ES Journal - 2012

Status
Not open for further replies.
Quote from ammo:

max could you post your chart
More...

2012-02-17_1537.png
 
Quote from Max618:

I'll try to explain this point. If a market opens outside its value area (where 70% of the prior session’s volume traded) and then trades into value for two consecutive 30 minute periods, there is an 80% chance that the market will rotate all the way to the other side of value.
For example, lets say today's value area in the /ES is 1356.75 – 1345.50 and the market opens at 1360. During the course of the morning it trades lower and two 30 minute consecutive periods trade below 1356.75. There is now an 80% chance that the market will trade to 1345.50 before the end of the day but you need to know some tips:

–Neither 30 minute period has to close inside of value in order for the rule to be satisfied, just needs to trade inside it. If the first 30 minute period closes inside of value, then the rule is automatically satisfied as that implies that the second one will open inside of value. You need not wait for the second 30 minute period to close.
–The rule works both ways, whether the market is moving down from above the value area or up from below it.
–If the market opens up inside of value and then trades out of value, the rule applies the same way. If the market can trade back inside value for two consecutive 30 minute periods, then it has an 80% chance of rotating to the other side of value.
–Context is extremely important. Do not trade this rule mechanically and expect to have good results. Always judge the strength of any directional move in terms of market internals, overall
More...

interesting theory.. may i ask where did you get that stat?
 
Quote from gmst:

+1

I read a thread on this topic on traderslaboratory around a month ago. As far as I remember, the conclusion was having a very strict and a particular criteria did had a >70% probability however when you make the criteria very particular, the number of days per year fell to like less 30 or so - if I remember correctly. I didn't pursue it much and left it on a to do list for future work - as anything that trades 25 days in 252 days in a year is way too less for my taste.

On a similar note of having a percentage chance for price to be at a particular point during the trading day - there is a webpage available on google search that states that for eurusd, there is a 95% probability that price will trade within 5 pips of its yesterday pivot. Sounds interesting - right to construct a trading strategy around this statistic - however, at least I have not been able to build one yet - although I think it might be possible. one of the big reasons is distance between yesterday pivot and today's open might be too small - thus leading to a very small average trade profit (compared to commissions/slippage) - and the strategy will endure massive drawdowns from time to time. Your R:R in such a vanilla strategy would be massively inverted like 1:6 or even 1:10 in some cases. In my experience, its much better to develop more generic strategies like momentum based, mean reversion based rather than on such stats. Only exception I can think of are gap strategies - which work well and are based on a stat.
More...

Maybe you could also combine with volatility bands. I did a study a few years back and 80% volatility bands contain price 80% of time, and 70% bands contain 70% of time, etc.

For stocks, volatility bands can be calculated here:

http://pitrading.com/volatility_bands.htm

Fur futures, you would have to construct your own using ivolatility atm options data or any platforms iv data. Kevin Haggerty shows how to construct your own here:

http://www.tradingmarkets.com/.site/stocks/education/strategies/01042000-3274.cfm
 
Max:

Two observations:

1. What about double distribution days? Would the market just test the opposite VA extreme of the closest cluster - meaning we virtually have two separate VAs - ?

2. Days that present a very elongated profile are harder to retrace, so intuitevely I'd think that most of the 20% of days that fail to retrace the whole VA derive from this factor?
 
Quote from fseitun:

Max:

Two observations:

1. What about double distribution days? Would the market just test the opposite VA extreme of the closest cluster - meaning we virtually have two separate VAs - ?

2. Days that present a very elongated profile are harder to retrace, so intuitevely I'd think that most of the 20% of days that fail to retrace the whole VA derive from this factor?
More...

Yes...so that's the reason to find quickly what kind of day we are playing...
 
Quote from HurricaneUS:

Maybe you could also combine with volatility bands. I did a study a few years back and 80% volatility bands contain price 80% of time, and 70% bands contain 70% of time, etc.

For stocks, volatility bands can be calculated here:

http://pitrading.com/volatility_bands.htm

Fur futures, you would have to construct your own using ivolatility atm options data or any platforms iv data. Kevin Haggerty shows how to construct your own here:

http://www.tradingmarkets.com/.site/stocks/education/strategies/01042000-3274.cfm
More...

Thanks, I looked at it. I think looking at last 10 days range and last couple days low/highs are a much better way to anticipate today's range and the point where selling/buying momentum is going to stop rather than implied vols. It basically does the same job, just complexity is reduced.
 
Status
Not open for further replies.
Back
Top