Currency Strength Meters: Built From 28 Pairs, and Where They Mislead

A currency strength meter splits pair prices into a score for each individual currency, to answer a simple question: did EURUSD rise because the euro got stronger, or because the dollar got weaker? It adds no new information. Every number comes from past prices of pairs you can already see. This article builds a meter step by step from 28 pairs, runs it on made-up numbers, and then looks at where it can mislead you.

Why 28 pairs

A currency pair is always a price of one currency in terms of another. EURUSD up 0.3% could mean a stronger euro, a weaker dollar, or both. One pair can’t tell them apart.

The fix is to look at the same currency against many others. Most meters use the 8 majors, EUR GBP AUD NZD USD CAD CHF JPY. They combine into 8 × 7 ÷ 2 = 28 pairs, and each currency appears in exactly 7 of them. If the euro is up against all 7, it’s fairly clear the euro itself is strong and it isn’t just a weak dollar.

Each pair is quoted only one way round. There is EURJPY but no JPYEUR. The first currency is the base, the second is the quote. When a pair rises, the base has strengthened against the quote.

Building the meter step by step

1. Choose a lookback, the window you measure over, such as 1 day, 5 days or 20 days
2. Work out each pair’s % change = (latest close ÷ close one lookback ago − 1) × 100
3. Collect each currency’s 7 pairs. Where the currency is the base, use the value as it is. Where it is the quote, flip the sign. EURAUD −0.45% means AUD gained about +0.45% against EUR
4. Average the 7 values. That’s the currency’s strength score

Strength of currency X = average % change of the 7 pairs that contain X (sign flipped where X is the quote currency)

Two properties to understand before you use one:

• The 8 scores always sum to zero. Every pair is counted once as a plus for its base and once as a minus for its quote. A meter can only rank currencies. It can never show all of them getting stronger at once
• Score ≈ the currency’s own move minus the average move of the other 7. It is purely relative

Meters you find online may work differently: scaled to 0–10, measured in pips, built from each pair’s RSI, or with some pairs weighted more. So numbers don’t match across programs. The core is the same, though: ranking currencies from the changes in the same set of pairs.

Worked example: recovering the hidden ranking

We generated made-up data in node, starting from a one-day move for each currency (the moves sum to zero): AUD +0.45%, NZD +0.30%, CAD +0.15%, GBP +0.05%, EUR −0.05%, CHF −0.15%, USD −0.25%, JPY −0.50%.

From these we worked out the % change of all 28 pairs from the moves of their two currencies, then added random noise of roughly ±0.05% to each pair, the way each real pair has its own order flow, and rounded to two decimals. Now pretend we don’t know the answer and compute the scores from those 28 numbers alone.

AUD’s 7 pairs:

Pair	% change	AUD is the	Value used
EURAUD	−0.45	quote	+0.45
GBPAUD	−0.39	quote	+0.39
AUDNZD	+0.16	base	+0.16
AUDUSD	+0.64	base	+0.64
AUDCAD	+0.31	base	+0.31
AUDCHF	+0.56	base	+0.56
AUDJPY	+0.88	base	+0.88

Total = 3.39, divided by 7 = +0.48, AUD’s score.

JPY is the quote in all 7 of its pairs (EURJPY +0.46, GBPJPY +0.50, AUDJPY +0.88, NZDJPY +0.79, USDJPY +0.29, CADJPY +0.65, CHFJPY +0.30). They total 3.87. Flip the sign and divide by 7 to get −0.55.

All 8 currencies:

Currency	Meter score	Move we set	Theory (× 8 ÷ 7)
AUD	+0.48	+0.45	+0.51
NZD	+0.35	+0.30	+0.34
CAD	+0.17	+0.15	+0.17
GBP	+0.04	+0.05	+0.06
EUR	−0.04	−0.05	−0.06
CHF	−0.17	−0.15	−0.17
USD	−0.27	−0.25	−0.29
JPY	−0.55	−0.50	−0.57

Check: the 8 scores add up to exactly 0.000, and the ranking matches the moves we set in every position. The theory column follows from the property above. AUD moved +0.45, and the other 7 averaged −0.45 ÷ 7 = −0.064, so its score should be 0.45 + 0.064 = 0.514, which is 0.45 × 8 ÷ 7. The small gaps left over are the noise. Note that GBP and EUR are only 0.08 apart, so a little noise could swap them.

The strongest-vs-weakest pair is AUD against JPY, AUDJPY, up +0.88%. It is also the biggest mover of all 28 pairs, followed by NZDJPY +0.79 and CADJPY +0.65. That’s no coincidence. The scores are relative moves, so the two currencies furthest apart on the meter form the pair that has already travelled furthest in that window.

Picking “strongest vs weakest” means picking the pair that has already moved the most and betting it keeps going. That is momentum, not foresight.

Choosing the lookback

The same meter with a different window gives a different ranking. We generated 20 days of made-up daily data (each currency with a small underlying drift plus daily random noise) and computed the scores over three windows.

Rank	Last 1 day	5 days	20 days
1	CHF +0.68	CHF +1.28	CHF +2.65
2	GBP +0.40	EUR +0.69	JPY +0.68
3	NZD +0.01	AUD +0.18	EUR +0.34
4	EUR −0.02	CAD +0.18	CAD −0.49
5	JPY −0.02	JPY −0.21	AUD −0.60
6	CAD −0.12	USD −0.39	USD −0.83
7	AUD −0.27	GBP −0.66	GBP −0.86
8	USD −0.66	NZD −1.07	NZD −0.89

Three things stand out:

• The middle of the table swings a lot. GBP is 2nd over one day but 7th over five. NZD is 3rd over one day but weakest over five
• The weakest currency depends on the window. One day points to USDCHF, while 5 and 20 days point to NZDCHF. Over 20 days, USD, GBP and NZD are less than 0.07 apart, so “weakest” is close to a coin toss
• Noise can beat the trend. In the data we built, EUR was given a slightly stronger underlying drift than CHF, yet 20 days of random noise put CHF on top in every window. In a real market you never see the underlying drift, only the noisy total

Rules of thumb:

• Match the lookback to how long you plan to hold. If you hold for days, don’t steer by a 1-hour meter
• Check two windows at once, the one you trade and a longer one, and only give weight to a reading when both agree. This is the same idea as multi-timeframe analysis
• Use closes of finished candles. A meter that uses the live price of an unfinished candle keeps changing until the candle closes (more in repainting indicators)
• Don’t try many lookbacks and keep whichever backtests best. That is hunting for luck. Read multiple testing and data snooping

Using the meter to pick pairs sensibly

Treat it as a filter, not an entry signal.

1. Look at the gap between scores, not just the ranks. Two currencies near zero in the middle of the table say very little. The gap should be bigger than the pair’s normal movement over the same window, which you can compare with its ATR
2. Confirm on the pair’s own chart. The meter gives no entry, no Stop loss and no support or resistance. You still need the price structure
3. Ask why the currency is strong. Rate news, economic data, or a risk-on, risk-off mood? If it comes from a single release, the move may fade quickly
4. Count the cost. Strongest-vs-weakest is often a cross such as AUDJPY or NZDCHF, where spreads are usually wider than on the majors. See spread cost per trade
5. Don’t open several pairs that share a currency. In the example, AUDJPY, NZDJPY and CADJPY all rose. Taking all three the same way is one yen bet at roughly three times the risk. Read correlation and portfolio risk and size every trade from its Stop loss (position size from Stop loss)

Where the meter misleads

1. Every meter is a lagging average of the same pairs. 28 pairs from 8 currencies contain only about 7 independent numbers: each currency’s move against one reference currency. Almost everything else can be worked out from those. A meter just rearranges them to be easier to read. It always lags price, like a moving average, and it holds nothing the 28 charts don’t.

2. Different lookbacks give different answers. As the table above shows, two websites whose meters disagree may both be right. They’re just measuring different windows.

3. Dollar news moves everything at once. The dollar is in 7 of the 28 pairs and is the most traded currency in the world. Take the example again and add a made-up news move that weakens USD by a further 0.6%, with nothing changing for any other currency.

USD falls from −0.27 to −0.88 and replaces JPY as the weakest currency, so the strongest-vs-weakest pair switches from AUDJPY to AUDUSD. The other 7 all rise by about +0.09 (0.6 ÷ 7) though nothing changed for them. EUR flips from negative to positive. Reading that as “the euro is turning strong” would be wrong. It’s only the dollar’s shadow. Around big US releases, keep an eye on the dollar index DXY and the economic calendar as well.

4. Extreme readings can continue or snap back. Picking strongest vs weakest is a momentum bet. Academic research has found cross-currency momentum at monthly horizons (for example Menkhoff, Sarno, Schmeling and Schrimpf, 2012), but the effect was not steady over time and trading costs ate part of it. Over very short horizons, extreme readings often come from a single release or a temporary flow, and sometimes they reverse. There is no fixed rule for which will happen, and the meter can’t tell you which case you are in.

5. Data sources close at different times. The daily candle’s close depends on each server, so “1 day” on two platforms may start at different times. Around the daily rollover (roughly 04:00–06:00 Thai time (UTC+7)) liquidity is thin and cross-pair spreads widen, so the prices a meter uses can jump oddly.

Common mistakes

• Treating the meter as a forecast. It summarises what already happened, no different from looking at 28 charts at once
• Reading ranks without sizes. 4th and 5th place 0.02 apart doesn’t mean one is really stronger than the other
• Entering every time a currency hits number 1. By the time it’s number 1, price has already moved. Without a Stop loss that makes sense, that’s chasing
• Ignoring dollar news. After a US release the whole table moves with USD. Currencies that look “stronger” may have nothing going on at all
• Opening several pairs with the same currency. Buying AUDJPY, NZDJPY and CADJPY together is one yen risk, tripled
• Changing the lookback until you see the answer you want. Set it once to fit your holding period, then use it the same way every time