- This topic has 120 replies, 12 voices, and was last updated 9 years, 12 months ago by
.
- Posts
Well, you know my opinion on currency strength, do you?!
Most CS indicators are lagging to some extent. And so do my early attempts to code a CS. All those discussions with @gg53 about this topic convinced me that this is not necessarily so, and my latest CIX is tuned to be much more reactive than every CS I coded earlier. But ok, it’s a bit choppy now. And since we want to calculate a slope, ‘some’ kind of lag will occur inevitably.
GOOD Currency-Strength indicator isn’t lagging, but precedes Price-Action by at least 1-2 bars.
I think I’m not quite there, but I’m working on it!
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
@simplex: My statements/thoughts are general observations, not only based on your CIX. Well, the delay is imho not only due to the calculation of slope. You can see it always when currency 1 is far from the 0 line (or curreny 2 far below the 0 line). But maybe we would get preceding signals using smaller TFs to current TF on which we trade.
@gg53: Would you please share a screenshot of your currency strength indicator showing preceding signals to reversals?! Thanks in advance!On Currency-Strength:
There are two kinds of Currency-Strength indicators, and people tend to confuse between them and even program indicators with a “mixed” logic.
Absolute Currency-Strength, measure Currency-Strength based on some anchor in past history, either date or some meaningful historical value.
Relative Currency-Strength is more responsive for trading and measures short-term interaction between the currencies.
The basics of Relative Currency-Strength is quite simple, although implementation is more complex.
In Relative Currency-Strength we need to measure the difference between current move against previous move, something like Close[0]-Close[1].
Problem 1: This will result in a very jagged indicator, and why use Close[x] value? it’s more logical to use Typical price ((H+L+C)/3).
Problem 2: Even Typical price will result in very jagged indicator. Using MA will result in lagging. Better to use DELTA between two MA’s of Typical price.
Problem 3: Not all currencies are created equal. A movement in one currency is not equal to same movement in other currency. Some currencies are more volatile than others, and also changed by market conditions. The solution is to devide the movement by ATR, which will result in TRUE relative movement that reflects interaction between currencies.
More will follow…
G.
2: Even Typical price will result in very jagged indicator. Using MA will result in lagging. Better to use DELTA between two MA’s of Typical price. Problem
Well, the difference between two MAs is also lagging as it is a function of the MA values itself
2: Even Typical price will result in very jagged indicator. Using MA will result in lagging. Better to use DELTA between two MA’s of Typical price. Problem
Well, the difference between two MAs is also lagging as it is a function of the MA values itself
What about ADAPTIVE MA’s?
G.
This reply has been reported for inappropriate content.
Yes, too. They all lag. Although adaptive MAs mimic the prize curve faster, but the only MAs that don’t lag are MAs with period 1. But perhaps this is a matter of definition of the term “lag”
There is no “definition” issue between us, just the way what and how you implement things in order to overcome “lag”, and make your indicator a predictive one.
If you use ONLY price-action in your calculations and considerations, your indicator will ALWAYS be lagging and you’ll always chase Price-Action.
Here is a sample of non-lag, predictive sample of currency-Strength (even on H4), and a proper way of using it:
On the pink vertical line, (against price-action):
CADJPY, AUDJPY – Short
EURCAD, EURAUD – Long
G.
Using MAs for trading, we always have to find some balance between two antagonistic mathematical methods: integration and differentiation.
Applying the MA (integration) to prices always introduces a certain amount of lag. Choosing a state of the art adaptive filter instead of a simple SMA can reduce the amount of lag while at the same time enhance smoothing performance, but the lag of this method alone will always be greater than zero.
Now we’re taking that smoothed (lagged) price line and differentiate it by some punctual slope calculation ( smoothed(x) – smoothed(x+y) ) or a linear regression slope LRS(x, y) or whatever. The specific choice of a differentiation algorithm and its parameters will determine how much lag introduced by the integration will be removed. And usually this will produce a more jagged signal line as compared to the integrated one. While the act of differentiation will remove lag from the original MA line, the choice of any y parameter greater than 1 will again introduce a bit of lag to the differentiation part.
I think that this is just one example of the most basic problem of engineering: finding the balanced point of optimum performance for an engine somewhere between zero performance and self destruction.
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
Strangely enough, above pics show predictive indicator and not a lagging one…
Quite amazing, since it’s on H4 TF, i.e. 4 hours advanced prediction… so don’t let the facts confuse your theories…
G.
Thank you both.
@gg53: Yes, it’s predictive. However, you don’t trade your rules. You said: only trade if currency 1 is above 0 and pointing up, currency 2 below 0 and pointing down. But at the place where your pink line is in your last pic, EUR is above 0 and pointing down and AUD is below 0 and pointing up … Surely, that can be a predictive signal. However, I assume that there are many false signals especially during low volatility times.Thank you both. @gg53: Yes, it’s predictive. However, you don’t trade your rules. You said: only trade if currency 1 is above 0 and pointing up, currency 2 below 0 and pointing down. But at the place where your pink line is in your last pic, EUR is above 0 and pointing down and AUD is below 0 and pointing up … Surely, that can be a predictive signal. However, I assume that there are many false signals especially during low volatility times.
I do trade my rules, I just gave only one rule of using Currency-Strength indicator.
With my indicator I can also trade Counter-Trends with high accuracy and low risk, I’m not so sure it can be done with other indicators, since I haven’t seen a good one yet.
G.
Strangely enough, above pics show predictive indicator and not a lagging one… Quite amazing, since it’s on H4 TF, i.e. 4 hours advanced prediction… so don’t let the facts confuse your theories…
Not confused! I just think you found a stable and well-balanced set of parameters for your system to work at optimum performance.
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
Ok, that makes a big difference.
No, it’s not. If you read and use Gadi_Currencies @FF, those are the exact examples that are given years ago.
The rule I gave is valid and accurate, and simplex proved it in one of his latest post.
The same logic that trade strong currency against weaker one can be applied to falling strong currency against rising weaker currency.
The only difference is that counter-trend are faster and doesn’t last long.
G.
Strangely enough, above pics show predictive indicator and not a lagging one… Quite amazing, since it’s on H4 TF, i.e. 4 hours advanced prediction… so don’t let the facts confuse your theories…
Not confused! I just think you found a stable and well-balanced set of parameters for your system to work at optimum performance. s.Oh, I thought you know me by now… What about some kind of Volume-Weighted MA? You already recognize the power of VZO, Tick-Volume, etc.
G.
i think this is all semantics. you can have the point of view anything other than tick is lag but you can also have a point of view such as a close to close on a 4 hour chart is a 4 MA on a 1 hour chart. is the 4 ma lag? it is and it’s not. that is just one quick example. it just depends on point of view. i consider using a MA to measure mean revision or any oscillator both lagging and predictive. you both are correct. :)
“I believe the very best money is made at the market turns. Everyone says you get killed trying to pick tops and bottoms and you make all
your money by playing the trend in the middle. Well for twelve years I have been missing the meat in the middle but I have made a lot of
money at tops and bottoms.”
– Paul Tudor Jonesi think this is all semantics.
I’m afraid you’re right!
It’s a question of the individual strategy preferred: fishing for exact swing reversals under controlled conditions or ‘good old’ waiting for confirmation. Like one of the ‘great’ traders said: I’m satisfied to cut the middle 30 or 40 % of a swing and let the others fight for the rest. I don’t remember who it was, but I think it was in the 1960’s.
And I’m afraid markets are faster today as compared to the 1960’s. So ‘good old’ wisdom might be obsolete to go fishing for those 40 %.
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
Two instances of CIX used simultaneously as a practical answer?
Top: short term, very reactive, RSI normalization (this is for entry timing by anticipating swing reversal)
Bottom: medium term, timeframe normalization (this is for selection of which pairs to trade)
Combined SELL entries are at the orange lines.
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
@simplex: You’re speaking from my heart!
You’re speaking from my heart!
Great! Any specific post, or just generally since we met here?
s.
A good trader is a realist who wants to grab a chunk from the body of a trend, leaving top- and bottom-fishing to people on an ego trip. (Dr. Alexander Elder)
I was referring to your post #11633. Nevertheless, I’m sure that you are also a nice guy in real life.2: Even Typical price will result in very jagged indicator. Using MA will result in lagging. Better to use DELTA between two MA’s of Typical price. Problem
Well, the difference between two MAs is also lagging as it is a function of the MA values itself
Oh, really?
Price(t1) = 6, Price(t2)=7 —> Diff=1.
MA(t1)={4,5,6), MA(t2)={4,5,6,7) —> MA(t1,3)=5, MA(t2,3)=6 —> Diff=1
So, in the above private case, MA’s Diff is equal to actual price Diff. and that’s just simple SMA. So your statement isn’t ALWAYS right…
It’s so easy to throw pseudo-math statements and theories, without actually digging into the facts…
G.
i think this is all semantics. you can have the point of view anything other than tick is lag but you can also have a point of view such as a close to close on a 4 hour chart is a 4 MA on a 1 hour chart. is the 4 ma lag? it is and it’s not. that is just one quick example. it just depends on point of view. i consider using a MA to measure mean revision or any oscillator both lagging and predictive. you both are correct. :)
It’s NOT semantics.
What’s the point in creating new indicator if you are just chasing price-action?
What’s the point of using just RSI (another Price-Action, and VERY lagging. BTW:CCI is better and faster for that purpose) to model Currency-Strength?
If you are into 30-40% of long-term trend, you don’t need all those fancy indicators.
The FACTS in FX market are:
Volume precede Price and move the market.
Currency-Strength will give you, with just one quick look, an overview of current market forces and individual currency direction.
Any Non-Predictive CS indicator is as good as staring on price-action bars themselves – you don’t need it….
….And, a Non-Predictive Currency-Strength will signal you to EXIT a trade only 30-40% AFTER trend reversal, which is basically where you entered that trend, making it a possibly loosing trade…
G.
2: Even Typical price will result in very jagged indicator. Using MA will result in lagging. Better to use DELTA between two MA’s of Typical price. Problem
Well, the difference between two MAs is also lagging as it is a function of the MA values itself
Oh, really? Price(t1) = 6, Price(t2)=7 —> Diff=1. MA(t1)={4,5,6), MA(t2)={4,5,6,7) —> MA(t1,3)=5, MA(t2,3)=6 —> Diff=1 So, in the above private case, MA’s Diff is equal to actual price Diff. and that’s just simple SMA. So your statement isn’t ALWAYS right… It’s so easy to throw pseudo-math statements and theories, without actually digging into the facts… G.
Surely, it’s not always true. You can always create situations in which MA difference is even leading. But in most cases it won’t. And you only use two means – not a moving average, as MA1 includes 3 values, but MA2 four. What kind of MA is that (how would you decide when to take the last 3, when to take the last 4 numbers into account)? Don’t say that most people don’t understand math!
Not confused! I just think you found a stable and well-balanced set of parameters for your system to work at optimum performance. s.
- You must be logged in to reply to this topic.