What Is Trading? How It Works, With 33 Years of SPY Data
Michael Whitman
·
Trading since 2017, full-time since 2020
·
YouTube creator, 137K subscribers
·
Published 2026-09-05
·
Updated
·
Reviewed
Trading is buying something with the intention of selling it later at a different price. Every decision reduces to three prices chosen before entry - where you get in, where you are wrong, and where you are done - and to how much of the account rides on each one.
Trading is buying something to sell it later at a different price, or selling first to buy back lower, with the entry, the exit and the size decided before the order goes in.
What separates it from investing is the holding period, and that difference can be measured: on 33 years of SPY, 54.03% of single days ended higher, dividends included, while 81.84% of one-year holds did.
That one comparison explains most of what is hard about trading, and the rest of this page works through it with real prices.
How it works
Every trade has two sides. When you buy, someone sells to you at the price you both accept; when you sell, someone buys. The chart reading page shows how those agreed prices become the candles on a screen.
You can be long or short. Long means you bought and want the price to rise. Short means you sold something you borrowed and want to buy it back cheaper, which the short selling page walks through.
The order goes through a broker, who routes it to an exchange or a market maker. On a stock or ETF you pay the gap between the bid and the ask on every round trip, which the bid-ask spread page measures.
The mechanics are the same in every market. Stocks, forex, crypto and futures differ in their hours, costs, leverage and how far they travel in a day, not in how a buy or a sell works.
Trading vs investing: the holding period changes the odds
An investor buys a share of a business and holds it for years. A trader holds for minutes, days or weeks and needs the price to move the right way inside that window. The trading vs investing page sets the two side by side.
The shorter the window, the closer the market is to a coin toss. Across every SPY window since 1993, a single day ended higher 54.03% of the time. A week ended higher 58.68% of the time, a month 65.74%, a year 81.84% and ten years 92.04%.
That is a description of the past, not a forecast. It still says something useful: over one day, the market’s own drift is small next to the noise, so a short-term trade has to be right about direction for reasons of its own. The table and the chart are in the data section below.
The four decisions
Three prices
Entry is where you get in, the stop is the price that proves the idea wrong, and the target is where you are done. All three are chosen while you are flat, which is the only time they can be chosen calmly. The entry and exit page is the long version.
A trade without a stop is not a trade with more room. It is a position with no defined loss.
How much
The risk per trade formula is one line: money at risk divided by the distance to the stop gives the size. A run of five losses at 2% each leaves 90% of the account; at 20% each it leaves 33% and needs +205% to recover. Same trades, same reads; only the size differed. What a run like that does to judgment is the subject of the trading psychology page.
How fast
Costs scale with the number of trades. Say a rule pays a fixed 0.10 per trade in costs. Run fast, it averages a 0.15 gross gain per trade and costs take 67% of it; run slow, it averages 3.00 and costs take 3%.
Speed is therefore a cost decision, not a style. The scalping, day trading, swing trading and position trading pages each work through one speed.
Which market
Pick on structure: hours, costs, real volume, normal daily range and default leverage. In the five-year sample SPY and gold futures traded about 250 sessions a year, EUR/USD about 253 and bitcoin all 365.
Leverage does nothing to the chart and everything to the account, and the forex or stocks page runs that comparison in full.
A worked example
Take SPY’s close on Friday 30 October 2026: $755.72. Over the five years to 26 October 2026, its median day spanned 0.99% from high to low, which at this price is $7.51.
Put the stop one ordinary day away, at $748.21, and the target twice as far, at $770.74. Those three prices are written down before anything is bought.
Now the size. A $10,000 account risking 1% has $100 to lose. $100 divided by $7.51 is 13.3, so the order is 13 shares: a position of $9,824.36, or 98.24% of the account, with $97.63 at risk. If the target fills, the gain is $195.26.
The point is the gap between those two numbers. Almost the whole account is in the trade, yet a normal-sized loss costs under 1% of it, because the stop sets the risk and the position size follows.
Halve the stop distance and the same $100 buys 26 shares, which no longer fits in a $10,000 account without borrowing, and that is where leverage usually enters.
The original data
How often SPY ended higher, by holding period
Measured on Yahoo Finance adjusted closes (dividends included) for SPY, 29 January 1993 to 30 October 2026: 8,497 sessions, no bars dropped. A window starts at every close and ends a fixed number of sessions later, so the windows overlap; “higher” means the ending close is above the starting one.
Because neighboring windows share most of their days, the long rows are not independent: the ten-year row rests on only about three separate decades. On closes without dividends, 53.58% of days and 79.14% of one-year windows ended higher.
| Holding period | Sessions | Windows | Ended higher | Median change |
|---|---|---|---|---|
| 1 day | 1 | 8,496 | 54.03% | +0.07% |
| 1 week | 5 | 8,492 | 58.68% | +0.37% |
| 1 month | 21 | 8,476 | 65.74% | +1.44% |
| 3 months | 63 | 8,434 | 71.99% | +3.73% |
| 1 year | 252 | 8,245 | 81.84% | +14.45% |
| 5 years | 1,260 | 7,237 | 84.84% | +72.03% |
| 10 years | 2,520 | 5,977 | 92.04% | +131.83% |
The one-day row is the trader’s row. Only 26.38% of those 8,496 days moved 1% or more either way, and just 44 moved 5% or more. The biggest rise was +14.52% on 13 October 2008; the biggest fall, -10.94% on 16 March 2020.
A second source agrees. On 2,511 consecutive-session pairs from 31 October 2016 to 28 October 2026, FRED’s S&P 500 index rose on 54.80% of days and SPY’s unadjusted close on 55.04%, and the two moved the same way on 97.53% of days. The windows are in the holding-period table.
How far four markets travel in a day
Measured on Yahoo Finance daily bars from 27 October 2021 to 26 October 2026, the range is each day’s high minus its low as a percentage of the close. Days with no prices, days whose open or close sat outside their own high and low (34 on EUR/USD), and gold days with zero volume (8) were left out.
| Market | Days kept | Median range | Middle half of days | At 20x, median day = |
|---|---|---|---|---|
| Forex: EUR/USD | 1,264 | 0.55% | 0.40% to 0.76% | 10.98% of the account |
| Stocks: SPY | 1,254 | 0.99% | 0.69% to 1.44% | 19.86% of the account |
| Futures: gold, front month | 1,249 | 1.45% | 1.06% to 2.01% | 28.99% of the account |
| Crypto: bitcoin, weekends included | 1,826 | 3.11% | 2.00% to 4.55% | 62.19% of the account |
A median bitcoin day traveled 5.66 times as far as a median EUR/USD day. Nasdaq’s SPY bars give the same 0.99% median over the same 1,254 days, and the full rows are in the daily range table.
At 20 times leverage, the range alone covered at least a third of the account on 12 days in 1,264 on EUR/USD, 235 on SPY, 475 on gold and 1,534 in 1,826 on bitcoin.
When it fails
There was no stop
The most common way a plan stops being a plan, and it happens after entry rather than before.
The size was decided by the opportunity
“This one looks good, so I will take more” is how an ordinary losing run becomes a permanent problem.
The leverage was copied from another market
A setting that leaves room on EUR/USD leaves almost none on bitcoin, because the median bitcoin day is more than five times wider. Leverage chosen without looking at the range is chosen blind.
You traded through the quiet
The S&P 500 went 1,897 trading days without a new record close after January 1973, so a signal that waited for fresh highs would have sat silent until July 1980. Waiting is part of the job.
You judged it from the finished chart
Every chart in every explanation has its answer already printed. The live one does not. A hundred trades pins a win rate to about ten percentage points either way; twenty-five pins it to twenty, which is why the trading journal records the decision, not just the result.
Questions people ask
What is trading in simple terms?
Buying something to sell it later at a different price, or selling first and buying back, with the entry, the stop and the size written down before the order goes in.
Can I start trading with $100?
The arithmetic is the same at any size. With $100 and 1% at risk, the SPY example above risks $1 with a stop $7.51 away, which is 0.13 of a share; that needs a broker offering fractional shares, and a fixed fee per trade would take a large bite.
Is $1,000 a day from trading realistic?
On a $10,000 account that is a 10% gain in one day. SPY closed up 10% or more on 3 of 8,496 sessions since 1993: 13 and 28 October 2008 and 9 April 2025. Anything that promises it routinely is using leverage, and leverage multiplies the losing days by the same amount.
Is trading like gambling?
Over one day, SPY ended higher 54.03% of the time since 1993, dividends included, not far from a coin toss. What makes trading different is the plan around each bet: a written stop, a size set from it, and a record long enough to tell skill from noise.
Related
Chart reading is where to start: an empty chart to a decision, in four steps.
Risk per trade is the formula behind the decision that matters most.
And the trading journal is how you find out whether any of it is working.
The decisions worth agonizing over are the ones I make while I am flat. Once I am in a position, every option available is worse than the one that existed before entry, and most bad results come from making a new decision after the position exists.
— Michael Whitman
This page is educational, not financial advice. Test every idea on your own charts before risking money.