the flipping coin argument :)

just another realization occurred to me while reading Greg Michalowski, sharing it to you guys

how many times you have heard this:

Markets either go up or down, flipping a coin should be enough to give 50% odds of success!

----
practically we know things are not as easy as the coin flip argument suggests!
do you know WHY?! Here is the explanation.

on a normal trade there will be a bid-ask spread for 2 pips.
when you place a long order, the ask price should move up 3 pips from
current ask price to book 1 pip profit.

i.e Here is long position:
when you place long order: bid = 1.2334 ask =1.2336
when you book 1 pip profit: bid = 1.2337 ask =1.2339

to allow 3 pips movement up we should accept the risk of 3 pips down
assuming markets go up/down 50-50 times. so a loss 3 pip SL is needed.
if you add in the 2 pip spread, when the SL is triggered you will incur 5 pips loss!

Lets say we have a magic coin flip technical indicator that's right on about whooping 84% and we make 100 trades!

84 times we make 1 pip
16 times we loss 5 pips !
Net gain = 84x1 - 16x5 =4 pips ! duh! after 100 trades (assuming you don't pay brokerage!)
=========

so there you go, even with a magical coin flipper with 84% accuracy,
and markets going up or down by 50% probability, the trader will go broke!

Hi BSM,

'Flip' is just a great talking point....
Let's say...you have lost all your Capital with the first five Flips... and your next 5 Flips will give you the winners ....where is the Capital to trade the next 5 Flips.....

What is not elaborated..is the Key Point of 'Winning Streaks' & 'Losing Streaks'

SG

Here is one more familiar dialog, lets call it "switching signals argument"

"damn, if i had just switched buy and sell signals for all the trades I have made, my trading account would be up by 200%"

who have not ever said some variation of the above statement !

well, if switching signals argument is right then 80% of the losing traders would have done that already

lets apply a transformation on "switching signal" and lets replace it with "switch calls and puts", we will be always long on 'call' or 'put'. All those 80% of losing traders will lose no matter whether they trade 'call' or 'put' !
The problem is NOT in the instrument but their risk-reward ratio for their entries and exits as explained in the first post!!!

Similarly even if a loser trader switches signals, he will always end up entering/exiting trades with poor risk-reward ratio!

so 'switching signals' won't work unless the trader gets a grip on his 'risk management' (plus money management, psychology etc)!!!

Argument in your post ignores dozens of parameters and circumstance...just to prove that 84 % accuracy does not matter.

that argument is valid for scalping/jobbing.....

but certainly invalid for swing trades...where you try to capture 50-150 pips....in that case 5 pips...does not matter......

also 50 - 50 argument should apply in market..
its like car driving.......what are the chances that you will hit someone...or some else will hit your...
The chances are high....very high...
but still we drive it successfully..........only because of skills...

What if indicator fails

16 times we make 5 pips
84 times we loss 1 pips

typical psychology lol

Interesting discussion. Let me contribute something to it.

I came up with a theory based on this coin flip idea only. Do a coin flip. Take long/short according to heads/tails. Probability of being correct = 50%, whatever be your timeframe.

Now, what if we can match this with a 1:2 or better risk:reward in our favour.. so let a coin decide what trade to take, then let there be fixed stop & profit target with profit being twice that of stop loss & leave the trade. It will hit either of the targets. Now if probability still stands of being right = 50% then you make money. But the problem is that the probability will now move below 50% cause trades will hit stop with greater frequency cause its nearer.

But will probability move to as low as 33%, which would be our breakeven point for a 1:2 risk:reward ratio. Well, the answer to that depends on how large the absolute values for stops & profit targets are. If they are so small that would count as noise, then chances of making money becomes little. For example. 10-20 points for Nifty. But if it is large enough, which would be hit only when there are good trends, the accuracy might turn out to be better than 33%. If we move the ratio to 1:3 then the breakeven moves down to being correct 25% of all trades.

This entire argument sounds very neat. But I have tested it extensively in excel. I started off with a very simple sheet and moved to make it very very complex.. but money wasn't to be made consistenly..

imo trading cannot be modeled on a coin flip behavior..however in the example you demonstrated you are playing for 3 pips ..what if you make it 30 pips on each side..in that case your profit with a 84% system is

84*28-16*32 = 1840 pips in 100 trades.. so compared to your earlier example you have an enormous advantage...

All i wanted to show is that commissions and cost of in and out takes a significant part of the profit so it makes sense to play the game on a little longer time frames.

You do have a very valid point bigger time frames contain nice bigger trends, which allow low risk high profit entries. Lower time frames profits are so low, it can't even offset the brokerage (for retail traders) most of the time.

anyways the point of illustration is how a best scalper won't be able to make money when his risk/reward ratio is poor. Obviously smaller time frames nearing the tick level has very poor risk/reward ratio for retail traders.