4
$\begingroup$

Find a diversified set of financial instruments by whatever method you like. Every day, buy each instrument at the open price. Historically, the open price is almost never the high. Sell immediately as soon as the price is at least one tick greater than the buy price.

Why doesn't this just print money?

$\endgroup$
  • 5
    $\begingroup$ You may be interested in comparing your strategy to the martingale betting system. $\endgroup$ – user2357112 supports Monica Sep 5 '14 at 19:30
  • $\begingroup$ This common technique called as "scalping". $\endgroup$ – jm666 Sep 6 '14 at 1:11
  • $\begingroup$ The open is also almost never the low, so you could aswell short the asset. But in both cases, you have only marginal profit with potentially unbounded losses if it does not cross tick+spread. $\endgroup$ – emcor Sep 6 '14 at 11:02
  • $\begingroup$ Similar to other feedback you're getting, I'd say you'd have to better define this strategy. For example, you might narrow down the diversified set of stocks. Maybe you would just choose the previous days biggest gainers or something like that. Or maybe only choose from stocks that are within 80% of their 52-week high (which some folks use as a base for quant strategies). $\endgroup$ – Unknown Coder Sep 8 '14 at 15:59
5
$\begingroup$

There is a general principle that the answer to "Would this extremely simple strategy make money?" is "No". This is the "no free lunch" or "no arbitrage" rule. It isn't exactly a physical law, but it is a pretty decent approximation of reality. (A more nuanced version is, maybe it can make some money, but only in proportion to the difficulty, risk, and expense of executing the strategy.)

Your strategy as stated is incomplete. You need to say what happens for a stock when the price goes down from the open. Do you keep holding it to close? Or set some stop-loss and sell then? Keep holding it overnight so you don't have to repurchase at the next day's open?

Lets say you do that last version: keep holding. Then consider the results. Classify stocks as "winners" if they generally trend up during a day, "mixed" if they bounce up and down, and "losers" if they generally trend down during a day. At the start of the day you have a portfolio of everything. But you soon sell out of the winners and mixed stocks for a tiny profit on each. By the end of the day you are only holding losers. Do you expect this to outperform the buy-and-hold-everything strategy? I would bet even without transaction costs it doesn't, and with transaction costs it throws away money like crazy.

$\endgroup$
  • 1
    $\begingroup$ +1 for the Strategy as stated is incomplete comment.... And to the original question.. Historically, the open price is almost never the high. Is that another way of saying the market has historically grown in value in the past X years? $\endgroup$ – zipzit Sep 6 '14 at 4:44
  • 3
    $\begingroup$ When this strategy works, it will make a small amount of money. When it fails it will lose oodles and oodles of money. $\endgroup$ – corsiKa Sep 6 '14 at 7:18
4
$\begingroup$

Like Good Guy Mike says in his answer, you have to account for transaction costs. But change "one tick" to "transaction costs + 1 tick" and you still have a question.

One issue is that while the open may not usually be the day's high, you'll have a few catastrophic losers that you hold all day, and they may be enough to wipe out your small steady gains.

If not, then why stop at the open? Any given price is probably not the stock's all-time high, so what about this: buy at the prevailing price, sell once the price reaches the price you bought at + transaction costs + 1 tick. Then buy again and repeat. Unless the occasional catastrophe wipes you out, this will make money. But this strategy is strictly worse than buy-and-hold because of transaction costs and the gaps in time between selling and the next buy.

It's easy to make positive profits; it's hard to beat the market.

$\endgroup$
1
$\begingroup$

Because you have to take transaction costs into account.

$\endgroup$
  • 1
    $\begingroup$ Okay so make it tick + transaction cost $\endgroup$ – crf Sep 5 '14 at 17:12
1
$\begingroup$

I just checked for S&P 500 from Yahoo Finance. From Jan 3rd, 2005 to April 14th, 2014 (2336 trading days), for 2014 days High is greater than Open and 322 days they are equal.

Assume no friction.

Suppose you are an almighty person to catch Highs on your trade, then in one day you will make %0.00675 on average. That is your ceiling.

Now add friction and be more realistic about your predictive power. Remember the 322 days where High is the Open (things get worse).

If you had had invested in S&P 500 on Jan 3rd, 2005 from the Open price and sold on April 14th, 2014 High price your return would be %51.3458 (minus some friction). This is your benchmark.

This is an interesting strategy. I came across a few more. It reminds me of %y filter of Alexander although it is different.

$\endgroup$
0
$\begingroup$

The open is also almost never the low, so you could aswell short the asset. But in both cases, you have only marginal profit with potentially unbounded losses if it does not cross tick+spread.

E.g. see this change in EUR/USD exchange rate: http://www.ariva.de/euro-dollar-kurs/chart?layout=neu&boerse_id=36&t=week

This event would have created a definitive extremely high loss, when going long before the ECB rate cut, which might easily counter all previous tick-gains.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.