. The private values of each license to each subject will be induced. That is, each license and several packages of licenses will have monetary values for the subjects. They get these values if they win the licenses. Of course they must pay their bids for the license package they win. In the table below, we supply the values for 15 bidders in two environments. Values are stated in experimental dollars that can be redeemed in our experiment at a rate of 1,000,000 experimental dollars equals one US dollar. Thus, in Table 1, subject 9 has a value of 200,000,000 experimental dollars for license 10MHz- Northeast. Table 1 Value Parameters (Values are stated in US Dollars) Warm-up Case Subjects --> License 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10-NE 100 100 100 100 100 100 100 100 200 100 100 100 100 100 100 10-MA 50 50 50 50 50 150 50 50 50 50 50 50 50 50 50 10-SE 150 150 150 250 150 150 150 150 150 150 150 150 150 150 150 10-GL 100 100 100 100 100 100 100 100 100 100 200 100 100 100 100 10-CM 75 75 75 75 75 75 75 75 75 75 75 175 75 75 75 10-P 120 120 120 120 220 120 120 120 120 120 120 120 120 120 120 20-NE 270 170 170 170 170 170 170 170 170 170 170 170 170 170 170 20-MA 200 200 200 200 200 200 300 200 200 200 200 200 200 200 200 20-SE 300 300 300 300 300 300 300 300 300 300 300 300 300 400 300 20-GL 50 50 50 50 50 50 50 150 50 50 50 50 50 50 50 20-CM 75 75 175 75 75 75 75 75 75 75 75 75 75 75 75 20-P 100 200 100 100 100 100 100 100 100 100 100 100 100 100 100 10-NE,SE,GL NA NA NA NA NA NA NA NA NA 350 NA NA NA NA NA 10-P, 20-NE,MA NA NA NA NA NA NA NA NA NA NA NA NA 490 NA NA 20-CM,P NA NA NA NA NA NA NA NA NA NA NA NA NA NA 275 Boundary Case License 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 10-NE 78 56 27 16 23 33 87 12 61 56 51 61 92 95 20 10-MA 99 107 4 103 87 22 33 69 128 90 2 57 122 138 58 10-SE 37 86 100 29 49 51 34 12 33 74 41 44 7 45 32 10-GL 47 150 83 79 1 14 2 151 69 195 124 151 54 129 55 10-CM 127 39 130 243 173 95 116 8 281 46 53 10 32 24 52 10-P 56 52 39 60 29 38 65 78 100 77 54 97 82 71 44 20-NE 120 124 73 22 125 122 129 91 101 64 87 115 19 146 71 20-MA 48 81 70 61 108 111 46 52 82 71 19 72 118 44 41 20-SE 96 44 43 76 48 95 87 93 52 50 16 29 80 98 92 20-GL 27 180 88 137 43 7 60 193 134 72 190 197 88 9 137 20-CM 285 273 142 246 144 235 196 286 184 289 263 195 277 61 239 20-P 59 63 66 29 67 59 36 12 67 81 16 73 15 57 18 NE,MA 35 171 138 59 162 137 151 55 130 219 117 NE,SE 145 137 85 107 131 161 43 97 174 172 172 99 NE,P 148 154 116 95 99 113 93 197 179 111 P2,GL 125 252 176 90 102 51 276 143 222 93 89 P2,GL2 160 186 194 144 125 115 235 287 163 107 204 P2,SE2 214 142 142 116 135 65 152 101 CM,NE2 437 273 382 247 280 164 196 199 239 83 181 144 CM,MA2 278 202 325 420 366 189 78 151 142 146 205 111 138 CM,CM2 596 458 350 421 516 538 368 469 310 313 158 481 MA,GL2 360 GL,CM2 525 SE2,P2 225 NE,GL,CM 421 401 302 409 240 158 388 306 529 290 295 328 435 210 NE,NE2,MA2 360 258 255 260 320 303 278 GL,CM,P 406 274 500 409 272 180 267 322 385 264 294 196 405 295 GL2,CM2,P2 363 507 398 371 456 505 571 466 247 503 P,NE2,CM2 312 384 371 455 513 325 465 NE2,MA2,SE2 356 246 259 295 352 225 168 296 314 SE,GL,CM 277 546 488 582 310 425 288 311 562 525 395 238 131 287 243 SE,CM,NE2 575 NE,P,MA2 337 NE,CM,P,CM2 619 360 525 763 551 722 729 571 787 638 673 288 531 NE,MA2,SE2,GL2 462 368 303 425 319 268 356 392 414 307 375 294 456 CM,SE2,P2,GL2 471 553 537 500 465 324 334 369 350 389 382 265 310 477 NE2,SE2,GL,P 523 502 275 249 260 531 433 419 498 495 387 361 517 365 SE,CM,P,CM2 416 443 649 465 691 635 734 545 652 779 290 397 NE,GL,P2,GL2,NE2 335 465 496 375 350 567 359 575 317 All Licenses 1920 IV. Reporting Process Cybernomics will issue a report after each auction day. The report will consist of any anomalies in the results and potential errors discovered by subjects. In addition, we will post results on our website so that the round files can be compared. The report will also consist of remarks by subjects on screen displays and any other hard to understand interfaces. Any rule inconsistencies or potential loopholes will be reported if found. V. Task Plan Table 2 below lists our planning and reporting timeline. Table 2 Timeline August Task 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Task Plan Protocols and Parameters Rule Analysis Backend Programming Subject Training Conduct Auction and Report