Before any indicator, you need to be precise about what a chart actually is, because several quiet choices are baked into it and each one changes the analysis.

• Open — the first traded price of the period. In markets with an opening auction, this is an auction price and can be far from the previous close.
• High and low — the extremes. Note that these are traded prices, so a single small trade at an extreme creates a wick identical to one produced by enormous volume. This is why a wick alone tells you very little.
• Close — the last traded price, or in most markets the closing auction price. The close is the most information-dense of the four, because it is where the market settled after the day's argument, and it is the price against which most participants are benchmarked. This is why so many rules are written on closes rather than on intrabar prices, and why a rule that triggers intrabar is a different rule from one that triggers on a close.

The same instrument on a 5-minute chart, a daily chart and a weekly chart tells three different stories, and none of them is more true than the others. What actually changes:

The relationship that matters: as you go shorter, the ratio of your trading costs to the size of the move you are trying to capture gets worse, and it gets worse faster than the number of opportunities improves. This single fact kills most retail intraday strategies, and it is arithmetic rather than opinion. A 5-minute strategy capturing 0.15% per trade against a 0.05% round-trip cost has given away a third of its gross edge before it starts.

• Candlestick — the default, and the right one for almost everything, because it shows all four numbers and the relationship between open and close.
• Line on close — underrated. Removes wick noise entirely and makes structure much easier to see. If you are unsure whether a trend is real, look at the close line; a great deal of apparent complexity disappears.
• Heikin-Ashi — averages the candles so trends look cleaner. Useful for the eye and dangerous for rules, because the displayed open, high, low and close are not prices you can trade at. A backtest that enters at a Heikin-Ashi open is entering at a price that never existed. This mistake is extremely common.
• Renko and point-and-figure — remove time entirely and plot only price movement. Genuinely useful for seeing structure; the same caution applies about backtesting on synthetic bars.
• Log versus linear scale — on any multi-year chart, use log. A trendline drawn on a linear scale over ten years is meaningless because equal vertical distances represent wildly different percentage moves. This is one of the most common errors in published long-term analysis.

Two silent data issues change what you are looking at, and both were covered in the market data chapter of the previous book because they matter even more there:

• Corporate-action adjustment. A split- or bonus-adjusted series rewrites history. It is correct for computing returns and wrong for identifying the price levels people actually traded at — a support level at ₹1,000 pre-split appears at ₹500 on an adjusted chart, where no one ever defended it.
• Futures continuation. A continuous futures chart stitches successive contracts together, and the stitch point has a discontinuity. Back-adjusted series remove the jump by shifting all history — which means old prices on the chart are not prices that ever traded, and can even go negative on some long crude oil series. For structure this is usually acceptable; for anything level-based it is not.