Would a crowd come up with better predictions than individuals?


I’m Brian Eastlake, new to the forums.

I was talking with a friend of mine a few weeks ago about trading. He doesn’t think that future price action can be assigned a probability based upon past price action. ie, he doesn’t believe in technical analysis whatsoever, and is somewhat suspicious of fundamental analysis too. Basically, he thinks price action is random.

I disagreed. I think both fundamental and technical analysis have their place - fundamental forces drive the higher timeframes, and I can clearly see technical patterns in price action, I’m just not always right about which pattern is dominating at any particular point in time. :slight_smile:

My friend - Will - and I are both also interested in psychology, and we’ve previously discussed the phenomenon of how crowd aggregated predictions tend to be better than the predictions made by individuals (even experts).

You may have heard about the jellybeans-in-a-jar contest, where you have to guess the number of beans and the person with the closest guess wins the jar. If so, you may also have heard that the average of all the guesses is [B]remarkably[/B] accurate, and consistently so.

Will pointed out that if price action wasn’t random, and future price action could be assigned a probability based upon past price action and/or fundamental analysis, that crowds should be able to consistently predict future price action better than most individuals.

So, what do you think? Would a crowd come up with better predictions than individuals?


Read this: StockTwits Research (@research) | StockTwits

1 Like

The same is true in the world of horse racing. Over the course of a race meet the horse most favored by the crowd in a 10 horse field (the horse with the lowest odds at post time) wins more often than any other betting choice. The horse with the second lowest odds wins more often than the third betting choice which wins more often than the fourth, etc.

Picking the winner in any single race? Your guess is as good as the next guy’s. Over thousands of races, however, the crowd’s choice for favorite wins about 33% of the time. The crowds second choice wins 20%, third choice wins 14%… the crowd selects the winner from the its top three collective choices 67% of the time. These stats have held up year after year for well over a over a century.

Broad market index funds that follow the S&P 500 or DJIA, for example, are really nothing more than aggregate crowd predictions and year after year these funds beat the returns of the typical managed mutual fund. Again, the crowd out-predicts the individual.

Can a large enough crowd come up with better predictions than most of the individuals that comprise that crowd in almost anything? Absolutely.

Why? I have no idea. They just do.

Awesome, thanks for the link! :slight_smile:

That was a fascinating read. I guess that particular style of prediction wouldn’t work with forex, because it’s impossible to know what someone would be referring to, whereas if someone refers to Microsoft then it’s clear what it’s about.

1 Like

Interesting, I hadn’t thought of it from the horse racing perspective.

So if price action isn’t random, and there are probabilities for the different outcomes at any point in time (and for a given timeframe, of course), then if you got a bunch of people to make a prediction, that single prediction may not be accurate, but with enough samples you’d expect to see the crowd’s prediction would be a lot more accurate than an individual’s prediction.

So, you think that ultimately the crowd’s prediction would tend to be more accurate than an individual’s, even though there would be times when the crowd’s prediction is simply wrong?

I’ve read the book by James Surowiecki called “The Wisdom Of Crowds”, in which he goes into a lot of detail about why/how. He identified 4 things that make a crowd “smart”. Basically it’s:

  1. Diversity: the members in the crowd must be from a range of backgrounds, ideally including both experts and non-experts, and have access to private information.
  2. Independence: the individuals in the crowd each have to use their own judgement to come to a personal decision, without being influenced by the other members.
  3. Decentralisation: the range of experience of the members should be decentralised. (This is the key to better diversity.)
  4. Aggregation: there must be some mechanism for converting the individual decision into a group decision.

If any of those criteria aren’t met, then the crowd tends not to come up with very good predictions. But when they are met… that’s when the crowd becomes “smart” (or “wise”), and does sensational things.

There are also certain types of problems that crowds are good at solving, and others that they aren’t so good at. For example, it turns out that crowds aren’t very good at creativity/innovation, but are really good at information processing (including estimating, like with the number of jellybeans in a jar, or in the case of horse racing: estimating probabilities).

The book is a fascinating read, if you’re interested, but the Wikipedia article gives a good summary.

The part that has me puzzled (and interested), is whether or not it would work for forex. If price action is too chaotic to make predictions for the next bar, then a crowd wouldn’t be able to make a worthwhile prediction. But, if price action is so chaotic that you can’t assign a probability to what will happen next, I’m not sure I should trade at all, because I don’t think I’m smarter than the crowd. :20:


Interesting discussion. I recommend a read of Flash boys.

The problem here is we are not recognizing the scale of the auction that is FX.

This so called no central exchange environment is actually getting centralized everyday by the HFT firms who will pay brokers to send them their order flow to trade against. We are fair game, whether we make money or not no one cares. They just need to have the order flow.

This can obviously lead many to believe price action is random but it is not. To retail it is but in the FX option world it sure is not when you have 1bn Euros worth at particular price level. The lack of understanding of markets lead to many erroneous claims. HFT is the reason for the deep liquidity even banks open their dark pools to HFT. IEX in stocks are trying to do something but FX is a free for all. So with prices been moved at the milli and micro second of course it will look random but it is the noisy in between the big futures and options positions. Why do you think COT and vanilla options are what guide institutional trades?

We all need more education.

And pays less than 3 to 1 making betting with the crowd a dumb strategy

The more I think about this, the more I’m certain I don’t want to bet on the horses. :20:

If the crowd is more likely to be right than anybody else (although on any given race it may not be right), then deviating from the crowd’s prediction comes with a higher risk of failure. The only reason I can imagine you’d want to do that is if you had some kind of information that the rest of the crowd didn’t know - maybe about the health/fitness of the horses, or similar.

Not having any particular knowledge about horses, I am well aware that my own predictions in horse races would certainly be worse than the crowd’s over the long term. I guess that’s why they call it gambling.

But, if I were to go with the crowd’s prediction, I’d also lose money, because of the odds, as you’ve pointed out.

The same isn’t necessarily true for trading, but I think a large part of the equation for success in trading comes down to how accurately you can assess the probabilities of possible future price action (in whatever timeframe you’re looking at).

For example, if you had a mechanism that was 60% accurate at making predictions, and that mechanism said the market would go up in the next week, then you’d have a tradeable edge to work with.

Now think of that in terms of the crowd. If you got a smart crowd, and had them make a prediction about what the market would do over the following hour/day/week/whatever timeframe, I’m wondering whether they would be able to uncover that probability. If, that is, the probability is there to find.

Will thinks the idea is silly, because he doesn’t agree with the underlying principle that possible future price action can be assigned probabilities. He thinks it all unfolds chaotically, and that we’re only looking at price action in retrospect and applying patterns to it after-the-fact.


Hey, thanks for the reference. I’d never heard of the IEX and wasn’t aware of the extent to which HFT had pervaded the markets. I read the Wikipedia articles for IEX and Flash Boys, and I’ll probably read the book when I have some spare time.

I don’t want to even try to compete on that type of timescale. It’s utterly dominated by people with insane amount of resources to throw at the problem.

Sure, if you can peek behind the curtain at the order flow then the price movements would probably make a lot more sense, but without the benefit of that insight then the really low timescale movements do start to look random (I have to admit that). Personally, it’s only on the higher timeframes that I can start to make sense of what I’m looking at. I see patterns in weekly charts, daily charts, and even hourly charts. But when I get down to 15-minute and below… there may be patterns but I have to squint to see them and there are a lot of what I’ll call “artefacts” that distort those patterns. It makes it a lot harder to assign probabilities to future price action.

I prefer starting high, and zooming in. So I’ll get an idea of what I think is happening on a monthly/weekly timeframe, for context, then take a look at the daily, 4-hourly, and hourly. For entries I might dip into the lower timeframes, but I don’t actually trade on them - I personally find that kind of timeframe too stressful. I prefer taking smaller positions (I’m not as leveraged as a lot of traders), and give the market a bit of room to move. Any particular trade may go against me, but if I can improve my ability to make predictions based upon the patterns and/or fundamental-data, then my trading improves. Because over the longer term, it’s my ability to predict what will happen next that determines how successful I am. Probably this is true of all traders, regardless of the timeframe(s) they work with…?