5 years and still no strategy | Page 63 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

WRB/MC strategy:Coder needed to semi-automate my strategy 11 replies

strategy builder/tester for new scalping strategy 1 reply

Is opposite of losing strategy a winning strategy? 21 replies

After years in trading, do you still believe in get-rich-fast BS? 98 replies

Fozzy - still using after all these years 2 replies

Attachments: 5 years and still no strategy
Exit Attachments

5 years and still no strategy

  • #1,241
  • May 10, 2017 11:03am May 10, 2017 11:03am
Quoting Copernicus
{quote} That's the thing.....where I feel the opportunity lies is in the possible small amount of non-random behaviour that in general is overlooked by a trader seeking the thrills. If there is a small presence of non-random price movement in an otherwise random price series, then like a casino with a small house edge, you can capitalise on this feature by busting through the randomness with a large sample size. For example, if I can find an array of long term return streams with a small edge (say 1-2% each on average over say the last 16 years...
The small presence of non-random (price) movement is a characteristic of ALL random movement.
There is no need to look for them to have a positive edge. (yes positive edge)
Namely because:
1) they are unpredictable (eventhough they do exist)
2) their pattern change over time.
  • #1,242
  • Edited 11:26am May 10, 2017 11:15am | Edited 11:26am
Quoting Mingary
{quote} The small presence of non-random (price) movement is a characteristic of ALL random movement. There is no need to look for them to have a positive edge.
That's where I believe we differ in opinion. I almost sit in your camp as in my opinion the FX market in particular is incredibly efficient but in terms of statistical randomness it is the very small bias that is present in a price series that is sufficient to make this game profitable through the application of simple principles of either momentum or mean reversion over a large sample size. While you might write off traders (i have a feeling your angst is against TA), it's harder to write off investors as both trader and investor are actually eating from the same trough albeit at different timeframes and with different time horizons in mind.
  • #1,243
  • May 10, 2017 11:32am May 10, 2017 11:32am
Quoting Copernicus
{quote} That's where I believe we differ in opinion. I almost sit in your camp but it in terms of statistical randomness it is the very small bias that is present in a price series that is sufficient to make this game profitable through the application of simple principles of either momentum or mean reversion. While you might write off traders (i have a feeling your angst is against TA), it's harder to write off investors as both trader and investor are actually eating from the same trough albeit at different timeframes and with different time horizons...
I agree overall.
I think it's not necessary to look for or develop tools to find the very small bias, simply because it's built in !.
All you would need is some algorithm that would match the position sizing and equity to the expected bias.
  • #1,244
  • May 10, 2017 11:37am May 10, 2017 11:37am
Quoting Mingary
{quote} I agree overall. I think it's not necessary to look for or develop tools to find the very small bias, simply because it's built in !. All you would need is some algorithm that would match the position sizing and equity to the expected bias.
Good point. I see where you are coming from. :-)
  • #1,245
  • Edited May 11, 2017 7:05am May 10, 2017 12:42pm | Edited May 11, 2017 7:05am
1
  • #1,246
  • Edited 2:38pm May 10, 2017 12:57pm | Edited 2:38pm
Quoting freakout
{quote} what I have learned? one very important thing - everyone who loses time on writing posts on the forums doesn't make money sad but true and I've seen this many times - the same nonsense is repeated everywhere and only a few posts in this thread were worth reading so the best what I can do is trying, trying and once again trying - nothing more this is my last post in this thread so thanks for all for participating in this discussion happy trading! cheers, freakout
Good thread you started. We got people arguing whether markets are random, or manipulated, or not predictable. In all 3 mindsets, these people are gamblers, nothing more or less and eventually will have to refill their accounts from monies earned from other sources.

And not to leave out the market is rigged and can't be beat posters. It dumbfounds me why would someone hang around in a trading forum posting the game can't be beat. What is their purpose here, boredom relief? That makes as much sense as an atheist going to religious worship. WTF?

The markets are sometimes predictable to those who have taken time to study them. But first, you have to get the right mindset. There are some in here that know how to trade and it's easy to spot. They don't know what will happen next and they have no pride.

Best of luck on your journey.
Ecclesiastes 1:9
  • #1,247
  • May 10, 2017 1:35pm May 10, 2017 1:35pm
Serial Correlation........
The lack of day-to-day serial correlation in auction markets is often the starting point for random market theory. Studies as early as Labys and Granger (1970) confirm the day-to-day randomness. Even in trending markets, the expectation of up price or down price for the next dayis still about 50-50. Even odds in non-trending markets is easy to understand; in trending ones it is not so obvious. A market that is trending up or down does have an enhanced probability in the trend direction; it is just pretty well masked by the random component that is always present. For example, a trend of one cent per day (on average) is virtually lost in the average daily trading range of, say, ten cents. By grouping, however, the random component can be averaged out and the underlying price movement can be seen. We will illustrate this point with a trend in soybeans.

Comparing one day's data to the previous days data (or comparing one candlestick to the next candlestick, or comparing one price to the next price) following the market in this way, are all examples of "linear measures".

Now if you follow the data, linearly, "day to day" in sequential/chronological order, the market appears to move randomly. There is no "day to day" evidence of a underlying trend.

This is where the random theory conspiracist's see this as some type of evidence that the markets are a random walk. When you look at the data in a linear, chronological , sequential, series (time series) it only appears to be random What this proves is that even in a trend there is no "day to day" serial correlation. Even in trending markets, the expectation of up price or down price for the next day is still about 50-50
Attached Image


Rather than "linearly" follow the data price to price. Lets start measuring the same exact data, using "non linear" measures used in auction market analysis. Here we will use a 2 day sample of data. Rather than following the data "day to day" in sequential order. The results look just as random as following the market day to day. There is still no evidence of the underlying downtrend Even with a 2 day sample of data.
Attached Image


Let's look at a 3 day sample of data......... Now things start to get interesting..Same exact data.. still using "non linear" measures.Now you start seeing some "evidence" of the underlying downtrend. By using a 3 day sample of data you began to filter out the "day to day" random fluctuations ("random component")
Attached Image


It also becomes more apparent with the 4 & 5 day sample of data.
Attached Image


So what does this prove?.......... do the math
Attached Image


Testing a Trend for Serial Correlation
There is one long downward trend from March 6, 1991, through July 10, 1991. This run covered 88 trading days (127 calendar days) or just over four months. The total price drop is 121 cents or 1.375 cents per trading day. The daily trading-range average over the 88 trading days is 8.17 cents, ranging between a high of 29.25 and a low of 3.25 cents. Table 1-5 has the prices and a list of higher/lower prices on the basis of 1 day, 2 days,. . . through 5 days.

Since a downtrend was Pre-selected, we have inserted a bias to the downside.
Therefore ties those cases where the compared prices were the same-will be awarded to the H, or higher count. That results in Table 1-6

One of the basic fundamentals in AMVT is an understanding that the markets are not linear. Using linear measures to describe & analyze a non linear market is like using a tape measure to measure how many gallons it takes to fill a bucket

If you group your data, group prices over the correct sample of time (price over time) rather than look at the data "linearly", day to day through a series of candlesticks, you see the markets are in no way a random walk.

"A market that is trending up or down does have an enhanced probability in the trend direction"

To test that statement in a downtrend using the data starting with the 1 day sample of data.
If you traded in the direction of the previous day's trend (following price) with the probability of trading with the underlying trend, you would have correctly traded in the direction of the underlying trend 52% of the time and you would have been wrong 48% of the time. Correctly trading in the direction of the underlying trend following price "day to day" is still pretty much a coin toss.

Let's look at the 2 day sample.had you traded in the direction of the underlying trend following the 2 day samples of data rather than following price, you would have correctly traded in the direction of the underlying trend 62% of the time and you would have been wrong 38% of the time.

Now things get interesting..

Let's look at the 3 day sample.had you traded in the direction of the underlying trend using the 3 day samples of data rather than following price, You would have correctly traded in the direction of the underlying trend, 70% of the time and wrong 30% of the time.

4 day sample you would have been correct 73% of the time and wrong 27% of the time.

5 day sample you would have been correct 76% of the time and wrong 24% of the time.etc.

In AMVT the used of TA is a non permitted practice, not because the methods are wrong..it's data validity of "price to price" (timeseries) data. is your data fit for the task at hand..are you using the correct data for the questions being asked.

Auction markets are not linear, you can not use linear measures to describe a "non linear environment"

This is what you call "proof of principle"...
Attached Image
Markets are not efficient, rather they are effective - Jones
  • #1,248
  • May 10, 2017 1:56pm May 10, 2017 1:56pm
Wow, very impressive!! I think your statistical prowess is unparalleled. As for me, the method I use is not as sophisticated.
I see the bulk of price action going north and figure I should trade north too, lol
  • #1,249
  • May 10, 2017 2:59pm May 10, 2017 2:59pm
How does one battle the 50/50% chance?
Say 50% price has chances of going 100 pips on average upwards, and 50% chance of price going 100 pips on average downwards at London open.
Say a trader was to open a trade at London open.
Now I assume that since we have a fixed time, which was chosen due to exponentially increasing volatility at such time, a fixed average range.
Our entry is not so random after all ?
The truth is hidden from you
  • #1,250
  • May 10, 2017 3:03pm May 10, 2017 3:03pm
Quoting MaxenshteinD
How does one battle the 50/50% chance? Say 50% price has chances of going 100 pips on average upwards, and 50% chance of price going 100 pips on average downwards at London open. Say a trader was to open a trade at London open. Now I assume that since we have a fixed time, which was chosen due to exponentially increasing volatility at such time, a fixed average range. Our entry is not so random after all ?
I was once successful with DIBS, quit it though because I seeked for bs which would allow me to predict each price swing with a 100% accuracy, because it gave me more pleasure to successfully predict price in order to show off (even though I had to pay the price of losing due to this). Now recently after discovering a few golden nuggets, I realised that why not try a statistical approach. Turned out that I was more successful with a 50/50% inside bar break at a relatively fixed time more than I was with all the complicated subjective bs (which even though allowed me to pick exact tops and bottoms at time made me a loser).
The truth is hidden from you
  • #1,251
  • Edited May 11, 2017 7:05am May 10, 2017 4:10pm | Edited May 11, 2017 7:05am
1
  • #1,252
  • Edited May 11, 2017 7:05am May 10, 2017 4:44pm | Edited May 11, 2017 7:05am
Last rant before my steak! In order to develop as a trader you must drop all mathematical models of trading including channel, slopes, three point touches of a line drawn randomly in space, fib levels clouds, ma's, etc. These are all based on some logical math formula that doesn't apply to a business driven by forces with huge sums of money at their disposal. Probabilities matter in that their are certain areas on a chart where something happened. If something significant happened at point A, their is a better than 50/50 chance that something significant will happen again when price returns to point A. These are the only places on a chart you should be trading. If you have the capital to last, you will eventually figure this out, and realize that 17 squiggly lines on your chart are of no use!! Happy trading.....
  • #1,253
  • May 10, 2017 4:46pm May 10, 2017 4:46pm
Quoting nealpek
Max, your logic is flawed. You are trying to attach 50/50 randomness to a situation that isn't random. This type of probability does not apply in this case. To you, there is a 50/50 chance of price going up or down 100 pips. To the market there is a 100% chance of the market going down 32 pips from london open then going 121 pips to the days close. The move has already been predetermined before london open even occurred. Although unknown to you, your visual clues, given to you in the form of candles or bars shows you where the market may go. It's...
Can a retail parasitic trader (trying to profit by anticipating "big boys" moves) determine objectively the aproximate odds of price closing above or below at the london's close though?
The truth is hidden from you
  • #1,254
  • May 10, 2017 4:53pm May 10, 2017 4:53pm
Quoting nealpek
If you have the capital to last
I have a couple of friends who insist that they are failing in trading only because they don't have sufficient capital.
But isn't it as easy to lose 1 million $ as a 100$ if one doesn't know what he is doing ?
Were you implying that one needs a sufficient capital to learn to trade by losing allot of money trading live? (Why not use a demo? One either knows how to trade or doesn't, live money might trigger a different approach which might allow a trader to learn new things, however the trader without an objective methodology would most likely be subjected to unnecessary stress and greed.
Or were you implying that profitable traders deal with huge draw downs? If that is the case, than the required capital (as well as profit) can be lowered by trading in smaller lot sizes.
The truth is hidden from you
  • #1,255
  • May 10, 2017 4:53pm May 10, 2017 4:53pm
Absolutely not! The only way to determine what will happen is to see what price does in and around the london open, not in 1 hour but over several hours of trading, to determine the direction they want to take price. You must look at what is above price and what is below price at any given point on a chart to determine likely direction!
I can never accurately predict from the london open where price will go. It depends on the nearest swings, and overall direction of the market at that time where price is likely to go. Hope this helps..cheers!
  • #1,256
  • May 10, 2017 4:57pm May 10, 2017 4:57pm
Quoting nealpek
Absolutely not! The only way to determine what will happen is to see what price does in and around the london open, not in 1 hour but over several hours of trading, to determine the direction they want to take price. You must look at what is above price and what is below price at any given point on a chart to determine likely direction! I can never accurately predict from the london open where price will go. It depends on the nearest swings, and overall direction of the market at that time where price is likely to go. Hope this helps..cheers!
So basically its easier to predict if the price will be lower or higher a few hours after the open of the session rather than minutes?
P.s. I really do hope that I don't sound sarcastic and etc. Because I'm honestly not
The truth is hidden from you
  • #1,257
  • Edited May 11, 2017 7:04am May 10, 2017 4:59pm | Edited May 11, 2017 7:04am
1
  • #1,258
  • Edited May 11, 2017 7:04am May 10, 2017 5:01pm | Edited May 11, 2017 7:04am
1
  • #1,259
  • May 10, 2017 5:04pm May 10, 2017 5:04pm
Quoting nealpek
Never ever put your money on an event that occurred in a couple of minutes. Pure suicide!!! Must use higher TF
why ?
  • #1,260
  • May 10, 2017 5:25pm May 10, 2017 5:25pm
Quoting mzvega
Serial Correlation........ The lack of day-to-day serial correlation in auction markets is often the starting point for random market theory. Studies as early as Labys and Granger (1970) confirm the day-to-day randomness. Even in trending markets, the expectation of up price or down price for the next dayis still about 50-50. Even odds in non-trending markets is easy to understand; in trending ones it is not so obvious. A market that is trending up or down does have an enhanced probability in the trend direction; it is just pretty well masked by...
1 trader viewing now