Algorithmic Trading Strategy: Strong (Signal) In, Weak (Signal) Out (SIWO)

Algorithmic Trading Strategy

Algorithmic Trading Strategy

This paper tackles the problem of making the correct portfolio allocation in a systematic manner. This allocation deals with the dilemma of having you funds invested or many investment opportunities at once, and deciding which one to make. Because the I Know First signal is an absolute figure with equal weight to any asset it is produced for (commodity, stocks, indices, currencies etc.) this strategy can be applied in the exact same manner to every portfolio. The idea is to always take the strongest buy signal into your portfolio, and throw the weakest one out. In a previous case study a strategy for identifying good buy and sell signal was created, this case builds on top of it.

Because the algorithms produce daily signals, it represents a bit of an update of the remaining money in the pool to earn. Biggers signals are thus less risky than lower ones. For this reason we exit a position when the signal goes below 10, the implied profit is just not enough to outweigh the implied loss.  This article talks about another systematic rule in which you could exit an active position, or maybe not enter one at all.

The signal’s properties

In this example let us imagine a fresh portfolio. You have $10,000 of which none has been invested and you begin tracking signals to find optimal trading points. Your rule is simple, if the signal is above 32 and the closing price is above the 5 day average buy. Similarly, if the signal is below -32 and the closing price is below the 5 day average sell short. Because the signal is an absolute number both the long and short positions are of the EXACT same weight. Following this logic, a signal of -50 is more powerful than a signal of 40. For this paper we will invest $2,500 on every trade, never more, never less.

Case 1: Too many investments

In this example you are faced with the dilemma of having too many investment opportunities. Your fund currently holds $10,000 and you are faced with 7 buy opportunities. Your entry rule is as simple as it gets, just pick the strongest 4 signals and invest in their corresponding assets (assuming no trading cost and your investment size being $2,500). Let us assume that on this day before market opening the moving averages are all aligned and you are faced with the following signals:

Algorithmic Trading Strategy

Algorithmic Trading Strategy

 

First we line them in chronological order:

line2

Because you can only invest in four assets you would pick the strongest four signals, thus you would buy

line3

Case 2: Signals below 10

The next morning no new investment opportunities arise; however, the signals on your active investments surely changed. After looking at your updated table you pick out the new updated signals.  Let us assume all the investments were accurate, and all prices went up for longs and down for shorts (thus the signals decreased in strength)

APOL (-8) GGP (-55) RAX (-100) GEL (120)

As can be seen above the first signal APOL (-8) has drastically gone down from the previous signal APOL (-66). Because -8 is within our exit range (-10<x>10) we would exit with a profit of $500 for a total of $3,000. We are now holding 3 investments (Valued at $7500+) and $3000 in cash to be re-invested.

Case 3: Strong (Signal) In, Weak (Signal) Out (SIWO)

Now that the fundamentals are clear SIWO should be very self-understandable. Following case 2 signals you are faced with the following signals from your active investments the next morning.

GGP (-3) RAX (-25) GEL (55) and (cash = $3,000)

You are also faced with 5 new investment opportunities.

XOMA (33) AAPL (-42) GOOG (44) TSLA (-55) IBM (320)

The question is now which investments you keep and which sell considering you are not able to hold all the assets at once. The first step would be to exit our GGP (-3) signal as it is within the exit rule range. Assume it was valued at $4,500. The new portfolio is as follows.

RAX (-25) GEL (55) and (cash = $7,500)

Because our cash base is $7,500 it allows us to first of all invest in 3/5 new opportunities. Thus you would purchase the underlying assets of the three strongest signals GOOG (44) TSLA (-55) IBM (320). Your portfolio now includes the following investments to the left, with the remaining possibilities to the right.

Holding RAX (-25) GOOG (44) GEL (55) TSLA (-55) IBM (320) and (cash = $0) | Open decision XOMA (33) AAPL (-42)

We will now re-organize the two sides according to the remaining assets’ strongest signal, in this case (-42). On the left side any stronger signal than that (in absolute value) will be erased, thus we are left with:

Holding RAX (-25) | Open decision XOMA (33) AAPL (-42)

Because our open decisions currently have a stronger signal than our current position we will replace it with the strongest signal, in this case AAPL (-42).

Sell RAX (-25) = $4,500

Buy AAPL (-42) = $2,500

If the RAX (-25) signal would have sold for over $5,000 you could have even bought both new investments at $2,500 each; however, due to our rule we decide to be left with the cash. Thus our final portfolio looks as such:

AAPL (-42) GOOG (44) GEL (55) TSLA (-55) IBM (320) and (cash =$2,000)

Assuming each investment is currently valued at a minimum of $2,500 the total portfolio is sitting around $14,500. Once your portfolio value doubles to $20,000 you can adjust the investment size to $5,000, and so you will always keep a tight portfolio of just the strongest signals.

Conclusion

This model has been extremely simplified. The original base could be anywhere from 5%-50%, and by no means is optimal at 25% as in this example. Also the returns are extremely high for the sake of the example and are not correlated with actual market earnings expectations. Finally the algorithm is unlikely to find 7 opportunities on one day (although this is not out of the question). This system however can help any investor systematically allocate funds within his portfolio, making it a very profitable systematic strategy of always holding a few very high expected yield and low risk assets. SIWO guarantees you will always maintain the absolute optimum expected value strategy according to whatever investment parameters you set when your decision making (buy/sell) variable is the I Know First Signal.

I Know First Research is the analytic branch of I Know First, a financial start-up company that specializes in quantitatively predicting the stock market. This model was designed Daniel Hai. We did not receive compensation for this article, and we have no business relationship with any company whose stock is mentioned in this article.