About the Two Corner Timing Models

There are two distinct Two Corner Timing models, developed idependently at different times, using different underlying strategies. What they share is being entirely technical, using price and volume data from the Nasdaq 100 index (NDX), and being geared toward short-term swing trading. Like all market timing systems worth considering, they are completely mechanical and based on publicly available information (in our case, index price and volume data). We developed the original model in 2006 and the current model a decade later, based on data going back to 1999. While newer signals and variants are occasionally added, a signal remains unchanged after it has been introduced. While the details of the models are proprietary, there is much to say about their general nature and behavior.

These are very active systems, with trades lasting on average two to four days. While medium-term market trends are taken into account, they are generally used only as a basis of comparison to the shorter term movements of the NDX. This focus on recent action means that not only can our models perform equally well in rising and falling markets, they also appear to make the exact same kinds of choices in both kinds of markets. In truth, the slightly different behavior in different trends is what makes the signals seem to remain steady.

The short-term nature of the Two Corner Timing signals means that they must be acted on quickly. Delaying one day cuts backtested gains down by more than half, and delaying two days actually loses money. It's possible to wait until the open (or at least morning) of the next trading day with a moderate decline from tested results, as mentioned in the "Trading Time" section of our How-To Guide.

Longer-term market timing methods generally try to find the current direction of the market, and short-term methods are generally counter-trend. Two Corner Timing falls almost completely in the latter category. Counter-trend is often counter-intuitive, leading to signals that just "seem wrong". Often enough, it is wrong - our models tend to pick about 55% of the days correctly. This can be frustrating, especially at the inevitable times when a signal has been wrong several days in a row while stubbornly refusing to change when it "obviously" should. But investing is a game of averages, and stepping back a bit tells a different story: while the 55% success rate isn't much better than the market as a whole, our average wins are bigger than the market's, and our average losses are smaller. Over the long run, that makes all the difference.

Despite whatever complexity there is in the models' internal workings, they're set up so that the one day's signal can always be decided from current value of NDX and the daily high and low. This allows us to tell investors in advance what to expect so they don't have to be surprised by a signal at the last minute - see "The Current Signal Page" in our How-To Guide for details.

The Composite Family

A period of unavailablity of the original Two Corner Timing signals (described below) led to the development of the new Composite and Aggressive signals. In that time the performance of the originals has slowed, making these the new flagships of Two Corner Timing.

They're both built around an adaptive system of short-to-meduim technical indicators, with parameters that are automatically adjusted to the current market conditions. This in combination with a few other standard metrics identifies short-term buying opportunities, buying on dips whose levels are defined by recent market bahavior.

Composite debuted at TimerTrac.com at the beginning of 2016, and Aggressive later that year, so they're still more backtested than real-time, but their confirmed performance so far has been just as expected.

Composite

The goal of Composite is to match the long-term gains of the market, while spending most of the time safely in CASH. It seeks out only the clearest of the aforementioned dips, staying in only as long as that opportunity remains. Most signals are given at the close, but it is combined with a volatility-based signal that will occasionally trade at the open. Composite also stays out of the market for all weekends and holidays, as those times offer little reward and relatively high risk.

Composite is LONG-only, with about 75% of the time spent in long stretches of CASH. These periods can last as long as a month with no action, often while the market rises. But a dip always shows up sometime, and will often be caught. Bull market or bear, there will be a time to catch a quick bounce-back, making Composite our most consistent signal — the market goes up and down, but Composite spends most of its time going up.

Aggressive

The Aggressive signal is built on top of Composite, taking all of its buying opportunities. But it's very different the rest of the time. The market usually rises, and Aggressive usually rises along with it; unlike most Two Corner signals, it has a degree of momentum built into it. It spends only about 10% of the time in CASH, and is usually LONG even when Composite isn't. It can also give SHORT signals, more often in bear markets but always a possibility. While it produces about twice the gains that Composite does, it has to work harder for these gains, sacrificing safety for performance. In the short term that means it's as volatile as the market, though in the long term it misses out on the worst of the bear periods.

"Classic" TCT Signal Families and Variants

The original set of Two Corner Timing signals is split broadly into two families, with different variants within those families. They are all related in some way to our original Quick signal. Some variants are tailored to different trading needs, and some present different risk/reward levels. The original Quick signal has the longest recorded real-time history: it has been logged to TimerTrac.com since 2006, the others since 2009 or 2010; but they all tie in to the same model developed in 2006. TimerTrac logging was discontinued when they became unavailable, and has only recently started back up.

The success of this model depends on its volatility, more than on the current direction of the Nasdaq 100. Volatility isn't the same thing as decline; while a falling market is generally more volatile than a rising one, a bull market can still have volatile times. A market may also be "flat" as far as long-term trend is concerned, but as long as there are sufficient day-to-day moves the system can still take advantage of them.

These signals are mostly LONG/SHORT, and much of their historic outperformance has come from the SHORT side. But a little after the start of the current bull market, something changed. The volatility they thrive on isn't nearly what it used to be, and performance has suffered to the point that SHORT signals have stopped working. They'll likely be the big money-maker again, at some future point when the markets are again in turmoil. In the meantime, while the classic signals are still offered, they're no longer the number-one performers they once were.

The Quick Family

The Quick signals use a combination of four different components: one medium-term trend, two short-term technical, and one (the Flow signal) that's a little of both. These independent components are combined into a single signal that is then further refined in the variants listed below. The Quick signals are "pessimistic", with an average 40/60 LONG/SHORT ratio (on a day-by-day basis).

Quick

While this is the original signal of the Two Corner Timing strategy, it isn't exactly the base from which the others derive (that would be Quick Zero, below). This signal is always either LONG or SHORT, never in cash. Signals last an average of 3.5 days, and are almost always at least two days long.

Quick Time Out

This is a variant of Quick which isn't always invested. About 10% of the days are a CASH signal, with the remainder split along the same 40/60 LONG/SHORT lines as Quick. Its gains have generally been a bit higher than Quick and its average signal length shorter at 2.7 days. The reason Quick Time Out wasn't the flagship signal is that the always-invested Quick is simpler to follow and has the longest real-time record.

Quick Zero

This is actually the signal on which the others are based. It's fairly evenly distributed between LONG, SHORT, and CASH, though still a little heavier on the SHORT side. The Quick Time Out and Quick variants were made by adding some post-processing steps to Quick Zero, which wasn't originally designed to be used by itself but was later found to do well in its own right. It is the most active trader of the bunch, with an average signal lasting only a little more than two days.

Quick Zero T3

This variant was designed for the most restrictive of accounts, cash (i.e. non-margin) accounts that can only trade stocks and used to require a three-day settlement period between the sale of one stock and the purchase of another (known as the "T3 settlement period"). As the name implies, it's based on Quick Zero, which already spends significant time in cash. Whenever a signal changes, Quick Zero T3 will stay in CASH for a full three days before opening a new position. This ends up spending less that 40% of the time invested, but makes up for it with a tendency to be invested on the more profitable days. While the three-day settlement period is no more, Quick Zero T3 remains.

The Flow Family

Originally created as a component of the Quick signals, the Flow signal ended up doing very well in its own right. Unlike the Quick family, these signals spend a little more time LONG than SHORT.

Flow

This signal is always either LONG or SHORT, never in cash. Signals last an average of 3 days, and are always at least two days long.

Flow Zero

Flow Zero is a close relative of Flow, which spends about 15% of the time in CASH. Performance is about the same, making this signal a little better for the safety of being in cash some of the time. It's a very active trader, with an average signal lasting just under two days.

Simple Flow

While the original Flow signals used both index price and volume data, the volume component was found to contribute little to the performance of the stand-alone signal. The Simple Flow variant has volume removed from the equation, giving a signal that is very similar but not quite the same. Since its introduction however, it has generally performed notably worse than the original Flow.

Simple Flow Zero

This is a version of Simple Flow that is sometimes in cash, the same way Flow Zero is related to Flow.

Flow Past

While all other signals are computed at the market close for immedate investment (or as close to it as feasible), Flow Past was tuned for next-day performance: the signal is known a full day in advance, making it easy to enter orders at the next day's close (such as with mutual funds that are priced at market close). This was done at the cost of performance: returns are lower than for other signals, as well as being less smooth.

The Hemi Signal

Though it uses three of the four components of the Quick signals, Hemi combines them in a very different way, leading to a signal that is very conservative and spends most of its time on the sidelines. It's in CASH 75% of time, then 20% LONG and only 5% SHORT. While it has much lower returns than the other signals, these gains are very consistent. This is similar to the newer Composite signal, which has outdone Hemi as the low-risk leader.

Mixed Signals

You can also combine any of the above into a single composite signal. This is effectively splitting your money into equal piles and then investing each pile in a different signal, though it's presented as a single signal that may be partially invested: if you blend two different signals, you may get not only LONG, SHORT, and CASH, but also 50% LONG and 50% SHORT. These mixed signals tend to trade often, as any one component signal change will trigger a different level.