I read through similar questions here, but feel they did not solve my confusion. I've been considering the prospects of day trading with my money verses investing. My confusion is

If we neglect broker fees, I'm assuming day trading (not investing) is zero sum. If someone were to trade randomly (without experience), it appears they'd win half the time and lose half the time. However day trading winners win at loser's expense and there is a small population of absurdly profitable day trading winners (some examples are given in Market Wizards by Schwager). These two ideas seem in direct contradiction.

I'll elaborate some. Although investing is not zero-sum as the market's value grows over a long period of time, buying and selling quickly is zero sum as the stock is only held a short time.

The Intelligent Investor by Graham and some people advises that although investing is great, day trading is a losing game for most. Graham attributes these losses to broker fees, market impact, and taxes. But even without them, could day trading still be a losing game for most? Particularly given the random nature of the market and the supposedly random betting of most people.

Edit: I don't quite understand the negative votes and perhaps you could help me. I tried my best to define the question well by highlighting it and tried providing where the question came from. The following seemed legitimate to me:

In a zero sum world where inexperienced traders bet randomly, how could the elite traders possibly take money from amateurs?

  • 1
    "If someone were to trade randomly"... but they don't trade randomly. It's always some Scheme.
    – RonJohn
    Commented Jun 20, 2019 at 16:35
  • "it appears they'd win half the time and lose half the time" - ah, probability is your problem there, for the same reason the following statement is problematic: if you play the lottery you can either win or you lose, two outcomes, so there's a 50% chance of winning! The problem is that there are many, many, many ways to lose (especially when you add in transaction fees, because zero change in stock price between buy-sell means you lose), and far fewer ways to win. Counting in probability is a tricky problem. That, and you can't ignore transaction costs, it turns zero sum to negative sum.
    – BrianH
    Commented Jun 20, 2019 at 23:27

6 Answers 6


Most day trading is based on some sort of technical analysis. You're looking for particular patterns in stock prices that provide a signal to buy or sell. The problem is that there are lots of very well capitalized firms that run very highly optimized algorithms to do high frequency trading based on technical analysis. These firms are getting the data with extremely low latency (i.e. they're spending lots of cash to get data microseconds earlier), they're able to process it faster than the human day trader, they're spending lots of money developing and testing their algorithms, and they're trading without human emotion so they're not tempted to do something stupid like making a big bet in the hopes of breaking even if they've had a run of bad luck. When you have a competition between a multi-billion dollar Wall Street firm and some guy in his basement, Wall Street is going to win well over half the time.

Then when you factor in the costs of day trading (financial, time, and stress), it is rare that someone comes out ahead. And it's extremely rare that someone comes out ahead consistently rather than simply being lucky. If you have 1024 traders each of which will win on 50% of the days and lose on 50% of the days, and you let them trade for 2 weeks (M-F), you'd expect that 1 of them would win 10 straight days. That particular trader will likely interpret that as an indication that he's really good at day trading when the reality is that he was simply the lucky one.

  • I think that you've conflated traditional day trading and high frequency trading as well as other issues At times, HFT has accounted for as much as 70% of market trading and it is not based on technical analysis. It uses algorithms that take advantage of NBBO, utilize collocation speed edge, etc. to achieve small arbs. Day traders are not competing against HFT. HFT is skimming from everyone. Commented Jun 20, 2019 at 18:24
  • @BobBaerker - You could certainly separate some of this into "high frequency trading" vs "algorithmic trading" but there is a lot of grey area between the two. High frequency traders doing arbitrage aren't the ones day traders are competing against, algorithmic traders are. But the same problems exist-- you're comparing a solitary human against a giant firm that can build a bot that responds almost instantly to whatever signal you want to define. Even if that means that they're holding positions for minutes or hours rather than milliseconds, they're able to front-run the day trader. Commented Jun 20, 2019 at 19:23

Even if day trading is a perfectly zero-sum game - by which I mean than every stock you buy has an equal chance of increasing or decreasing in price, and there are no trading costs - you're still facing the "gambler's ruin" problem: https://en.wikipedia.org/wiki/Gambler%27s_ruin That is, you have a finite amount of money to invest, while the rest of the market has an infinite (in reality, just very large) amount. Eventually you will experience a string of losses that wipes out your capital.


One Yes, it can be a losing game depending on human nature right? If you double down on losers too often or if you are too greedy and you lose your gains.

Two You can lose money by investing in individual stocks which perform badly.If you were just day trading a sector which is in decline, you would lose money over time as it declined (small loses each day of trading).

Three It is also hard to predict what will happen with a company on a day to day basis especially because macro factors affect stocks more than the companies themselves on a day to day basis.

I suppose if you actually randomly traded stocks and held all of factors constant then maybe you could get to zero sum. The loses come when you change things up in order to try and make profit.

  • 1
    But you can only get to zero sum in this question’s fantasy world where there are no broker fees, bid/offer spreads or taxes. In the real world, the transaction costs of frequent trading are much larger than the average market gain on the same timescale, and so most day traders lose money and a random strategy is pretty much guaranteed to lose money.
    – Mike Scott
    Commented Jun 20, 2019 at 16:26
  • 1
    @MikeScott Taxes aren't relevant, if you hit zero-sum you'd have zero tax, but good points otherwise.
    – Hart CO
    Commented Jun 20, 2019 at 16:29
  • @HartCO Zero sum doesn’t mean zero every day. You might make some money in one tax period and lose some in another, and depending on your jurisdiction’s tax laws you might then be liable for tax on the money that you made without being able to offset your losses from the different period.
    – Mike Scott
    Commented Jun 20, 2019 at 16:38
  • @MikeScott Of course, but the US and many other countries tax based on annual gains. In the US wash-sale disallowed losses affect basis for currently held shares. I suppose If there's a country that ignores losses or collects taxes at very short intervals then taxes would be relevant, I'm not familiar with such a country.
    – Hart CO
    Commented Jun 20, 2019 at 17:07
  • 1
    @HartCO And also, for example, in the UK stamp duty of 0.5% is payable on the purchase of shares. Not all taxes are based on profits, some are pure transaction costs.
    – Mike Scott
    Commented Jun 20, 2019 at 17:12

Lets imagine that you pick them at random, and sell them x minutes later also at random. Note that x could be single digit minutes or hundreds of minutes later, and you do dozens of transactions each day. Each pair of transactions might make or lose money. If everything is truly random, then over the long term you would match the motions of the overall market. If the market for the year was up, you would also be up. If the market was down then you would also trend down.

Of course somebody would also be able to get the same result by investing in a broad index fund. That index fund would have low expenses, and is often tax efficient by minimizing capital gains.

The day trader expended time and money chasing the average. Of course if your method of picking wasn't random, but was a flawed method then you could be tricked into bad choices.

Once you factor in the cost of each transaction, then your random method would fall behind the general market. That is the equivalent of what the house takes at the casino.

  • Yes, if it's much more than a finite sample and the results are truly random then you achieve the market performance less the additional slippage. The only way to outperform is to have an edge and/or practice disciplined risk management. Commented Jun 20, 2019 at 17:28

I believe you're overthinking this. Imagine for a moment that stock movements are completely random and there are no fees. Then you can expect that day traders will achieve a fairly standard bell curve, with the peak at 0. Many people will break even, many will gain or lose just a little, and a few will gain or lose a lot.

Now if you add fees you simply shift the entire curve slightly to the left, making the average a slight loss rather than 0.

Many people believe that there is an element of skill in day trading, and if so, this would help determine where on the curve you land (the more skillful you are, the better your chances of landing further to the right), but the overall curve doesn't really change.


AFAIC, the stock market is zero sum.

For every buyer, there is a seller and the amount of money involved remains the same. Money just changes hands, regardless of what happens to the price of the stock. Here's a simple scenario:

There are 3 people (A, B, and C). Each one has $100 and A owns the stock. That’s $300 in existence. Let’s assume no slippage or transactional costs.

B buys the stock from A for $60. Now B has the stock. A has $160, B has $40 and the stock, and C has $100. That’s still $300 in existence.

If B now sells the stock to C for $80 then A has $160, B has $120, and C has $20 and the stock. the stock appreciated $20 and yet there's still only $300 in existence.

Now suppose the company goes bankrupt and the stock is delisted. A still has $160, B still has $120, and C still has $20 but no stock. Again, no money vanished and $300 still exists.

When share price rises significantly, paper wealth is created (not money). When share price drops significantly, paper wealth is wiped out (not money).

It's not an easy concept to swallow but when you buy a stock, your money is lost (someone else now has it). If share price appreciates and you sell, you get your money back and then some. That's where the tree analogy in your link fails. The author wrote:

If I sell that tree to a sawmill, they can cut the tree into usable lumber, and sell that lumber at a profit. They added their efforts and increased the value. A carpenter can increase the value even further by making something useful (a door, for example). A retail store can make that door more useful by transporting it to a location with a buyer, and a builder can make it even more useful by installing it on a house.

No one lost any money in any of these transactions. They bought something valuable, and made it more valuable by adding their effort.

The last person in the sequence lost his money when he paid for the final product (the house) and it is gone until someone else gives up his money for the house. If the house burns down and the home owner has no insurance, it's a total loss. In this tree to house example, the money supply also remained constant.

  • No, it’s not zero sum, because shares pay dividends. That $300-worth of shares will generate another $15-ish of dividends every year that they are held. Sell out after a year and you have turned $300 into perhaps $315.
    – Mike Scott
    Commented Jun 20, 2019 at 17:15
  • 1
    Dividends don't matter, earnings do. A stock price goes up when it earns money, and down when it pays dividends. A firm that pays no dividends can still grow in value due to their earnings (or the prospect of future earnings). But I agree, on the timescale of hours this doesn't make a difference and you can consider the market as zero-sum.
    – Earth
    Commented Jun 20, 2019 at 17:20
  • 1
    @Mike Scott - First off, it's $300 worth of cash and a stock that hit a high of $80 in the example. So it pays nearly a 20% yield? Nice! OK, apart from that gaffe, so because the shareholder received a $15 dividend, $15 was created out of thin air and the US money supply increased by that times the outstanding float? I think not. The dividend came corporate earnings which in turn came from customer purchases ad infinitum. Sorry, but all you did was complicate the example with more variables not refute it. Commented Jun 20, 2019 at 17:22

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .