I've talked at some length about backup plans on these pages, including backup plans if a computer crashes or an ISP goes down. In the early days of DSL and cable Internet, I kept my dialup ISP for a while as a backup plan, for example. Now mobile platforms for my brokerages are accessible through a Blackberry and a regular platform through a netbook that I can tether to that Blackberry or use with an aircard connection. They, coupled with alerts that warn me to get to the computer, constitute my backup plan for those occurrences.
I've mentioned that another possibility is setting catastrophe-type contingency orders to be triggered if the market crashes and you subsequently crash to the floor, paralyzed with fear. This article was originally roughed out prior to the Thursday, May 6, debacle, but now we all understand what can happen to the markets, even those who didn't understand previously.
Would our contingent or conditional orders have been triggered in such an event? Would we have wanted them to have been triggered? I've heard of some traders whose orders to sell at a loss were triggered just before the huge rebound last Thursday, to their dismay. I've also heard about others who weren't savvy enough about the way volatility would impact prices and whose triggered orders weren't filled, also to their dismay.
And what happens if a market move is occurring when your brokerage's server is down? I didn't hear of that happening on May 6, although I did hear of many platforms that were slow or just not working at all. However, a week or two before the May 6 debacle, one popular platform did fail.
The moment it did, a light bulb went off in my head. I thought I remembered literature from my other brokerage that stated that contingent orders were held on their servers until the specified conditions were met and only then sent out. I immediately began questioning other experienced traders, asking if that was their understanding, too. If I'd had contingent orders placed, and the brokerage's server or servers were down, would the order even be submitted if the appropriate parameters were met? Would those orders have done me any good?
I didn't try to call the brokerage's trading desk that morning, but I did hear from others who did, and who got voice mail. It didn't take long to get the platform up and running again, but those critical few minutes were scary for some in trades that were in trouble. Fortunately, I wasn't in that position.
However, I was left with that question. I wrote the brokerage in question, and received a reply stating, in part, "We do hold conditional orders on our servers. If there was ever an issue w/our system, we do have multiple back-up platforms. If an order was missed because of a system issue, we will always make it good." I'm not specifying the brokerage in question, because I don't think this handling of contingent or conditional orders is specific to that brokerage. My other brokerage also specifies that the order is placed only if the specifications for placing the order are met.
The brokerage in question assured me that if an order was missed because of a problem with their system, they'd "make it good," but what does that mean, exactly? If my criteria had been met, and I'd specified that the order go in at a certain price, a price that turned out to be midway between the bid and the ask or even at what I considered a fair amount above or below that mid price, could the brokerage have insisted--rightly--that it was by no means guaranteed that such an order would have been filled, even if submitted? If such an event had occurred on May 6, for example, it wasn't guaranteed that anything other than a market order would be filled. Not even an order to buy at the ask and sell at the bid is guaranteed, at least according to my understanding. It seems that, depending on the underlying and the market conditions at the time, the brokerage might have a valid argument saying that they hadn't caused any harm and that my order wouldn't have been filled.
So, what can we as traders do, especially if it's possible that we might find our frantic telephone calls met by an immediate transfer to voice mail? Should we even have contingent or conditional orders in, if they're likely to be filled at the extremely adverse levels that were some traders' experiences on May 6?
Speaking in a presentation on May 14, Dan Sheridan counseled that we can't forsake all use of conditional or contingency orders because of one day's events. However, a day or two after the debacle, Tom Sosnoff of think-or-swim counseled that we can't get too cute--my words, not his--with contingency or conditional orders. It's tempting with a platform as versatile as TOS's, of course, but all kinds of difficulties can result from such conditions. It's possible, for example, to imagine that both orders that are part of a OCO series of orders could be filled in a fast-moving market, with the second one filled so shortly after the first one that there's no time to cancel the second order.
First, we can verify with our brokerages how our contingent or conditional orders are treated and how OCO orders are managed during such sharp moves. My primary platform won't let me put through OCO orders if I've violated some formula for how closely the orders can be set. I've been irritated at times when I felt that there was no way both could be triggered in quick succession, but after May 6, can we say that it's impossible for two orders, no matter how far apart we've set them, to be triggered in quick succession?
As many recommend, we should carefully consider what types of contingent or conditional orders we're setting. Maybe orders for spreads aren't a good idea in all market conditions. As we've written on these pages for years, it may not be possible to get spread orders filled in fast-moving market conditions, although I was able to do so on May 6. Sheridan suggests long puts or calls as the most expedient type of order to get through in fast market conditions. His advice points to the need to think of protective measures that can be taken even in fast-moving markets.
For example, with the markets already trending down sharply on May 6, we could have considered in-case-of-catastrophe orders for long puts to help offset the losses we were suffering in other parts of our portfolios. Perhaps those should have been one-triggers-other orders, though, to sell back the protective puts purchased on the way down if the markets reversed sharply, as they did. And then there's the question of whether those should have been market orders. Puts were of course inflated in price, and their values came out quickly on the sharp bounce. It might have been possible to lose quite a bit of money if a protective put had been bought at market and then sold at market as the original trigger level was bypassed on the upside.
There are no magic answers, however. This wasn't an in-depth discussion of how to set conditional or contingent orders. It wasn't meant to be as there are differences among the brokers in the types of orders and ways they're set. This is evidenced by the very fact that they're called "contingent orders" on one of my brokerages and "conditional orders" at the other, one small difference that points to bigger differences in the way these orders might be set at separate brokerages. However, the topics in this article are matters to consider and perhaps to discuss with your broker or the trading desk at your brokerage. How are conditional or contingency orders handled if their servers go down? In adverse market conditions? What were their experiences on May 6? Talk to other traders about their experiences. If your platform has a forum, go to it and read about the experiences of others.
We might take another step that I recently discussed. We can consider having a backup brokerage at which we deposit enough money that we can take defensive action until our other account is accessible. How much money would that be? That would depend on the account and the size of the trades, of course. Some traders might need only a small amount of money, enough to buy a couple of IBM, WMT, MNX, IWM or SPY calls or puts to hedge their 2- or 3-contract butterflies or calendars in those underlyings. Others might need a much bigger account with approved trading levels that provide the ability to trade futures, so that they more fully hedge a bigger position.
In all honesty, we're going to need some of these backup plans in place if we trade more than a few months. Even if we discount what happened on May 6 as an anomaly and not a hint of a new possible reality, I've had landscapers cut through a cable line, fast-market conditions overwhelm a brokerage's platform, and a hurricane that wiped out electrical connections--and my modem--for days on end. I've had a computer freeze at the worst possible moment, and so have many of you. These considerations are not meant to frighten traders but rather to reassure them. Preventative measures can be taken.