Scalping and The Winning Combination

I had an interesting conversation with someone about scalping. By now most readers will know what I think about scalping and would rather stick a knife in the side of their heads than hear me repeat myself.

Part of the discussion centred around a trade that tries to take advantage of a volatile market that has no particular direction. The discussion led to combinatorics (the calculation of combinations), profit and loss stops, and trend trading.

Their argument was as follows

With regards to scalping being 50/50, let's say 10 minutes before the off I offer £10 at 6.0 and my offer gets accepted, and I place a back order at 6.2. Would you agree with me that there is a greater than 50% chance of my trade being completed successfully before the off if I don't have a stop loss in place.

Intuitively, I knew this to be 50/50. My answer was

If you put in a random lay at 6.0 and a back (at) 6.2 then in the long run 50% of the backs will be taken and with no stop loss you will lose money. If people are backing at 6.0 and taking your lay then I would think there is a slightly greater chance of the market going down than up. You are hoping for market volatility and that the spread marches up and down the ladder. The chances of that are again 50/50.

If they had asked, "In an upwardly trending market." Then I might have said something different. However, because I thought the question was about a volatile market then my answer is correct.

A further email was received

The market's possible next 4 up/down movements are as follows:

DDDD
DDUU
DDUD
DDDU
UUUU
UUDU
UUUD
UUDD
UDDD
UDUD
UDUU
UDDU

The first 4 scenarios result in my trade failing (assuming that I'm not matched fully when the market is at 6.0 - 6.2). The latter 8 scenarios are ones where my back order would be taken out. Does that not suggest that, if the market decides to go on a random walk, I'm far more likely to have a successful trade than not?

Do you see the error? How many combinations should there be? I make it 2^4 (2 directions to the power of 4 successive time periods) but I only see 12, which is why the questioner thought that they had a greater chance of the market being up after 4 moves. Here is the full list of combinations with statistics that will be explained later.

DDDD - 5.4, 5.3, 5.2, 5.1 - lose 8 (-1)
UDDD - 5.6 - win 1 (+1)
DUDD - 5.4, 5.5, 5.4, 5.3 - lose 4 (-1)
DDUD - 5.4, 5.3, 5.4, 5.3 - lose 4 (-1)
DDDU - 5.4, 5.3, 5.2, 5.3 - lose 6 (-1)
UUDD - 5.6 - win 1 (+1)
DUUD - 5.4, 5.5, 5.6 - win 1 (-1)
DDUU - 5.4, 5.3, 5.4, 5.5 - scratch (-1)
UDDU - 5.6 - win 1 (+1)
DUDU - 5.4, 5.5, 5.4, 5.5 - scratch (-1)
UDUD - 5.6 - win 1 (+1)
DUUU - 5.4, 5.5, 5.6 - win 1 (-1)
UDUU - 5.6 - win 1 (+1)
UUDU - 5.6 - win 1 (+1)
UUUD - 5.6 - win 1 (+1)
UUUU - 5.6 - win 1 (+1)

The questioner also assumed that each move would be equal (i.e. 1 tick) so I have additionally marked the combinations with the consequences of 1-tick profit taking and no stop. We also assume that the trader offered 5.5 and was taken and then offered to back at 5.6 rather than taking prices on the other side of the spread.

As we can see the trader had 10 wins of 1 tick but lost 22 ticks without the stop. If there had been a stop (let's say 1 tick) then the result in parenthesis would have been the outcome and we have the 50/50 that I expected; 8 ticks of profit cancelled out by 8 ticks of losses. The mathematics of probabilitity must be fully understood otherwise some serious mistakes can be made.

If you are thinking of scalping (and I don't recommend it) then you must understand that in a volatile market you will lose money (break even minus commission or worse if you are manually trading). In a trending market you shouldn't be tick scalping at all but trading the trend for multiple ticks. Feel free to print this article out and take it to one of those useless trading courses you are about to waste your money on.

8 comments:

  1. Would you lose money in this example? 6.0 is not 5.0 or 5.5 it has something else going for it!

    ReplyDelete
    Replies
    1. No. Same result. A random walk is a random walk.

      Try working it out. Good practice.

      Delete
  2. Most scalpers I dont think recommend doing so in a volatile market. So for horse racing markets that doesn't really leave many markets if any (depending on what the max level of volatility is recommended which is normally just left to the user to work out. When it fails the user is to blame.)

    ReplyDelete
    Replies
    1. That leaves us with trending markets and I wouldn't scalp them either.

      Ride the wave.

      Delete
    2. A random walk is a random walk but a crossover is a crossover. A 50% strike rate on a crossover will render a profit if your entry is a lay of one.

      Delete
    3. And a reversal is a reversal.

      Good luck with your trading rules.

      Delete
  3. Your insight is refreshing. I have always thought, scalping is basically a con-trick, advocated by software developers, to help them shift product. Maybe, some churn can be automated, to help keep up comm levels, for PC payers, but, for anyone to spend useful time, manually trading .. I would be shaking my head.

    ReplyDelete
    Replies
    1. Scalping looks easy, is easy to "teach" and is a great hook.

      Software shifters don't care what you do with the software so long as you keep up with the subscription payments.

      Scalping is like trying to juggle on a rug, next to a ticking bomb, whilst people are pulling the rug from under you.

      There are plenty of better strategies but they are difficult to teach (doesn't shift as much software) and require better trading infrastructure (doesn't shift any software).

      That's why I produce my own software and keep my strategies to myself.

      Delete