Technical Analysis for Beginners: What Charts Can and Cannot Tell You

A daily price bar is five numbers: the open, the high, the low, the close, and the shares that changed hands. Everything else is arithmetic on those five. Technical analysis is the practice of choosing which arithmetic to perform and then being strict about what the answer is allowed to mean, which is the harder half of the job and the half that beginners skip.

Start with what is in the file

There is no sixth number hiding behind the drawing. A 200-day moving average, an RSI reading, a MACD histogram and a Bollinger Band are all computed from the same column of closing prices, and each one rearranges that column into a shape the eye reads faster than a list of numbers. A rearrangement can make a feature easier to see. It cannot put information into data that never held it.

Sit with that, because it explains a mistake nearly everyone makes. When four indicators agree, the feeling is corroboration. The accurate description is that one series was transformed four ways and the four answers came out pointing the same direction, which they were always going to do.

Two things the chart genuinely offers. Exchange price and volume data is the same quality for every listed security, while fundamental information varies enormously between a mega-cap with forty analysts and a small company with two. And the record updates with every trade. A decision made ten minutes ago sits in it alongside one made last quarter.

Trend, defined so that two people would agree

A swing high is a bar whose high stands above the bars on either side of it, and a swing low is the mirror image, a bar whose low sits under both of its neighbours. Those two definitions are enough to name the structure. They are also checkable, which most chart vocabulary is not.

Sequence of swing points	Description	Name
Higher highs and higher lows	Each advance clears the last, each pullback holds above the last	Uptrend
Lower highs and lower lows	Each advance stalls sooner, each decline goes further	Downtrend
Highs and lows roughly level	Price oscillates between two areas	Range

Trend belongs to a timeframe and to nothing else. A stock can show a daily downtrend inside a weekly uptrend, and both readings are correct, because they answer different questions about different windows of the same data. Pick the timeframe you intend to hold for. Take your context from the one above it.

Settle this first. Every other tool on the chart reverses meaning depending on the answer, so a reading taken before you know the trend is a reading you will end up interpreting to suit whatever you already wanted to do. A momentum reading of 75 in a sustained advance reports weeks of gains outweighing losses, while the same 75 at the top of a nine-month range reports that price has arrived where sellers have acted before. Same number, opposite implication. The indicator cannot tell you which of the two you are in.

Levels are a claim about orders nobody has filled yet

Support is an area where buying previously stopped a decline. Resistance is where selling stopped an advance. Both are records. The useful question about any line you draw is whether the orders that made it are still sitting there, because a level is worth exactly as much as the unfilled demand or supply behind it.

They form for reasons that have no bearing on the company. Traders who bought at $42 and watched a rally to $48 will consider $42 again, and anyone who bought at $48 and held all the way down spends the whole recovery waiting to get out at cost, which puts supply back at $48. Round numbers work the same way. People choose round numbers. More size sits at 50.00 than at 49.63.

A level earns its keep in one way. It gives you a specific price at which your reasoning was wrong. Say you buy at $43 because $42 held twice. Then $41.50 is your number: below it, the demand you were relying on has gone, and the reason you entered no longer exists. You can write that rule down before you enter. It works whether or not the level holds. The support and resistance guide covers how to draw zones wide enough to survive noise, and how to tell a false break from a real one.

Volume is the only series that is not price

Volume is the count of shares that changed hands. It is the one input on the chart that is not derived from price, which makes it the only genuinely independent second opinion the chart has to offer.

Raw share counts are useless across stocks. Use the ratio: relative volume = today's volume / average daily volume over the last 50 days. A stock that normally trades 1.4 million shares and trades 4.2 million gives 4.2 / 1.4 = 3.0. A breakout on three times normal participation involved a lot of people. The same breakout at 0.6 times normal involved almost nobody. That is usually why it gets given back.

Volume carries no direction inside it. Every share sold was bought by somebody, so a heavy down day is no evidence that sellers outnumbered buyers, only that the two sides disagreed strongly enough to move a lot of stock between them. The volume analysis guide works through climax volume, on-balance volume and VWAP.

One indicator, done by hand

The simple moving average is the mean of the last N closes. Recalculate it each day. Five days, in full:

Day	Close	Window sum	5-day SMA
1	48.20
2	49.10
3	48.60
4	50.40
5	51.20	48.20 + 49.10 + 48.60 + 50.40 + 51.20 = 247.50	247.50 / 5 = 49.50
6	52.00	247.50 - 48.20 + 52.00 = 251.30	251.30 / 5 = 50.26

That is the whole indicator. The oldest close leaves the window. The newest one enters. Divide by N. A 200-day average is the identical sum with 200 terms in it, which is why a single session moves it so little: one close is one two-hundredth of the total. It also lags price by roughly (N - 1) / 2 bars, which is about 100 bars for the 200-day average, so what you are reading is where price has been, at a delay you can calculate exactly.

Momentum is one step further on. The Relative Strength Index compares average gains with average losses over 14 periods, using RS = average gain / average loss and RSI = 100 - (100 / (1 + RS)). With an average gain of 0.80 and an average loss of 0.40, RS = 2.0 and RSI = 100 - (100 / 3) = 66.7. Gains outweighed losses two to one across the smoothed window. The reading contains nothing about valuation and nothing about tomorrow.

MACD subtracts one exponential moving average from another: MACD line = 12-period EMA - 26-period EMA. The signal line is a 9-period EMA of that difference. Both inputs are averages of the same closes, so MACD and a 12/26 moving average crossover are close relatives, which matters a great deal before you put both on one chart and describe their agreement as confirmation. The moving averages guide has the EMA arithmetic. The momentum guide has the rest of RSI.

The calculation tells you when the output is meaningless

This is the part that gets skipped. It is the part that pays.

RSI cannot exceed 100, and it approaches 100 whenever the average loss approaches zero. A stock with no down close in 14 sessions pins the oscillator near its ceiling, regardless of how small the daily gains were, because the denominator of the ratio has collapsed towards nothing. The reading is arithmetically correct and informationally empty.

MACD is unbounded and scales with price. A MACD line of 1.40 on a $50 stock describes a separation of 2.8% of price, and the same 1.40 on a $500 stock describes 0.28%. Comparing MACD values across two names without dividing by price compares nothing.

A 20-day standard deviation sets the width of a Bollinger Band. It keeps one violent session in its window for 20 sessions, so a stock can look twice as volatile as it is for a month after a single earnings gap, then look artificially calm on the day that session drops out. The average true range measure behaves differently. Both are worked in full in the volatility indicators guide.

None of those three failures is visible on the chart. They are visible in the formula.

Where the method fails, including the evidence

News gaps straight through every level. A stock that opens 14% lower on guidance did not pause at support, and a stop order resting there fills at whatever price the market reopens at.

Trend tools whipsaw in ranges by construction. A crossover system in a sideways market buys near the top of the range and sells near the bottom, repeatedly, paying a spread each time, and none of those trades was a lapse in judgement. Oscillators fail in the opposite conditions. Fade every overbought reading and you will short a strong advance the entire way up, adding to the position each time the reading gets more extreme.

Then there is the sample you are looking at. The formation that preceded a 20% rally is visible because the rally happened. The identical formation that went nowhere sits unlabelled in the same data. Nobody screenshots that one. Confidence built from examples found after the fact is confidence in a sample selected on its outcome, and staring at more of those examples will not correct it.

Which brings up the statistics. Pattern and candle success rates circulate in enormous volume without a sample, a date range, a qualifying test or a scored outcome behind any of them. Take a figure such as “this formation works 70% of the time”. Nobody can check it, including whoever published it. This site does not repeat those numbers. A pattern can be described completely by its break level, its projection and the price at which it has failed, and those are facts about geometry rather than claims about the future.

The order I work in

The order is fixed on purpose. Shopping for the one indicator that agrees with what you already wanted to do is the default human behaviour here, and a fixed sequence is the cheapest defence against it.

Step	Question	Tool
1	What is the trend on the daily chart?	Swing points, slope of the 200-day average
2	Where are the levels above and below?	Prior swing highs and lows, round numbers
3	How many people are taking part?	Relative volume against the 50-day average
4	What does one momentum reading say, given step 1?	RSI or MACD
5	Where am I wrong, and what does that cost?	Invalidation price, then share count

Steps 1 to 4 are a filtering problem, which is what the stock screener is for: it sorts on RSI, position relative to the moving averages, distance from the 52-week high and relative volume together. The candlestick scanner locates the single-bar and two-bar formations so you can inspect where each one actually printed, and market breadth answers the question no individual chart can answer, which is how many stocks are taking part in an index move.

Only step 5 works whether or not the chart was right. Read how to read candlestick charts next for what a single bar records, then the chart patterns guide for measured moves and invalidation levels, and use the technical analysis basics quiz to find the parts that did not stick.