A trade you have an even chance of getting right is not a break-even bet. It loses money, and you can work out roughly how much before you place it.
Here is where the money goes. At any moment a share has two prices rather than one: the lowest anyone will sell at, and the highest anyone will buy at. The gap between them is how the firms standing in the middle earn their keep. When you buy you pay the seller’s number, and when you sell you accept the buyer’s number, so you cross that gap twice. Then add what the broker charges, and in India a tax levied on the transaction itself. None of it is much on any single trade. All of it is certain, which is more than can be said for the reason you traded.
So selling one holding to buy another is not a bet that the second beats the first. It is a bet that it beats it by more than the gap and the fees put together. Be right about the direction half the time and you still end the year behind.
That puts the interesting question somewhere I did not expect. When people argue about whether better information makes better investors, they argue about quality: is the fact true, did I get it early, do I understand it. The arithmetic above is indifferent to all of that. What it responds to is how many times you act.
There is real research on this, though it is older than the phone in your pocket. Brad Barber and Terrance Odean went through the accounts of tens of thousands of customers at a large American discount broker in the 1990s, and found that the households which traded most actively ended up with the worst returns after costs. It was not that the busy traders picked badly; their choices before costs were not far off everyone else’s. The trading itself consumed the difference. The same pair later found that investors who moved to online trading traded more after the switch, and did worse than before it. That is one country, one broker’s customers, one era before smartphones. But the direction is hard to wave away.
Which leaves the psychological half: why does more information produce more action rather than better? Part of it, I think, is that a new fact never arrives on its own. It comes with a question attached, and the question is never “is this any of your business.” It is “what are you going to do about it.” Look at the screen where the fact is delivered. There is a button to buy and one to sell, and no control anywhere that means nothing has changed. The interface has an opinion about what facts are for.
Here is what seems new to me, and this is reasoning rather than evidence. Until recently a feed handed you the raw fact and left you to build the interpretation yourself. That was work. Turning “the rupee slipped today” into “so I should sell this particular thing I own” took effort, and the effort was friction — most facts died there, because nobody can sustain it forty times a day. The step is now free. Ask a model what a headline means for what you hold and it will tell you in seconds, fluently, with reasons that hang together. What it will almost never say is that the honest answer is nothing, this is noise, come back next year. Answering is what it is for. So the supply of plausible reasons to act has gone from scarce to effectively unlimited, which is a strange thing to have happened to a decision whose costs scale with how often you make it.
Now the best case against me, and I find it genuinely strong. Technology has already made doing nothing far easier than it used to be. A monthly amount that leaves your account automatically and buys a broad fund is engineering whose whole purpose is to remove decisions, and it worked — vast sums now sit in arrangements that ask their owner to do nothing at all. Same industry, same tooling, same decade. So “better tools make people busier” is plainly not a law of nature, and my worry may describe one product category rather than a direction of travel.
Holding both, here is where I come out. What decides which one a person meets is not the technology. It is who is paying for it. A business that takes a slice of each transaction needs its users to keep finding reasons; a business paid a fee on money held needs them to sit still. The same model, with the same capabilities, will write either the reason or the reassurance depending on which one the invoice rewards.
So, my answer: faster information does not mostly make people better or worse investors. It makes them more decisive, and decisiveness is not accuracy — it is frequency, and frequency has a price list. What follows is a test I would apply to anything that talks to people about money: is this thing capable of saying that nothing has happened? Not whether it is accurate, not whether it is fast. Whether the sentence “there is nothing here for you today” is one it can produce at all. A tool that cannot say that will say the other thing instead, every hour, for as long as you keep asking.