Silver Dragon and i wanted to discuss this topic and we both trade futures, so i have added this topic here because i have used the emini's as an example in our discussion. i also use the same analysis for currencies, metals, crude and natural gas.
this is what Silver Dragon said and i was interested in his thoughts on the subject.
I daytraded stocks for 10 years and switched over to futures 13 years ago.
I am a trend trader and analyze volatility, expected move and trends every Friday after the close. My analysis tends to end up with percentages, averages, probabilities and ranges.
Most traders have different definitions for volatility, expected move and trends, so I will try to define them as I go.
Volatility is the distance that price moves over some time period. For emini futures I tend to trade the first 2 hours after the equities open. So my volatility number is the price move over those 2 hours. The price move over those 2 hours has 2 components, trends and chop. I focus on the trends. My definition of a trend is a price move in one direction from the trend start to when price retraces against the trend more than my initial stoploss. I analyze those trends and calculate the average percent of trends that move more than 2 times my stoploss. Based on Friday’s analysis, during that time period, for the emini, 38% of them moved more than 2 times my stoploss. For those 38% of trends, i calculated an average move of 3.1 to 4.2 times my stoploss.
During the trading day when I add support/resistance and market context, I can raise that ratio from 38% to over 50% for the actual trades that I take and achieve more than a 2:1 profit/loss ratio. The 38%/50% and the average move is used to setup my charts for trading. They are also used for trade management, determining the initial stoploss and contract sizing.
Silver Dragon, How do you define volatility, expected move and reversion to the mean and how do you analyze and use them in your trading.
Looking forward to our discussion
toucan
The following 8 users say Thank You to toucan94506bm for this post:
Let me start by saying you are head and shoulders of where I am at. Salud!
I am very visual trader. I keep and eye on various chart intervals and keep and eye on standard deviation and wait for price to reach 1 to 2 std deviations. Then I buy or sell accordingly. The idea being that price hit its expected move and will revert back to its mean. My favorite chart setup is the 1 hour chart at a 3 week time frame.
Once I place the trade I wait for price to revert back to the channel and sell where congestion is or some other support/resistance level occurs. Sometimes I wait for price to move back to the opposite std dev channel line. Below is a trade I took last week. The target was hit within a few hours of placing the trade.
I use volatility to determine if my expectations of the price movement are in in line with what I see on the chart.
What I like about playing the std dev channels is that if I am wrong I am more willing to hold because more than likely it will revert back to my entry on the second or third std dev channel.
I would be interested to see your Friday analysis for any given product.
Robert
Small 10year
nosce te ipsum
You make your own opportunities in life.
The following 4 users say Thank You to Silver Dragon for this post:
Here is a trade I entered tonight. Price made it to the first upper std deviation line. Price has been declining over the last few days. This is a high probability trade. Taking the short position. Defined entry and exit.
Robert
nosce te ipsum
You make your own opportunities in life.
The following 2 users say Thank You to Silver Dragon for this post:
thanks for the info and the chart. for the chart... are those green/blue lines longer term standard deviation lines? i assume red is zero deviation line. not sure what you mean by using volatility to determine if expectations are in line with your chart. thanks
today is FOMC day and someone must have leaked information because starting at the equities open all my currency charts moved down at the same time(i always look at currencies for moves at the this time) ... i trade futures us dollar vs euro, pound, yen, swiss franc, australian dollar, new zealand dollar, canadian dollar and peso .... this doesn't happen very often, but when it does, it can make for a great day trading. i was able to short yen swiss franc, pound and australian dollar. attached is my chart for australian dollar. its messy, but everything on the chart is used for my daytrading analysis of market context and setup identification. after i noticed all currencies dropping, i shorted as many currencies as i could. for the australian dollar trade, when price started moving back down, i shorted thinking that if it ran below the red line low of yesterday, it would run for a while (ie no congestion).
if you look at the right side of the chart you can see the results of my weekly analysis and real time info for today. the green numbers say that there is an average of 5.1 trends from 5am to 8am pst, and 2.2 of those trends average greater than 2 stoploss. the black numbers below that say that for those 2.2 trends greater than 2 stoploss, they average 2.8 - 3.9 stoploss. they are averages and the trend that i traded today ran 4.5 times my stoploss versus the average range of 2.8-3.9. the cyan numbers below that indicate average volatility for each time period in stoploss terms. for example, i trade currencies from 5am to 8am and the average move is 3.5 times my stoploss, the number right after that is a running number of the volatility happening in real time today so it measures the high-low from 5am to 8am. the other cyan numbers are for 12am-5am and 8am to 11am.
The following 2 users say Thank You to toucan94506bm for this post:
interesting topic, does any of you look at option prices to have a feel for the expected move of the futures? I am not an expert on options but I know that options traders use this concept of expected moves all the time. I have always planned to look into it more deeply.
The following user says Thank You to SBtrader82 for this post:
RTY trade today... Price had been below the low of yesterday for most of this session and then a new trend up started with the long setup signal about 9:40 on my chart. you can see my analysis on the right hand side of the chart.
The following 2 users say Thank You to toucan94506bm for this post:
Yes, if I am trading for longer term trades I get the expected move and volatility from the closest option to expiration. It confirms if my expectations of where I think it will move and the target are realistic.
If the volatility is high then I have to take into consideration volatility contraction. So if the expected move is 20 dollars 3 weeks out with a volatility ranking of 50, I may want to set my target to be half of expected move in the event volatility contracts and the daily moves and overall expected move become smaller.
Robert
nosce te ipsum
You make your own opportunities in life.
The following user says Thank You to Silver Dragon for this post:
You were looking forward to this Friday’s analysis. Here is a description of that analysis. I perform this analysis on futures contracts for currencies, crude, natural gas, metals and indexes.
My definition of a trend is a price move in one direction from the trend start to when price retraces against the trend more than my initial stoploss.
The first step in my analysis is to evaluate what the best initial stoploss is for the next trading week. if the stoploss is too small then the trends tend to be very short and the risk of stopping out is high. If the stoploss is too large then the trends tend to be much longer, but the risk of holding through retraces is very high. However, there is a sweet spot for an initial stoploss somewhere in between these two extremes.
I start with the current initial stoploss and adjust it up/down to find the best fit for the number of trends larger than 2 times my stoploss. this is the first line of data shown on the upper right hand side of all my charts. Ideally, I am looking for fewer trends that move more than 2*stoploss and a ratio of (trends>=2*stoploss) / (total trends) that is > 40%. You can see this on the RTY chart in my post above where the total trend average is 13.2, the average (trends>2*stoploss) is 5.8 with a ratio of 44%. That means if I traded every trend at the start and got out of every trend at the end, I would be profitable 44% of the time. This analysis determines the initial stoploss that I will use for the next week.
Once I have the best fit for trends and initial stoploss, I then look at trade risk sizing. Some people can live with a $100 stoploss and others may be happy with a $1,000 stoploss. My current risk per trade is $300-$500. On the RTY chart, below the green trend analysis numbers, you can see next line of black numbers that represent the trend length range for (trends>= 2*stoploss). for this week, that range is 3.1-4.2 times my initial stoploss. So I know that I have a 44% chance of getting into a good trade with a profit/loss ratio of 3.1-4.2. that is if i got in at the start of the trend and got out and the end of the trend.
When trading in real time, I add market context and support/resistance to improve the win rate from 44% to above 50%. I usually get in after the trend starts and get out before the trend ends, so my profit/loss ratio is probably greater than 2:1 and not the 3.1-4.2 if I had entered when the trend started and exited when the trade ended.
My trend analysis above produces a stoploss for 1 contract. then I adjust the number of contracts so my initial risk is somewhere in my risk range. This usually ends up being 1-10 contracts. If the risk of 1 contract is over $500 then I will not trade that future. If it is less than $300, I will add contracts so that I end up in the $300-500 range.
The following 2 users say Thank You to toucan94506bm for this post:
@Silver Dragon
It´s interesting that you mentioned the std. deviation channels. Not many chart softwares have them included.
I tried them for future trading. I know tos has them and tradingview (member created indicator). May i ask which software you are using
for them?
The following user says Thank You to Antares66 for this post:
Just curious since I see sloping std channel in charts posted here, generally its from open or some sort of pivot point from where I'm measuring it. Resulting in horizontal lines and not sloped ones.
I´m using Motivewave. MW has regression channels with standard deviations. I set it up to use 1000 bars backwards to create the regression channel.
It´s really amazing, not only for futures but also for stocks (5min, 1 min for daytrading).
The most programms i know don´t offer auto standard deviations channels. I know TOS, Motivewave and now tasty works.
Tasty works was created by the guys who also created TOS in the past.
The following user says Thank You to Antares66 for this post:
That´s great. Did you publish your code on tradingview? I found something similar and use it with tradingview.
But i am glad MW has such a tool on board because i can´t trade with tradingview because it doesn´t connect to IB.
MW does it. I put all regression lines on a chart with different colors and created a layout.
I set up 100 bars to calculate the regression lines on the one minute chart and 999 on the 5 minute but still trying to find the best setups.
Do you have a recommendation for the bar lenght to calculate the regression lines?
Some only trade with order flow but i still prefer the standard charting with regression.
I've checked inbuilt regression channel code on TV, it seems to be doing same, its regression line +/- 2 stdev channel, settings are adjustable. So I'm saved on effort of coding one from scratch. If you are interested you can check that one also, if you want to modify how stdev is calculated in it, you can simply swap that part of the code. However, I don't see any need for it,
I've not used regression channel enough to comment on recommendation but this is going to be highly subjective. I feel 100 is fairly decent length for it so far.
I've been looking through TV and have found many Regression Channel options. Which one was the one that you found most closely resembled the one Silver Dragon is using on his chart? The built-in option only seems to give one line for STD on the upper and lower, however that line can be adjusted to which STD you want.
I was looking for one that has multiple lines, one for 2nd STD, another for 3rd STD etc.
15.874 is just the square root of time. 252 trading days in a year will give the expected daily move. (Some peeps adjust here, 16 is commonly used for calcs.)
Wow!!! @WoodyFox thanks a lot!! I have been looking for something like this for months. I once heard Tom Sosnoff mention the "expected move" but I could not figure out how he was computing it. This is extremely interesting!!
I don't know how to thank you enough, also I really appreciate that you find my posts interesting.
If you have any other resources from "the option world" please share them. I would really like to incorporate implied vol. considerations while trading futures.
The following 4 users say Thank You to SBtrader82 for this post:
I have been having the weekly expected move up when trading /ES. It can be a strong support/resistance with confluence. Did not look too deep into other application of IV though. The previous replies are interesting to read. Thank you.
I have been studying this since early last year. From my observations I am confident that levels based on projected ranges have some significance because when we cross those levels you tend to see explosions of volume stopping out. However, the connection between implied volatility and realized volatility is complicated as you can see from the chart. Here I took the expected range of the day plotted against the realized range of the day.
Obviously what point you use for the price of the instrument and the price of your volatility index matters, but my testing did not show statistical significance. This chart was based on using the NYSE opening prices.
The following 3 users say Thank You to TWDsje for this post:
Really interesting, thanks for sharing. How were you testing for statistical significance?
I've been inactive here (and trading) for a long time but am considering taking on a big analysis project. Implied volatility and to what extent it can be used as a tool to gauge targets was one of the things that I wanted to look at. Coincidentally have also just read through the thread on orderflow and MBO data you posted in last year as that is probably the other key thing I wanted to explore.
The following user says Thank You to NoSquigglyLines for this post:
Many strategies can be built on this very premise.
Some traders think this is nothing more than random line theory, but they simple do not get it.
These are high areas of importance statistically. After all, if you know how often they are broke, wouldn't you place a bet.
The connection between implied and realized in nothing more than probability. (Although this may be hard for some to understand and only complicated in that way).
68.2% of the time your range will be within the expected based on one standard deviation.
Use this and a directional filter...Boom you have the basis of a statistical strategy. A place were edges are found. LOL
There is no reason to reinvent the wheel here. JMHO
As for your price point (Control), I would use the prior day close of both. But that's me.
The following user says Thank You to WoodyFox for this post:
Well it's not a random line. It's the price options are predicting as likely ranges for the day. Whether we do or don't get to that point matter significantly to traders with options positions on.
The connection between implied and realized volatility is a lot more than just probability. That's a gross oversimplification that misses some of the greater insights we get from this data.
Implied volatility and realized volatility aren't very well correlated. Otherwise we would get more of a diagonal line across the chart. Realized volatility tends to underperform implied volatility which is why they cluster on the lower half of the chart. When realized volatility does overperform it can do so substantially creating the randomly scattered points at the top of the chart. However there is also a lower bound meaning that based on implied volatility there is at least some volatility that you can expect creating the empty space below the lines I've drawn on the chart. Periods of low volatility are more common than high volatility creating the huge cluster of points in the bottom left. Most of those points were from 2019 in this image. This year we've had more volatility than normal meaning that if you were relying on that line holding for your trades in 2020 you probably had a pretty bad year.
These are all things that if you're going to trade with volatility should be taken into consideration. Otherwise you'll end up with something that probably makes money overall, but suffers very painful drawdowns and inconsistency. Or rather your risk adjusted returns won't be very good.
Hence will happen 68.2 percent of the time or less. (Looking at one Standard Deviation)
2019 and 2020 are not the same year but they are actually statistically no different.
You can get your straight line.
To do this make your years relevant in terms of a smaller integral.
Look here:
A chart of VXN from Jan. 1 2019 till now. Clearly looks different, but comparing volatility from contract to contract you can see by the bottom plot you get periods of low volatility in 2019 and 2020.
You can do this by comparing Volatility over short periods of time (using IV Percentile, which is used daily by options traders all over the world. Tasty Trade has lots of info on how to use it)
There several other ways to do this...keep that in mind. Just thought I would use something Option traders like.
Here I use 52 Days.
So anything under 50 percentile I would call low volatility relevant to the last 52 days no matter which year you are in.
Capture 10
Also, a high Volatility number just changes the range of the bell curve, it does not change its probabilities.
Capture 11
So just bigger 2020 ranges (implied and realized) with the same statistics behind the connection between them.
You wouldn't be able to compare 2019 to 2020 independently the way you are doing it.
Yes, and the other 31.8% of the time? And what if those all happen in a row? It's more than just the percentage, and you're doing anybody that reads this a disservice by ignoring all nuance. That's how we end up with things like February 2018.
I am talking purely about the statistical properties of implied volatility vs realized volatility. That is the number of days where actual range > expected range is higher than the past. Hence why you see all these retail traders hitting it big with OTM options. I guess I shouldn't have phrased as "we've had more volatility".
But this is why I plotted it on a scatter plot. If you just look at a distribution histogram you don't get the whole story.
Over time will the realized range fall within the expected range, based of implied volatility, 68.2% of the time or less?
If your answer is yes. Good.
If you answer no and can prove it, I will personally enter you for the David Hilbert Award. LOL just sayin.
Past performance is not indicative of future results. Without a mathematical proof to show why that percentage is constant we should always assume that such stats can change.
But there are periods where it is less than that, and it's up to the trader to determine for how long that can go wrong before they can't trade that strategy.