back start next


[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [ 142 ] [143] [144] [145] [146] [147] [148] [149] [150]


142

The Screen in the Pit

Getting a screen installed in the pit was no easy matter. Although it was in the next building from us, (we were in 440 S. LaSalle and the is at 400 S. LaSalle), turf issues kept us from wiring directly from our Telco closet to their Telco closet only six feet away!

We had to order a dedicated 9600 baud line from Illinois Bell that ran from the 440 building to the switching station on Wabash and back to the 400 building. While we were waiting for the installation, we split the pricer in half. One half would use a 386 PC to paint the screens down on the floor and the other half would listen to current markets on a 286 machine in our upstairs trading room. Thus traffic on the dedicated 9600 line would be kept to items by request only.

We had supposed that the 386 computer would be installed near where the monitors would be installed in the SPX pit. This was not the case. The had a separate computer room in the undercroft of the raised floor specifically for this purpose. It was about 30 feet away. The other trading firms used a computer that had the CPU built into the monitor, but we didnt want to do this because the built-in CPU was only a very slow 8088 and we needed the 286. To overcome the distance problem between the 386 and its monitors, I called around to the electronics supply houses and came up with a set of amplifiers. The manufacturer would not guarantee their use for monitors and the engineer thought it would only work for 20 feet. We tried it anyway and it worked fine. I took the opportunity to remove the floppy drive in the computer in the undercroft for security reasons. All file transfers were done remote via the 9600 baud line. I also taped a black trash bag to the ceiling to prevent dirt from falling through the above floor onto the computer. Two weeks after the line was installed, we were ready to test live on the floor.

the computer sharks

The first week was fraught with system problems. In that week, we lost more money than all but one week of market vagaries. The crowd in the SPX pit was quick to pick up and capitalize on any computer glitch. It was a game to them akin to taking candy from a baby. It was a hard week for the programmers. They worked late into the night to make changes that were furiously tested the next day. In addition, we had not anticipated that some members of the crowd would stand right on top of us and scrutinize everything we did.

The terminal in the pit controlled the market quote screens. The human operator could control the spreads, best bid/offer, behind the best bid/offer, the bias, long or short, the amount of deltas, long or short, in essence, the bet size, which would cause the bias to be less and less as it got closer to the delta limits. I am not sure if we were able to control the other Greek letters. All this information had to be available to the human operator and to many of the members of the crowd as well. I remember standing behind Jim Sauser as another market maker explained to his



buddy where the delta information was located on our screen, and so on. He said it in a matter-of-fact manner as if he was directing a newcomer to the restrooms. He generously gave us credit for things we had not yet implemented and probably would not.

To maintain some secrecy, we came up with a scheme to put all position information in one long number at the bottom of the screen. For example, character positions 3-5 would contain the delta and positions 10-12 would contain the bias. To hide the real information, additional numbers were added. These numbers would change based on a formula off of the real numbers. We tried using pure random numbers, but they would change when the others did not and the randomness became apparent to the crowd. Instead, we programmed a routine that would multiply the Greek letter by linear equations and put the result in the spare digits, and it would all move in sync together either up or down, but nobody else would know which was the real number. We also rotated the character position orders around a bit just to make sure. Before the trading day started, the human operators would go talk to the programmers to get the latest character location information as it usually was changed overnight.

Soon the computer failures decreased and the computer sharks laid off. We gradually got our nerves up, turning on the best/bid offer for a while and then turning it off when it got too hot. We played like this for a few weeks, in the market, out of the market, in the market, always getting braver, staying in the market longer, adding more bias, setting larger deltas, always getting better.

Risk Management

At the same time, work was progressing on the risk management side, which was located in the upstairs office. We added on-line position reports, smile charts, all the Greek letters, everything we could think about to measure our exposure. We also developed coded messages that could be sent down to the floor. These messages occupied a certain two-digit position. These were some of the messages:

• Everything OK.

• Check your screens-I think I see something.

• We are going to lunch.

• Dont feel well.

• Someone is giving me a hard time.

• Someones coming downstairs to see you.

• The boss wants to see you.

• We are experiencing minor technical problems.

• We are experiencing more serious technical problems.

• Panic-pull the plug immediately.



Trading cards were made with all these codes translated and most of the staff kept them handy in their shirt pockets. My favorite message announced "Were eating donuts/bagels/muffins/pizza and theyre going fast."

Enter Cary Grant

Cary Grant came on the scene about this time. Not the Cary Grant, the actor, but Cary Grant the famous technical trader. I never did find out if it was true, but Cary told me that he figured it was worth about $10,000 a year to him to have this name, in terms of people remembering his name, doors opening to him, and especially people not cashing checks he wrote. Cary later went on to become the head of the Third Market Corporation.

Cary was a successful SPU trader (pronounced "spooze," S&P 500 Index futures traded at the Chicago Mercantile exchange). He used a technique based on George Lanes stochastics, Elliott waves and Gann angles. He was using FutureSource to trade SPUs from his desk and doing quite well. Somehow, he knew Steve Fossett and was on very friendly terms. An idea came up at one of our staff meetings, and Steve immediately put us both together and we were totally enthusiastic about the idea.

The idea was this: Nimble up to this point was making money following the market up and down, making money on allowing the orders to come to us. Cary was making money predicting short-term movements of the S&P, splitting his profits with the market makers and the brokers. We would have Cary call the market and instead of picking up his phone to place his order, he would call the upstairs Nimble risk management desk and they would adjust the bias as if he had entered a position in the underlying. We then would be trading without commissions, gaining the bid/offer spread from each transaction, instead of giving it up.

This was easy to implement and success came quickly. The only drawback was that execution was slower than futures directly. Cary would enter an order and it would take a few minutes to become long, and more than a few minutes on a slow day. This didnt seem to affect his trading very much. He took it on faith that orders would be "filled" and when it took time, well, that was the price one had to pay to trade without friction.

Cary did very well and the Nimble team did its best to keep the system operating smoothly and the rest is history. It was like being a runner who trains with ankle weights and takes the weights off before an important race. Carys trading ran hot and stayed hot, thanks to the buffer of accumulated spreads and commissions he did not need to pay.

How the Crowd reacted

Here we were with a successful trading system in the SPX pit and soon were one of the biggest market makers in the pit. We had expected more than the usual hostilities from



[start] [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] [42] [43] [44] [45] [46] [47] [48] [49] [50] [51] [52] [53] [54] [55] [56] [57] [58] [59] [60] [61] [62] [63] [64] [65] [66] [67] [68] [69] [70] [71] [72] [73] [74] [75] [76] [77] [78] [79] [80] [81] [82] [83] [84] [85] [86] [87] [88] [89] [90] [91] [92] [93] [94] [95] [96] [97] [98] [99] [100] [101] [102] [103] [104] [105] [106] [107] [108] [109] [110] [111] [112] [113] [114] [115] [116] [117] [118] [119] [120] [121] [122] [123] [124] [125] [126] [127] [128] [129] [130] [131] [132] [133] [134] [135] [136] [137] [138] [139] [140] [141] [ 142 ] [143] [144] [145] [146] [147] [148] [149] [150]