Identifying Sharp Price Moves
- Thread starter dima777
- Start date
For am stock gaps you can use STOCH.
Do the buy in the late PM the prior day.
Use the fast line on the STOCH and use a very short STOCH default.
Do the buy in the late PM the prior day.
Use the fast line on the STOCH and use a very short STOCH default.
I use a standard deviation of closing prices for 13 periods, which I update daily after the close and then exponentially average.
I prefer not to but you can multiply the square root of time to the standard deviation before averaging. Once you have a number make a percentage out of it and multiply it by the market price. You have a target.
At the beginning of a correction or at the end of one, you can target a key quant mechanical sell off before it happens. It's usually the second to the last pullback, where smart money exits orderly on indication of being overbought or oversold. Scalp the rest of the day, possibly into the next day and do it again.
I use daily bars to cast, to include wick extremities if using candlesticks. I hate bar charts, they cross my eyes.
It's nothing new and all quants exploit it, so, join in if you can craft the numbers; if not, shoot me a pm and I'll answer your questions.
I prefer not to but you can multiply the square root of time to the standard deviation before averaging. Once you have a number make a percentage out of it and multiply it by the market price. You have a target.
At the beginning of a correction or at the end of one, you can target a key quant mechanical sell off before it happens. It's usually the second to the last pullback, where smart money exits orderly on indication of being overbought or oversold. Scalp the rest of the day, possibly into the next day and do it again.
I use daily bars to cast, to include wick extremities if using candlesticks. I hate bar charts, they cross my eyes.
It's nothing new and all quants exploit it, so, join in if you can craft the numbers; if not, shoot me a pm and I'll answer your questions.
price action
Set your bollinger bands at 1,2, 3, 4 standard deviations, with a 20 period simple moving average. When the bands start to widen, you can see an expansion of volatility. In a good trend day, you will see price action bounce back and forth between 1 and 2, take profits when it hits 3.
Also, use it on a 3 min, 30 min, and 60 min to get the overall picture.
Also, use it on a 3 min, 30 min, and 60 min to get the overall picture.
I left the door open too wide there. 
Subsequently as a result to save some time and effort I'll just go ahead and address everyone interested at once.
It sounds much more complicated than it is.
A standard deviation is simply the square root of a variance; and a variance is an average squared deviation from the mean. The mean is a simple, static average; add up all values sampled, then divide by the number sampled. The variance is found by subtracting the mean average from each individual sample (resulting sometimes in negative numbers), squaring each new value, adding all values together and dividing by the sampled degrees of freedom; which means one less than the population sampled. The reason for this is factoring infinite and dynamic room for variance to occur out of static and finite points of reference. The variance however doesn't have cohesion to the original data. Finding the square root of the variance takes care of this problem. There are however limits to this data coupling in practice when trying to forecast into the future how volatility will change, but someone more versed in implied volatility could if they wished expound. I only have a general knowledge of it.
But for measuring current moves based on past volatility (not historical volatility, that's found by finding the square root of time, and multiplying it to the standard deviation), this use of standard deviation will suffice.
No matter which used, past or historical volatility, you have to exponentially weight average the standard deviation to constantly reflect current fluctuations in volatility.
To do that first you need to inject a value to represent the sample. After deciding on the sample period you exponentially average the sample (finite). This forms the basis of the EMA. But let's backtrack. What's needed is a value we can weight to current prices or volatility, while by infinite degrees removing less weight from the oldest value. Those removed values once found in the moving average however always contribute to the current data set.
To find the exponent: you divide 2 by the period sampled, resulting in a negative number, and then round that number to the nearest hundreth.
The simplified equation is as follows:
Today's Exponential Moving Average=(current day's closing price (or whatever) x Exponent) + (previous day's EMA x (1-Exponent)).
I told some of you last night I'd give you my new target but I honestly can't tell which way this consolidation period will ultimately break to. If it breaks to the upside again, then a new volatility period begins at 4% still: 1312.48; then again it may be 1238.37.
That's the art part. Use it at your discretion.
Enjoy!
Subsequently as a result to save some time and effort I'll just go ahead and address everyone interested at once.It sounds much more complicated than it is.
A standard deviation is simply the square root of a variance; and a variance is an average squared deviation from the mean. The mean is a simple, static average; add up all values sampled, then divide by the number sampled. The variance is found by subtracting the mean average from each individual sample (resulting sometimes in negative numbers), squaring each new value, adding all values together and dividing by the sampled degrees of freedom; which means one less than the population sampled. The reason for this is factoring infinite and dynamic room for variance to occur out of static and finite points of reference. The variance however doesn't have cohesion to the original data. Finding the square root of the variance takes care of this problem. There are however limits to this data coupling in practice when trying to forecast into the future how volatility will change, but someone more versed in implied volatility could if they wished expound. I only have a general knowledge of it.
But for measuring current moves based on past volatility (not historical volatility, that's found by finding the square root of time, and multiplying it to the standard deviation), this use of standard deviation will suffice.
No matter which used, past or historical volatility, you have to exponentially weight average the standard deviation to constantly reflect current fluctuations in volatility.
To do that first you need to inject a value to represent the sample. After deciding on the sample period you exponentially average the sample (finite). This forms the basis of the EMA. But let's backtrack. What's needed is a value we can weight to current prices or volatility, while by infinite degrees removing less weight from the oldest value. Those removed values once found in the moving average however always contribute to the current data set.
To find the exponent: you divide 2 by the period sampled, resulting in a negative number, and then round that number to the nearest hundreth.
The simplified equation is as follows:
Today's Exponential Moving Average=(current day's closing price (or whatever) x Exponent) + (previous day's EMA x (1-Exponent)).
I told some of you last night I'd give you my new target but I honestly can't tell which way this consolidation period will ultimately break to. If it breaks to the upside again, then a new volatility period begins at 4% still: 1312.48; then again it may be 1238.37.
That's the art part. Use it at your discretion.
Enjoy!
For front running stock gaps (which I take to be unusually sharp thrusts), the tell is seen way below a 13 type length of measure and things like sig deviations.
Second you change the point the signal is generated.
Recently an example of NTRS doing a run up for 2 to 3 day position trading was posted.
A STOCH (5, 2, 3) is used and the signal is the fast line crossing the 50% in the direction you wish to trade. Use a 30 min graph and pick off the entry during the three common activities that end each day.
In commodities, for trading ahead of the quants, you need leading signals of price. I front run on ES simply because of the markets great capacity.
Several indicators are required. And the unusually sharp thrusts are caused by the high risk and poor trading strategies combos.
The 3BR is a good example around BO's and FBO's of how things go wrong.
Sometimes these unusually sharp thrusts are linked, especially in three's. Long ago it was possible to take 3 300 point thrusts on the large DJXX. (the good old days before quants and the e- mini's drew interest in chicken feed trading). So you shift to the e minis and front run as mentioned above.
The issues revolve around "harmonics" trading. Spikes are odd and square waves have the vertical walls.
I use 27FEB as one of the better examples(in jem's terms it was a 100,000 dollar day for retail amateurs).
So what makes for unusually sharpe thrusts is people getting caught off guard and frozen for the duration until their broker lets them blow out. the best ones for this are when the series of events goes past single digit traders to double digit traders. That was what made the 27FEB soo enjoyable.
In one of these trades it is the most fun to watch the T&S and see those opposite you on the trade getting blown out. In trading, there is an IB overriding going on as blow out is happening.
I have a camtasia for the 27th and an audio standalone as well. It is helpful since you see tha display and you get to hear where to look for what is next (we front run).
I am speaking about much higher sigs than were mentioned by others.
Good Q.