Charts, Timeframes & Market Watch

Press Ctrl + N or go to View > Navigator.

You will see categories: Trend, Oscillators, Volumes, Bill Williams, and Custom.

A settings dialog will appear where you can adjust parameters, colors, and the timeframe.

Go: open a terminal in a folder you can write to — cd ~/Desktop on macOS or Linux, cd %USERPROFILE%\Desktop on Windows.

Do: save this as candle.py and run python3 candle.py.

"""A candle is four numbers chosen from hundreds. See what it discards."""
import random

random.seed(4)

def ticks_for_bar(n=60):
    t, p = [], 100.0
    for _ in range(n):
        p += random.uniform(-0.20, 0.20)
        t.append(round(p, 2))
    return t

ticks = ticks_for_bar()
o, h, l, c = ticks[0], max(ticks), min(ticks), ticks[-1]

print("ticks inside this one bar :", len(ticks))
print("the candle keeps          : 4 numbers")
print("   open  %.2f    high  %.2f    low  %.2f    close  %.2f" % (o, h, l, c))
print()
print("what the four numbers CANNOT tell you:")
print("   did the high come before the low?  ->",
      "high first" if ticks.index(h) < ticks.index(l) else "low first",
      "  (index %d vs %d)" % (ticks.index(h), ticks.index(l)))
print("   how many times did price cross the open?")
crossings = sum(1 for i in range(1, len(ticks))
                if (ticks[i-1] < o) != (ticks[i] < o))
print("      ->", crossings, "times")
print("   how long was it above the open?")
above = sum(1 for p in ticks if p > o)
print("      -> %d of %d ticks (%.0f%%)" % (above, len(ticks), 100.0*above/len(ticks)))
print()
print("A candle is a summary, and every summary discards. The four numbers")
print("survive; the ORDER of everything that happened does not.")
print()
print("This is exactly the information a strategy tester lacks when it uses")
print("OHLC modelling -- and exactly what decides whether your stop or your")
print("target was reached first.")

You should see: sixty ticks reduced to four numbers:

ticks inside this one bar : 60
the candle keeps          : 4 numbers
   open  99.89    high  100.33    low  99.15    close  99.64

what the four numbers CANNOT tell you:
   did the high come before the low?  -> low first   (index 42 vs 14)
   how many times did price cross the open?
      -> 11 times
   how long was it above the open?
      -> 17 of 60 ticks (28%)

A candle is a summary, and every summary discards. The four numbers
survive; the ORDER of everything that happened does not.

This is exactly the information a strategy tester lacks when it uses
OHLC modelling -- and exactly what decides whether your stop or your
target was reached first.

Sixty prices become four. What survives is the extremes and the endpoints; what is lost is the order of everything that happened — and the three questions the script asks are all questions about order.

The first one is the consequential one. The candle cannot say whether the high or the low came first, and that is precisely the fact that decides whether your stop or your target was reached when both were inside a single bar's range. It is the same gap that makes a strategy tester's cheaper modelling modes guess rather than know.

The other two are worth knowing about for a different reason: a bar in which price crossed its own open eleven times and spent only 28% of its life above it is not a “bullish” or “bearish” bar in any meaningful sense, whatever colour it is drawn.

If not: python3: command not found on Windows means Python was installed without “Add python.exe to PATH”; try py candle.py. The figures are seeded and will match this page.

Go: the same folder.

Do: save this as timeframe.py and run python3 timeframe.py.

"""The same data at three timeframes. The pattern changes; the data does not."""
import random

random.seed(4)

m1, p = [], 100.0
for _ in range(240):
    o = p
    p += random.uniform(-0.20, 0.20)
    m1.append((o, max(o, p) + 0.05, min(o, p) - 0.05, p))

def group(bars, factor):
    out = []
    for i in range(0, len(bars) - factor + 1, factor):
        chunk = bars[i:i+factor]
        out.append((chunk[0][0], max(b[1] for b in chunk),
                    min(b[2] for b in chunk), chunk[-1][3]))
    return out

def count_bullish(bars):
    return sum(1 for o, h, l, c in bars if c > o)

def biggest_range(bars):
    return max(h - l for o, h, l, c in bars)

for name, factor in (("M1", 1), ("M15", 15), ("H1", 60)):
    bars = group(m1, factor)
    print("%-4s  bars %4d   bullish %4d (%.0f%%)   biggest range %.2f"
          % (name, len(bars), count_bullish(bars),
             100.0 * count_bullish(bars) / len(bars), biggest_range(bars)))

print()
print("Identical price history. Four hours of it, three ways.")
print()
print("The M1 chart has 240 decisions to make and 240 spreads to pay. The H1")
print("chart has 4. Neither is more 'true' -- they are the same ticks grouped")
print("differently, and the grouping is a choice about how often you act.")
print()
print("The biggest-range figure is the one worth noticing: a move that looks")
print("dramatic on M1 is a small part of one H1 candle. 'Volatility' as an")
print("eye impression is mostly a statement about the timeframe you chose.")

You should see: 240 bars, 16 bars and 4 bars from one set of prices:

M1    bars  240   bullish  117 (49%)   biggest range 0.30
M15   bars   16   bullish    7 (44%)   biggest range 1.14
H1    bars    4   bullish    2 (50%)   biggest range 2.19

Identical price history. Four hours of it, three ways.

The M1 chart has 240 decisions to make and 240 spreads to pay. The H1
chart has 4. Neither is more 'true' -- they are the same ticks grouped
differently, and the grouping is a choice about how often you act.

The biggest-range figure is the one worth noticing: a move that looks
dramatic on M1 is a small part of one H1 candle. 'Volatility' as an
eye impression is mostly a statement about the timeframe you chose.

None of these is more real than the others. They are the same ticks grouped differently, so choosing a timeframe is not choosing a view of the market — it is choosing how often you act, and therefore how many spreads you pay.

The biggest-range column is the one worth sitting with. A 0.30 move fills an M1 candle and looks dramatic; the same move is a fraction of one H1 candle. Most of what people mean by “this market is volatile” is a statement about the timeframe they happen to be looking at, and zooming out is the cheapest way to test whether a move matters.

It also explains why the same strategy behaves differently on different timeframes even with identical settings: the edge per trade shrinks as the timeframe does, and the cost per trade does not.

If not: if all three rows show the same bar count, the grouping factor is not being applied — the loop steps by factor and takes a chunk of that size.

Go: the same folder.

Do: save this as bidask.py and run python3 bidask.py.

"""The chart draws one line. There are always two prices."""

QUOTES = [("EURUSD", 1.10248, 1.10251), ("GBPUSD", 1.27010, 1.27016),
          ("XAUUSD", 2412.15, 2412.48), ("US500",  5432.1,  5432.6)]

print("%-9s %10s %10s %9s" % ("SYMBOL", "BID", "ASK", "SPREAD"))
print("-" * 42)
for sym, bid, ask in QUOTES:
    print("%-9s %10.5f %10.5f %9.5f" % (sym, bid, ask, ask - bid))

print()
print("You BUY at the ask and SELL at the bid. So a position is negative the")
print("instant it opens, by exactly the spread.")
print()
sym, bid, ask = QUOTES[0]
print("open a buy on %s:" % sym)
print("   you pay        : %.5f  (the ask)" % ask)
print("   it is worth    : %.5f  (the bid, what you could sell at)" % bid)
print("   immediate P/L  : %+.5f per unit" % (bid - ask))
print()
print("MetaTrader draws BID by default, so the candle you are looking at is")
print("not the price you will pay to buy. Turn on 'Show Ask line' in the")
print("chart properties and you will see a second line above it.")
print()
print("This matters most for stops. A BUY is closed at the BID, so a stop")
print("below the bid line is hit when the visible chart reaches it. A SELL")
print("is closed at the ASK, which is above the line you can see -- so a sell")
print("stop is triggered before the candle appears to touch it.")
print()
print("Anyone who has said 'my stop was hit but the chart never got there'")
print("has met this, and the chart was not lying.")

You should see: two prices per instrument and a position that starts negative:

SYMBOL           BID        ASK    SPREAD
------------------------------------------
EURUSD       1.10248    1.10251   0.00003
GBPUSD       1.27010    1.27016   0.00006
XAUUSD    2412.15000 2412.48000   0.33000
US500     5432.10000 5432.60000   0.50000

You BUY at the ask and SELL at the bid. So a position is negative the
instant it opens, by exactly the spread.

open a buy on EURUSD:
   you pay        : 1.10251  (the ask)
   it is worth    : 1.10248  (the bid, what you could sell at)
   immediate P/L  : -0.00003 per unit

MetaTrader draws BID by default, so the candle you are looking at is
not the price you will pay to buy. Turn on 'Show Ask line' in the
chart properties and you will see a second line above it.

This matters most for stops. A BUY is closed at the BID, so a stop
below the bid line is hit when the visible chart reaches it. A SELL
is closed at the ASK, which is above the line you can see -- so a sell
stop is triggered before the candle appears to touch it.

Anyone who has said 'my stop was hit but the chart never got there'
has met this, and the chart was not lying.

There are always two prices: the bid, at which you can sell, and the ask, at which you can buy. MetaTrader draws the bid by default, so the candles you are looking at are not the prices you will pay to buy.

That single default explains one of the most common complaints in trading: “my stop was hit but the chart never got there”. A sell position is closed at the ask, which sits above the visible bid line by the spread — so its stop triggers before the candle appears to reach it. Nothing went wrong and the chart was not lying; it was drawing the other price.

Two things to do about it. Turn on the ask line in the chart's properties so you can see both, which takes seconds and permanently removes the confusion. And when placing a stop on a sell, allow for the spread — especially at the times step 3 of the backtesting page measured, when spreads widen to several times their usual size.

If not: these are fixed quotes and cannot vary. If the immediate profit prints as positive, the bid and ask were transposed — the ask is always the higher of the two.