What is considered an edge? - page 5
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
That begs a greater question... Do most future data patterns already exist in the historic sample tested, or are most future patterns truly unique?
Based on my own work with total tick history and custom charts, I'm in the "already exists" camp.
Yes if price reacts to the values of your indicator(s), which they don’t.
You somehow imported "indicator(s)" into your reply to my post regarding custom charts. Most traders are not familiar with custom charts, so I'll explain a bit. A custom chart is not an indicator. More accurately, a custom chart is a restructuring (or reorganization) of prices, volumes, and/or times of ticks (and/or traditional timeframes). Custom chart data is basically cast onto a "blank slate" chart wherein only the traditional time and price scales remain. Custom chart generators are usually EA's and in my case, sometimes MQL5 services.
For my purposes, the range of custom chart structures include traditional Renko bars, offset Renko, rail-to-rail Renko, wick'ed Renko, and every possible combination (204 total including offset percentages <= 50%) of those functions/features; range bars; and equal volume bars (sometimes referred to as tick volume bars). Those are all charts─not indicators.
My trading EA's run on those custom charts and are coded to trade the "source symbols" of those custom charts.
In my case, the custom chart is structured based on the entire tick history available which can go back several decades. Even if I apply an indicator to such a custom chart (which is not always the case), it is all of the tick data that has been score optimized into the most uniform structure possible─not an indicator nor a trading EA.
The underlying premise is the fact that it's tough to discern anything meaningful from the chaotic mess of traditional timeframes. When the data is mathematically restructured into discernable patterns, anything applied to that data is inherently smoother and more reliable.
You somehow imported "indicator(s)" into your reply to my post regarding custom charts. Most traders are not familiar with custom charts, so I'll explain a bit. A custom chart is not an indicator. More accurately, a custom chart is a restructuring (or reorganization) of prices, volumes, and/or times of ticks (and/or traditional timeframes). Custom chart data is basically cast onto a "blank slate" chart wherein only the traditional time and price scales remain. Custom chart generators are usually EA's and in my case, sometimes MQL5 services.
For my purposes, the range of custom chart structures include traditional Renko bars, offset Renko, rail-to-rail Renko, wick'ed Renko, and every possible combination (204 total including offset percentages <= 50%) of those functions/features; range bars; and equal volume bars (sometimes referred to as tick volume bars). Those are all charts─not indicators.
My trading EA's run on those custom charts and are coded to trade the "source symbols" of those custom charts.
In my case, the custom chart is structured based on the entire tick history available which can go back several decades. Even if I apply an indicator to such a custom chart (which is not always the case), it is all of the tick data that has been score optimized into the most uniform structure possible─not an indicator nor a trading EA.
The underlying premise is the fact that it's tough to discern anything meaningful from the chaotic mess of traditional timeframes. When the data is mathematically restructured into discernable patterns, anything applied to that data is inherently smoother and more reliable.
I know what custom charts are, but in my eyes they are still a form of indicator. Same as timeframe based charts are.
To the extent that any data that is printed whatsoever is an indicator, ok, I guess. I was thinking more in terms of MT5 terminology. "When in Rome, do as the Romans do"?
But they will still not show you where most stop/limit orders are located.
That seems to be a recurring theme here in your thread. If you want to read limit orders, you can do that with a data subscription that provides order book access. For example, real futures listed on a centralized exchange such as the CME. Stop orders are invisible to prevent manipulation.
I read elsewhere that you trade the AUDCAD. That's an OTC instrument─having no order book access unless you have real DMA/wholesale account. This brings up an important distinction between exchange markets and OTC markets─what who has access to, if you will. But I digress...
[Technical analysis] will never work all the time and because of that it’s risky.
As another poster has previously said in this thread, nothing works "all the time." Anyone who claims that they never incur a loss is a poorly skilled BS artist─certainly, not me.
I simply don't know how people trade without having run the numbers first. Some do, do it. I, on the other hand, must have something to hang my hat on prior to putting on a trade. Put simply, the likelihood that any given trade taken in its context will be profitable.
I also read elsewhere that you're grid trading. Are you saying that grid trading has no risk? I guess that a grid is an indicator too... along with the chart on which it runs.🤷♂️
To the extent that any data that is printed whatsoever is an indicator, ok, I guess. I was thinking more in terms of MT5 terminology. "When in Rome, do as the Romans do"?
By using candle formation/pattern, tops/bottoms of previous high/low price levels and so on it is by definition an indication of some sort, hence indicator.
Stop orders are invisible to prevent manipulation.
I understand that this will be hard for retail traders to take in, but the markets are unfair to the little guy.
As I have said before, large banks sell this information for large amount of money to hedge-funds and larger players.
Having the information from banks when/where iceberg orders will be deployed is also a play to use, price usually halt at those levels and reverse because of market order exhaustion.
I read elsewhere that you trade the AUDCAD. That's an OTC instrument─having no order book access unless you have real DMA/wholesale account. This brings up an important distinction between exchange markets and OTC markets─what who has access to, if you will. But I digress...
Trading it isn't the same as getting my information from that specific currency pair, more than just using the chart that its on and taking the trades using a forex broker. I get my trade information/decision elsewhere.
As another poster has previously said in this thread, nothing works "all the time." Anyone who claims that they never incur a loss is a poorly skilled BS artist─certainly, not me.
The reason I know that banks sell their information is that many years ago I had the privilege to use that said information in my trading. By using that information 99/100 trades were winners. (not completely 100%)
The problem is knowing how long you can stretch the winners before price reverses on you. Other large players will influence the direction of the price too.
I simply don't know how people trade without having run the numbers first. Some do, do it. I, on the other hand, must have something to hang my hat on prior to putting on a trade. Put simply, the likelihood that any given trade taken in its context will be profitable.
Yes, you can run the numbers and make some backtests to see where price usually move to your likings. But you will never know for sure that there will be any icebergorders or stoporders at those levels to push/halt the price, unless you have that information from elsewhere.
I also read elsewhere that you're grid trading. Are you saying that grid trading has no risk? I guess that a grid is an indicator too... along with the chart on which it runs.🤷♂️
I use several different EAs on several different platforms/brokers. Some are grid based EAs, some are not. What I do not have is scalping strategies.
The trades on those EAs are not taken because of any retail indicators, but I do use retail indicators in some other ways inside the EAs though.
With this said the markets are becoming harder to read using my "information" and mistakes will be made, we will see what will happen going forward.
I have discussed a couple of times in threads about EAs and if they have any real edge or not. But what is really considered an edge?
Is it a combination of indicators that on historical prices of a specific market can predict the direction of price most of the time?
Will it suffice to have a backtest to prove an edge, or does it have to be a live track record of a year back or more with a specific amount of trades that show overall profit?
Or does the edge need to be more of a predicting nature, like knowing news before it is official or some kind of insider information?
Can an edge be a money management on a some what ok strategy?
Can the edge be proven without showing the code behind the EA?
Or is it something completely different?
I would really like to hear what people here think.
I myself don't think any EAs in the Market section here have any real edge, even if some claim this. They are just curve fitted on historical prices, over and over again.
Measure it on per-trade results, not on the equity curve. Take net P/L per trade after commission and swap, and compute the mean, the standard deviation and t = mean / (stdev / sqrt(N)). Roughly t above 2 on a few hundred independent trades is suggestive, and parameter searching lowers that, so require it on data you did not optimise on. Then compare live with backtest over the same dates: same trade count, win rate, average win/loss and average cost per trade. If live average trade is within about one standard error of the backtest, the edge is plausibly intact. If it sits far below, check costs (spread and slippage per fill) first, before blaming the market. Also track a rolling window, for example the last 50 trades, so you can see decay as it happens rather than guess afterwards.
This only measures if your "edge" is working or not. You still need an edge to begin with, which most people here don't have.