You must adapt to the local volatilities and have pockets deep enough to support those volatilities. You will trade in the direction of momentum to avoid being taken out on a stretcher. You live by the sword you die by the sword with RTM trading.
jnbadger, your thinking is correct. very good. I advise applying a trend momentum if your doing RTM. Doing RTM against the momentum is suicide unless you've got unlimited capital like the generals who just wait for the bounce to breakeven or better.
I do. Not willing to share though Took a lot of research, but the basic components are a Cauchy dist. and historical ATR. Personally I'd suggest looking into Fuzzy Logic if you have the time. Price % from the mean is useless IMO. Mike
Standard Deviation is used to determine where the average prices (or dataset) falls. When surveys are done for just about anything, 95% of the values fall within 2 Standard Deviations. When you start to go further out to + or -2.5 you have hit extremes and the prices will soon "return to the mean." When you start to hit the +2 or -2 Standard Deviations such as on Bollinger Bands there is a good chance the trend will start to pull back, but of course with investing this is not 100%. Another way to look at it is a Z-Score Indicator that plots the changes of standard deviation. The chart below shows reversal points.
I did not mark every +2 -2 so as to just highlight some trades, so I understand what you are saying about bias. The other cases where it reached these levels the market turned, it may not have been a win in each case, but the wins are large enough that when using a stop loss, referencing other indicators as on the chart, and overall market can provided successful returns.
bump! Important part is how we define the mean. There was some discussion about prev day POC level(market profile) or pivots... how to factor these in function of mean? some hints for the quant-scared kid please...
This is an interesting thread that has a lot of potential. In my experience, when trading correlated instruments, it is important to understand the nature of their relationship (oscillation vs. trend) when assigning a mean. If the instrument is a trender, 200 SMA, if it is a chopper, 20 SMA. This is overly simplistic and not at all "optimized" but it's a starting point. There are several ways that you can define trend vs. chop. I'm not talking about one-sided market chop vs. trend, I'm referring to a REAL correlated relationship. The actual nature of the pair is related to what is actually going on with the drivers of the relationship (product cycles, underlying commodity price change, liquidity advantages, etc.) I'm sure some more mathematically inclined can expand on a way to quantify this. Sometimes as traders we calculate too much and THINK too little. Often, simply trading/observing the pair is the most foolproof way of determining the nature of the correlated pair you are trading. Or just look at the chart... does it go back and fourth quite a bit, or does it move in long stretches without any retracement?
My personality prefers RTM, as it seems intuitively logical based on the ebb & flow patterns of the market. However, as of late, I've been bewildered at the market's ability to continue with a strong trend on a frequent basis. I wish there was a steadfast rule for RTM, such as +/- 3 std dev; but such a rule often will not apply to a runaway stock that trending up or trending down, perhaps due to a major news event. Walt
Many instruments change from trending to congested depending upon market conditions. How do you anticipate and adapt to it?