Shifting Bias

Depending on your time frame when using the 5 assets of TLT,EEM,UUP,GLD,USO. (or if you are a minimalist and use TLT,GLD,SPY (and cash on the side), you may wish to try to "time" the market.
Picking exact tops and bottoms is too difficult to where I do not want to suggest it. Trading for "bottom picking" is possible; as is averaging in as an investment. The gameplan is much, much different for each one.

First, lets look at the strategy of legging into an investment using TLT,EEM,UUP,GLD,USO. The strategy works GREAT as just owning 20% of each of those 5 assets. However, if you believe you have an adge you may wish to increase and decrease according to bias. One such way is using the weekly chart's RSI What you can do is take 1 divided by the RSI for each etf on the weekly chart and add up the total. Divide each amount by the total to get your allocation. For example right now the RSI is
GLD 21.99. 1/21.99=.045475216
UUP 53.43 1/53.43=.018716077
TLT 31.93 1/31.93=.031318509
EEM 31.41 1/31.41=.031836995
USO 50.41 1/50.41=.019837334
Add them up and you get .14718413... take each of those numbers divided by this total and you get allocation of
30.89% GLD
12.72% UUP
21.28% TLT
21.63% EEM
13.48% USO.

I really don't like this plan on it's own because it adds to those in a downtrend. But it's not a terrible start. I would prefer only shifting into aggressively increasing allocation once the RSI dips below 40, and only shift at reducing (and not so aggressively as RSI goes above 60 or 70.

You could structure an individual bet as 40.01-64.99%=17.5%-22.5% investment. Above 40 is a bit backwards because as RSI increases it suggests a bull market so allocation can increase. Below 40 it is oversold and suggests the move is overplayed and is more likely to bounce.
35-40RSI=22.25-25% investment
30-35RSI=25-27.5%
25-30RSI=27.5-30%
20-25RSI=30%-32.5%
15-20RSI=32.5%-35%
10-15RSI=35-40% (from this point the allocation increases aggressively)
5-10RSI=40%-45%
0-5RSI=45-55%

Now, if we did this for multiple, we instead would be assigning a percentage, adding them up and taking the individual asset estimation divided by the sum.
For example the total of all 5 is 124
GLD 31/124=25%
UUP 20.1/124=16.2%
TLT 26.55/124=21.4%
EEM 26.75/124=21.6%
USO 19.6% 19.6/124=15.8%
This seems much more reasonable overall.

However, There are some other important indicators one may wish to consider. For example, an uptrend could add to a "score" a MACD crossover or histogram going positive or negative could trigger a "score" as well that applies to an individual assets correlation. The trick is understanding that with MULTIPLE assets comes the problem that if everything is oversold, you still have to apply more capital somewhere. Additionally, if there are commissions, you don't really want to adjust too much unless the adjustments are very significant from current levels.

Now you can construct mechanical systems like this, and that can provide your guidelines, but when it comes down to it, you may have better skill applying your own bias according to certain limits. You may not want to exceed a certain amount or allocate less than a certain amount per asset.


As a trader you can behave much more differently. What you can do as a trader is if an area is oversold, start a position either as soon as you also get an oversold on the 15m or 30m or 10m chart in addition to the daily and/or weekly chart. If you don't trade mid-day, you can wait for a candlestick closing ABOVE the previous candles high, and put a stop below the previous candle's low. You will probably incorrectly bottom pick a few times before you get it right, but if it is in a down trend, you miss most of the trend, and wait until the asset actually starts showing the V shape bottom on a intraday chart (will probably look like a hammer candlestick on a daily chart). As a trader, you might still use that mentality to acquire investment shares. You can establish a very long term position provided your low holds. On the other hand, you can also trade it for 2 or 3 times your downside risk, and if you really believe in the bottom, put on a 20% trailing stop after the stock has really rallied.

Another thing you can wait for, is an EQUAL high to be formed. Now you can either scale in at this point and if it starts to dip add, but then use the low as the stop. Or you can hold that position if it rallies from there and make the short term trade. The higher low follows the equal high and then a break of the equal high to higher highs tends to happen.

If you wait for these type of moments to establish these long term positions, or at least to add to them significantly, then you will possibly really boost this strategy.

Of course, you can also look for individual patterns to trade directionally in either direction with any of these assets. You might add to a position when a chart pattern sets up, and use it likea normal trade, but perhaps rather than sell all of the position, you leave a share or two behind. Then you repeat Basically the way you look at it as, you go in with 1k in the trade, you make 1.1k and you sell out your 1k and go about putting it to work again. When you lose, you may have to wait until you accumulate the capital.

You might also choose to use your technical skills to assign a "confidence" interval 0-100 in each asset class. 0-25 is very bearish 25-50 is neutral to slightly bearish 50-75 is neutral to slightly bullish and 75-100 is very bullish. Then add those up and find the sum and average of each will be your allocation. This requires that you actually have a good track record.

I personally like using long term charts. I look for price patterns, I look for support, I look at volume profits, I look at trends and weekly and monthly parabolic SAR. I look at MACD. And for oversold I look for bollinger bands, Slow Stochastics and RSI. Overbought only RSI concerns me. I look for divergences in the Slow Stoch and RSI as well. I also have a "triple oversold" indicator that works really well on longer term time-frames. Finally The McClellan Oscillator works great on a 6 month time horizon, and on a very short (few days) as well. If you don't want information overload, you can construct a spreadsheet and map out the suggested allocations 1 indicator at a time, and then produce an average or weighted average of them all together to produce a "suggestion" The slow stoch oversold triggers first, followed by RSI oversold and bollinger bands around the same time, then typically (unless uptrend is intact and you still get an oversold) you then get a crossover of momentum followed by a parabolic SAR break followed by positive relative strength (if no overbought is made) Typically you can ignore the initial overbought of the slow stochastics but after that the other indicators start to become more relevant.. Even though all of the oversold triggers happen first, I still might apply more capital to those that have BOTH oversold AND have just triggered parabolic SAR upside reversal. But it depends. Go back 10 years and is there any sort of advantageous signal? Such as support? Such as a monthly MACD crossover to the upside? Such as a 50day and 200 day moving average cross (or longer)

A Shifting Bias is about becoming exposed to one area over the other. You could also find assets that have greater correlation to the assets you re bullish on, giving you more exposure that way. When I show you a correlation matrix you can do some math and figure out which areas you have the most exposure to. For example with TLT everything is negative JJC(instead of GLD), UUP,USO,EEM. That means that in an equal weighted portfolio you are still biased against the movement of TLT. If you are bullish on bond principal (think interest rates are going lower). You may wish to find an asset that exposes you more towards it, such as AGG,SHY,PCY,BOND,CORP,etc. You can add up the correlations TO one specific asset to see that indirectly you have the most exposure in USO for example (with equal weight) because copper and EEM are both strongly correlated with it.

I have no problem with that bet, but at individual timeframes you might have more of a problem with it than others.