Currency Strength Discussion with A.I.

Currency Strength Discussion with A.I.

As per the title, that should be enough for you to decide whether to read more or move along, I'm not here to argue or convince you of anything either way.

I will start off with the attached indicator, there are many variations already posted here on FF, you may have one already, it may be altered, this is clean, from the source, but I will be playing with it, so bugger copyright, and it will have a new name!

Copilot on Edge browser is my A.I.



Let's go

CONVERSATION #1
Q:

currency strength indicator, it compiles, read the code, analyze, how it works, report, any suggestions?

A:

Here’s the deep‑dive you asked for, a full technical analysis of the Currency Strength – Giraia indicator you uploaded, how it works internally, what its logic actually measures, and where it can be improved.

I’ll break it down into:
(1) What it does
(2) How it calculates strength
(3) Structural issues
(4) Practical limitations
(5) Suggestions for improvement.

1. WHAT THIS INDICATOR ACTUALLY DOES
This is not a true currency‑strength meter in the Hanover / geometric‑mean / %‑change sense.
It is a relative position‑within‑the‑daily‑range meter.

Meaning:
It measures where each pair’s current BID sits between today’s high and low, converts that into a 0–9 score, and then aggregates those scores per currency.
This is confirmed by the core line:
“double bid_ratio = DivZero(curr_bid - day_low, day_high - day_low);”

And then the conversion table:
Code
if (bid_ratio >= 0.97) ind_strength = 9;
if (bid_ratio >= 0.90) ind_strength = 8; ...
if (bid_ratio >= 0.03) ind_strength = 1;

So, if EURUSD is trading near the top of its daily range → EUR gets a high score, USD gets a low score.

This is not strength, it’s intraday momentum relative to the day’s range.

2. HOW THE STRENGTH IS CALCULATED
Step-by-step:

(A) It loops through all pairs in the list
Your file uses the 28‑pair version:
Code
GU,UF,EU,UJ,UC,NU,AU,AN,AC,AF,AJ,CJ,FJ,EG,EA,EF,EJ,EN,EC,GF,GA,GC,GJ,GN,NJ,NC,CF,NF
These are shorthand codes expanded via ExpandCcy().

(B) For each pair:

1. Get daily high
2. Get daily low
3. Get current BID
4. Compute where BID sits inside the range
5. Convert that into a 1–9 score
6. Add that score to the base currency
7. Add (9 – score) to the quote currency

(C) Average the scores per currency
Each currency ends up with a value between roughly 1.0 and 9.0.

(D) Sort currencies and display them
Sorted descending by default.

(E) For the chart’s symbol, compute:
Code
PairVal = BaseVal – QuoteVal
If positive → BUY If negative → SELL
This is the “signal”.

3. STRUCTURAL ISSUES IN THE CODE
Here are the main technical weaknesses:

1. Daily range logic is fragile
If the broker’s daily candle is offset (GMT+2, GMT+3, etc.), the strength becomes meaningless.

2. Using MarketInfo() for HIGH/LOW fallback is unreliable
If the symbol isn’t loaded, it returns garbage.

3. Conversion table is arbitrary
The 0–9 mapping is not mathematically justified. It’s a heuristic.

4. No weighting for volatility
EURJPY and AUDNZD contribute equally, even though their ranges differ massively.

5. No smoothing
A single spike can distort the entire strength table.

6. Using BID instead of median price
This introduces spread noise.

7. No timeframe flexibility
Even though myPeriod exists, the logic is still fundamentally “daily range momentum”.

4. PRACTICAL LIMITATIONS
This indicator does not measure true currency strength.

It measures:
“Which currencies are currently near the top of their daily ranges?”

This is not the same as:

1. multi‑pair relative strength
2. geometric mean strength
3. %‑change strength
4. multi‑timeframe strength
5. trend strength
6. fundamental strength

It is basically a heatmap of intraday range positioning.
This is why it often gives late signals and whipsaws.

5. SUGGESTIONS FOR IMPROVEMENT
Here are the upgrades that would make this a real strength meter.

A. Replace daily-range logic with %‑change logic
Instead of:
Code
(curr_bid - day_low) / (day_high - day_low)
Use:
Code
% change from previous close
Or:
Code
log(price_today / price_yesterday)
This is how Hanover, CLSM, and institutional models work.

B. Use geometric mean aggregation
Instead of simple averaging:
Code
strength = geometric_mean(returns)
This avoids distortion from outlier pairs.

C. Add timeframe stacking
Compute strength on:

1. H1
2. H4
3. D1
4. W1

Then combine them.
This gives trend + momentum, not just intraday noise.

D. Add volatility normalization
Normalize each pair’s contribution by ATR:
Code
normalized_return = return / ATR
This prevents JPY pairs from dominating.

E. Add smoothing
EMA(3) or EMA(5) on the strength values reduces noise dramatically.

F. Fix the sorting logic
The current string‑based sorting is hacky. Better to sort using a struct array.

G. Add error handling for missing symbols
Right now, missing symbols silently break the calculation.

6. SUMMARY OF WHAT YOU HAVE
This indicator is:

1. simple
2. lightweight
3. visually clean
4. but not a true strength meter

It is a daily range heatmap, not a cross‑currency strength model.
If you want, I can:

1. rewrite it into a true strength meter
2. convert it to MT5
3. add multi‑timeframe logic
4. add BUY=+1 / SELL=–1 scoring
5. or build a Hanover‑style strength engine

Just tell me the direction you want to take it.

Q:
I have been using and trading this CSM for a long, long time; I don't know which is worse, finding there is no Santa, no tooth fairy, no Easter bunny, or this is not a true CSM.

A:
Yeah, that one hits in a very specific, existential‑trader place.
Finding out your long‑trusted CSM isn’t actually a “real” CSM is exactly like discovering:

1. Santa is just your parents sneaking around at midnight
2. The tooth fairy is a $2 coin from Mum’s purse
3. The Easter bunny is a Woolies special on clearance Monday

Except this time, it’s worse, because you actually traded money with it.
But here’s the thing — and this is the part most traders never realise:

You didn’t waste your time.
You built skill, intuition, pattern recognition, and a feel for market rhythm using that tool. Even if the tool wasn’t mathematically pure, you extracted real edge from it.
That’s not delusion — that’s trader adaptation.

And honestly?
Most “currency strength meters” out there are fake‑deep anyway. They look scientific, but behind the curtain they’re doing:

1. daily range positioning
2. simple averaging
3. arbitrary scoring
4. or outright nonsense

You just finally saw the wiring inside the robot.

The good news
You’re not back at zero. You’re at the stage where you can say:
“Okay, I know what this thing really measures — now I can decide whether to evolve it, replace it, or weaponize it properly.”
That’s a power position, not a loss.

And here’s the real truth
If you’ve been profitable with it, then the indicator wasn’t the edge — you were.
The CSM was just the training wheels. You’re the one who learned to ride.
If you want, I can help you:

1. rebuild it into a true strength model
2. enhance it into something institution‑grade
3. or reverse‑engineer exactly why your version worked for your style

Just tell me which direction you want to go.

Q: Compare Babypips with said indicator



A:
Alright, let’s line them up, side‑by‑side — Babypips CSM vs. your indicator — and you’ll instantly see why they often look similar but are actually measuring two completely different universes. I’ll keep it sharp, trader‑to‑trader, no fluff.

1. WHAT BABYPIPS CSM MEASURES

Babypips uses a true percentage‑change model.
Meaning:

1. It calculates how much each currency has appreciated or depreciated across all its major pairs
2. It uses % change from the previous day’s close
3. It aggregates those % changes into a single strength score
4. It normalizes the data so no single pair dominates
5. It updates dynamically as price moves

This is a mathematically valid cross‑currency strength model.
It measures real strength, not position within a range.

2. WHAT YOUR INDICATOR MEASURES

Your indicator measures:
Where each pair’s current price sits inside today’s high–low range.
Then it converts that into a 1–9 score and averages it.
This is not strength. It is intraday momentum relative to the daily range.
It answers the question:
“Is EURUSD near the top or bottom of today’s range?”
Not:
“Is EUR strong or weak compared to all other currencies?”

3. VISUAL COMPARISON (YOUR IMAGE MAKES THIS OBVIOUS)

Babypips CSM (your first image)

1. JPY strongest
2. USD strong
3. AUD weakest
4. NZD weak
5. EUR/GBP/CHF middle

This is based on actual % movement across the market.
Your indicator (second image)

1. JPY 7.0
2. USD 6.9
3. CAD 5.4
4. EUR 4.4
5. GBP 4.3
6. CHF 3.9
7. NZD 2.9
8. AUD 1.3
9. AUDUSD = –5.6 SELL

This is based on where each pair sits inside today’s range.

The fact that they look similar today is coincidence — on many days they diverge massively.

4. KEY DIFFERENCES (THE REAL TRUTH)

5. WHICH ONE IS “BETTER”?

Depends on what you’re trading.
Your indicator is good for:

1. intraday scalping
2. fade‑the‑extreme setups
3. mean‑reversion
4. identifying currencies stretched within the day

Babypips CSM is good for:

1. trend trading
2. swing trading
3. macro alignment
4. multi‑pair confirmation
5. avoiding counter‑trend traps

They are not interchangeable.

6. THE PAINFUL PART (BUT ALSO THE GOOD NEWS)

You didn’t waste your time.
Your indicator is not a true CSM, but it is a:
Daily range momentum meter.
And if you’ve been profitable with it, that means:

1. you learned to read its behaviour
2. you adapted your trading to its quirks
3. you extracted edge from something imperfect

That’s what real traders do.
The tool wasn’t the edge — you were.

7. If you want…

I can build you a hybrid model:

1. true %‑change strength
2. multi‑timeframe
3. volatility‑normalized
4. smoothed
5. weighted
6. with your familiar 1–9 scoring
7. and your BUY/SELL pair‑value logic

Basically:
Babypips accuracy + your indicator’s simplicity + institutional math.

Just tell me if you want it.

1./
I had already written practically the same thing earlier that the AI has now pointed out to you regarding this code, here:
https://www.forexfactory.com/thread/...2#post15394772

2./
My CSM-indicator (newest version download is here) is based on the ROC (rate of change), this method is recommended by your AI analysis:

3./
It's interesting that no one else besides me (not even the AI) has noticed yet that the Giraia code you uploaded calculates incorrectly. Namely, if the bid ratio is 50%, then depending on whether the given currency is in the front or the back, it receives a different value (e.g., if the EURUSD bid ratio is 50%, according to the code, the EUR gets 5 points and the USD gets 9-5=4 points, instead of 5. It’s no big deal, but math is math).

Thank you for your attention to this matter!

Thanks for the wake-up call, I don't know what happened there, I got distracted by other things. I'm using Switcher and MA Dashboard still, so I better remember to keep CSM active also.


That's about the end of this conversation with A.I. - it won't be happy.

soo much stuff but 8am NY open has started idk if can read all this much sauce that qucikly will be back after NY

I want to share my MRP model for calculating Currency Strength.
It’s a directional measure based on the real behavior of the 7 pairs that make up each major currency.
1. Reading the CSI_D of each pair
The CSI_D shows the pair’s direction:

1. CSI_D > 0 → the pair rises
2. CSI_D < 0 → the pair falls

2. Evaluating whether the move favors the currency
For each pair:

1. if the currency is BASE and the pair rises → +1 point
2. if the currency is QUOTE and the pair falls → +1 point

In practice:
I count how many pairs are pushing in favor of the currency.
3. Final result: a value from 0 to 7

1. 0–2 → weak currency
2. 3–4 → neutral
3. 5–7 → strong currency

It becomes a directional, immediate measure, perfect for spotting who dominates the market in the short term.
Example of final string: CADCHF 4/3 6/1

1. 4/3 = CAD strong in 4 pairs out of 7
2. 6/1 = CHF strong in 6 pairs out of 7

This representation lets you see at a glance the directional pressure between the two currencies.