Let's say that I want to buy a stock like MOMO and I want to buy it today (this decision was made before the market open). The duration of my position will be between a week and several month. When should I buy the stock so that I'm nearly sure that I won't lost money in that day (because of speculation).

  • 2
    It should not make a difference. You don't lose money that day.
    – Ross
    Commented Aug 24, 2016 at 15:50
  • 18
    There is no way to predict this, unless you believe that you can beat the pros in day-trading. And you believe that, I have a bridge you might be interested in buying...
    – keshlam
    Commented Aug 24, 2016 at 15:54
  • 1
    I think it would be best to buy early in the day. This will reduce your average time holding the investment before your objectives are meet. Keep in mind there are more efficient ways to lose money.
    – Pete B.
    Commented Aug 24, 2016 at 16:21
  • 5
    Oh, I know the answer. "Before you get drunk." If you wait until after you're drunk, the odds of making an error go up dramatically. Commented Aug 24, 2016 at 20:32
  • 3
    @keshlam: What time of day should I buy the bridge?
    – BrenBarn
    Commented Aug 25, 2016 at 5:39

6 Answers 6


The best time to buy a stock is the time of day when the stock price is lowest!

Obviously you learned nothing from that sentence, but unfortunately you won't get a much better answer than that. Here's a question that is very similar to yours: "Is it better to have a picnic for lunch or for dinner to minimize the chance of getting rained out?" Every day is different...

  • 1
    I'd add to this that some research could be done. I like the picnic scenario along with it. Historically, does the rain come more often around lunch or around dinner? You can look through the history of the stock and see if it typically falls later in the day compared to opening. It's never a guarantee, but could at least be a slightly more educated decision.
    – BobbyScon
    Commented Aug 24, 2016 at 17:09
  • 3
    @BobbyScon - I doubt that information would help very much, but I admit it probably couldn't hurt to investigate it to see if there's a pattern that is anything more than statistical noise.
    – TTT
    Commented Aug 24, 2016 at 17:30
  • I completely agree. Some stocks I own have fairly regular patterns as far as trading higher before noon, but typically by fractions of a percent. Certainly not something I'd bet against, but at least it might make you feel better about committing to a time to purchase.
    – BobbyScon
    Commented Aug 24, 2016 at 17:48
  • 1
    @user2023861 Every day is different.
    – Xalorous
    Commented Aug 24, 2016 at 22:22
  • 1
    @Victor - You can't know when it will be the "lowest" i.e. "the best time to buy", that's the point.
    – TTT
    Commented Aug 26, 2016 at 16:19

You can't predict when to buy a stock during the day to guarantee not having a loss for the day. In the short run stock prices are really pretty random. There are many day traders who try to accomplish exactly this and most of them lose money.

If you don't believe me, create an account on Investopedia and use their free stock market simulator and try day trading for a few months.


One of the biggest laws in economics is that if an opportunity is very profitable and is very easily exploitable even by complete beginners, then it will very soon stop being profitable.

That's how the market works. If you buy stock when it is at the lowest, then you are making money, but most of the time someone else is losing money. And if there was a magic hour of the day when buying would be the most profitable, then soon everybody would want to buy at that time and no one would want to sell anything, so the scheme would collapse.

  • 2
    This isn't entirely accurate. The market as a whole is generally increasing over time, so there do not have to be losers. Of course there are, but it isn't a necessary condition. You can imagine a hypothetical stock which averages an increase of exactly 0.01% per day (which is pretty close to the US stock market average since 1900 after adjusting for inflation). In that case no matter when you buy, you'll always earn a modest amount, and there would be no losers.
    – TTT
    Commented Aug 24, 2016 at 22:10
  • 1
    @TTT : I agree, but the question was about a hypothetical "best time of the day" (to by stocks at a lower value than at any other time of the day), which, if existed, would be used by so may people that it would cease to be the best time of the day.
    – vsz
    Commented Aug 25, 2016 at 4:24
  • I agree with the short life of a possible best time of day, I was only disagreeing with the part of your comment that said it must be true that "someone else is losing money". Sorry for the confusion.
    – TTT
    Commented Aug 25, 2016 at 13:54
  • 1
    @TTT While not technically accurate, I think it is generally appropriate to say that in day trading, winners and losers cancel out. The value of the underlying companies changes slowly over time, as corporations use investor funds to increase their businesses, leading to a net economic gain. I don't think that concept is particularly relevant to day trading, where buyers and sellers are more strictly at odds. Commented Aug 25, 2016 at 17:28
  • 1
    @Grade'Eh'Bacon - I'll buy that. :D
    – TTT
    Commented Aug 25, 2016 at 19:09

The best thing to do is not worry about what time is best to buy but put in a conditional order before the market opens. If your conditions are met during the trading day your order will go through and you will buy the shares.

This keeps your emotions out of your trading and will stop you from either chasing the market or buying when you consider the wrong time.

As you have already done your analysis and made your decision before market open, thus you should place your conditional orders and stop losses before market opens as well.


Buy it at the close. That way you won't lose money (even if marked to market) on the day.

  • If you are the type of person that checks your stock quotes every 5 minutes, then buying at market close is also a good idea because you'll be able to get more work done that day. :D
    – TTT
    Commented Aug 25, 2016 at 14:00

You want to buy when the stock market is at an all-time low for that day. Unfortunately, you don't know the lowest time until the end of the day, and then you, uh can't buy the stock...

Now the stock market is not random, but for your case, we can say that effectively, it is.

So, when should you buy the stock to hopefully get the lowest price for the day?

You should wait for 37% of the day, and then buy when it is lower than it has been for all of that day.

Here is a quick example (with fake data):

10 9 7 5 8 12 14 23 14 8 14 3 9 11 1 12 3 12

We have 18 points, and 37% of 18 is close to 7. So we discard the first 7 points - and just remember the lowest of those 7.

10 9 7 5 8 12 14 23 14 8 14 3 9 11 1 12 3 12
Discard this 37%

We bear in mind that the lowest for the first 37% was 5.

Now we wait until we find a stock which is lower than 5, and we buy at that point:

10 9 7 5 8 12 14 23 14 8 14 3 9 11 1 12 3 12
└───────┬──────┘ └──────┬─┘ | 
Discard this 37% Too High   Choose this one

This system is optimal for buying the stock at the lowest price for the day.


We want to find the best position to stop automatically ignoring. Why 37%?

P(K) = Σ P(Being in position n) x P(Being chosen given in position n)

0 1 2 ... K K+1 K+2 ... N-1 N 

We know the answer to P(Being in position n) - it's 1/N as there are N toilets, and we can select just 1.

Now, what is the chance we select them, given we're in position n?

The chance of selecting any of the toilets from 0 to K is 0 - remember we're never going to buy then.

So let's move on to the toilets from K+1 and onwards. If K+1 is better than all before it, we have this:

P(Being in position n) = 1/N
P(Being chosen given in position n) = 1

But, K+1 might not be the best price from all past and future prices. Maybe K+2 is better. Let's look at K+2

P(Being in position n) = 1/N
P(Being chosen given in position n) = 1 - ( P(NOT being chosen given in position n) ) = 1 - ( 1/K+1 ) = K/K+1

For K+2 we have K/K+1, for K+3 we have K/K+2...

So we have:

P(K) = 1/N x 0 + 1/N x 0 + ... + 1/N x 1 + 1/N x K/K+1 + 1/N x K/K+2 + ... + 1/N x K/N-1
     = 0 + 0 + ... + (K/N) x (1/K + 1/K+1 + 1/K+2 ... + 1/N-1)

This is a close approximation of the area under 1/x - especially as x → ∞

∫ 1/x dx    

= [ln(x)]

= ln(N) - ln(K)

= ln(N/K)

So 0 + 0 + ... + (K/N) x (1/K + 1/K+1 + 1/K+2 ... + 1/N-1) ≈ (K/N) x ln(N/K) and so P(K) ≈ (K/N) x ln(N/K)

Now to simplify, say that x = K/N

P(x) ≈ (x) x ln(1/x)

     ≈ -xln(x)

We can graph this, and find the maximum point so we know the maximum P(K) - or we can use calculus. Here's the graph:

Here's the calculus:

P'(K) = -ln(x) - x(1/x)
      = -ln(x) - 1

    0 = -ln(x) - 1
   -1 = ln(x)
    x = 1/e

    1/e = 1/ 2.718281828 = 0.367879441 ≈ 0.37 = 37%

To apply this back to your situation with the stocks, if your stock updates every 30 seconds, and is open between 09:30 and 16:00, we have 6.5 hours = 390 minutes = 780 refreshes. You should keep track of the lowest price for the first 289 refreshes, and then buy your stock on the next best price.

Because x = K/N, the chance of you choosing the best price is 37%. However, the chance of you choosing better than the average stock is above 50% for the day. Remember, this method just tries to mean you don't loose money within the day - if you want to try to minimise losses within the whole trading period, you should scale this up, so you wait 37% of the trading period (e.g. 37% of 3 months) and then select.

The maths is taken from Numberphile - Mathematical Way to Choose a Toilet.

Finally, one way to lose money a little slower and do some good is with Kiva.org - giving loans to people is developing countries. It's like a bank account with a -1% interest - which is only 1% lower than a lot of banks, and you do some good. I have no affiliation with them.

  • 7
    You're using the secretary problem, which just doesn't apply here. For this to work you must assume that the prices during the day are sampled from some independent distribution (closer to reality is price changes are randomly sampled). It works because sitting back for the first 37% gives you a sense for the that assumed distribution. But you already have yesterday's information, and a super low price at midday doesn't mean a great buy because morning prices are not as relevant to afternoon prices to the extent that the current price is. I'm sorry to say, but your advice is simply wrong.
    – Dan
    Commented Aug 25, 2016 at 5:10
  • 4
    I should add though, that your answer is a very thorough and well written explanation of the secretary problem, that I do appreciate and enjoy. The downvotes don't detract from that, but simply represent that it isn't a good answer for this question since it doesn't apply.
    – Dan
    Commented Aug 25, 2016 at 6:39

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