On Wall Street in the era when more and more work was being done on the trading floor on the computers, all sorts of new problems arose that required us to make adjustments. I was working on the trading desk in 1982/83 when Apples and PC started coming into the mainstream. Prior to that there were computers, but they were either mainframes in some remote location of mini-computers located on a different floor wired into small screens on the trading turrets that gave you minimal input capability. I literally brought the first PC onto the trading floor because what I was doing in the futures and options area (I was setting up BT Futures to trade on the various global futures exchanges) was all new and required a lot more on-the-spot programming than routine trading required. The Vax system programmers were busy doing things for the high volume trading operation and our nascent effort was running in between the legs of the big T-Bill auction boys, where they were the programmers were charged to focus their efforts. No one seemed to care enough about what we were doing and it was a fringe effort so we just did what we wanted to do including buying a PC at some computer store, lugging it in and setting it behind us on some random makeshift counter meant for a printer or something. We had that PC running pricing sheets overnight so we could quote prices and those runs were all based off spreadsheets that we had created ourselves…I think we had moved from SuperCalc onto the new whiz-bang Lotus 1-2-3 just recently (Excel is not yet out there and ubiquitous as it is now). But let’s the thing…there was no one checking our work and we had to eyeball our own spreadsheet programs in hopes that we hadn’t made any errors. One minor formulaic error in one cell would then cascade down through multiple cells and could cause a hidden pricing disaster in the making. The trade-offs between timeliness (the markets were always moving….) and accuracy were ever-present and spine-tingling when big trades were on the hoof.
You get to the point where you do everything you can to check and recheck numbers, but you can go blind and dizzy trying to validate rows and columns, looking for mistakes hidden perniciously in the details. There is no more gut-wrenching feeling than having a counterparty say, “you’re done at that level, how much more can I do there?” That’s when you would immediately stand back wondering why he liked your trade so much and what had you likely done wrong to misprice the thing. That made for nightmares that made waking up naked at school seem like child’s play. What you learned early on was hat the only salvation was to check your numbers as best you could and then come back quickly to a counterparty if you found a mistake and declare that it was a mistake and would they be so kind as to unwind that trade. It couldn’t happen too often, but your chances of survival improved dramatically if you were forthright and direct about having made a boneheaded mistake…something everyone could relate to. I sure am glad those days are behind me…..
Well, not so fast Abernathy…..my Australian client has been in a fit of preparation for a court hearing and for several nights they have been calling me to produce revisions to the modeling numbers we have prepared for this case. As I’m explained, some of that has been working very well since my analytics team is in Sicily and once the Aussies call me in the early evening, I can get the request into Sicily to redo the model so that its ready when I wake up. The problem is, there has been an urgency to getting these numbers as the Aussies were preparing for their hearing. And for most of the model adjustments or runs that they wanted, I simply had to wait for the Sicilians to wake up, rerun the number and get them back to me…I could not tickle the keys to the model and make the changes myself. At the beginning of the week that worked fine to get them the next morning, but then later in the week, they needed the numbers faster. That meant that onWednesday night I got my instructions to them at 9pm PDT, they got into the office at 10:30pm PDT (7:30am in Palermo) and then got me the revisions at 12:30am PDT so that I could relay them to the Aussies before toddling off to bed. When I awoke at 4am I checked my email and found that the Aussies wanted another change done ASAP. I shot it to Palermo, but somewhere in the shuffle they missed it and sent me other work for the client at 7am PDT…Yikes. Lucky for me the Aussies were asleeep, so we had the time to get it to them before wakey-wakey time (as they say down under). I gifted we were done.
Then, last night at 9pm PDT the Aussies’s called again looking for yet another calculation, but they needed it ASAP. This time I thought I could do it without Palermo. I did the math and sent the six numbers they needed back. As I ponders the outcome, I suddenly realized I had made a calculation error. I quickly redid the numbers using my Hewlett Packard 12C Financial Calculator, the reliable desktop standard for us finance guys for the last 45 years. It did for me what I had failed to get right in Excel, so I emailed it to the Aussies within 15 minutes of the first erroneous remit. The Aussies asked for a call at midnight just as I was sending the revised numbers. When we got on the call I opened with the ever-self-effacing “now don’t I feel like an old horse’s ass?” They had just seen my new numbers so they understood immediately and we all had a good laugh . I explained that that’s why I had an analyst team…to stop me from making stupid mistakes. One of the senior Aussie guys could relate and said that’s why he never went anywhere without his analyst. All was well and they bid me another late goodnight, thanking me for indulging their Australian fire-drill. Turns out I hadn’t left those days of all-nighters behind me altogether…