Coin Flip Trading

I think you are right.
Although there are 2 outcomes, the staging means I have to consider the individual outcomes separately.

I may need to write up all the outcomes.

Scose: thanks for xlsx file.
 
Maybe you should ask experienced coin t0ssers. There were some good observations regarding r/r - how to skew all that random entries in you favour from The Hare

BTW I haven't seen him posting for some time
 
"No sh1t sherlock! He said excluding expenses/commissions etc"
Click to expand...

A comment from Shakone via "recommendation"

The spread is not an expense.

The fact your 'co-equal' stop/target requires less movement to hit your stop than yout target is also not an expense.

The problems is the argument that there are 2 outcomes +100 or -100 is FALSE.

With a 100 tick stop & target, there are 2 outcomes on a long trade -99 (stop) or +101 (target).

You have to presume +101 to get filled on your target because you cannot presume front of queue.

I mean - you do actually want to get filled on these trades, right?
 
hi DT,

I appreciate the -99 and +101 comment.

I wanted to ignore those components initially, and prove whether the idea would work, in principle.

If the idea cannot work in principle, any further explorations become meaningless.

If the idea can work conceptually, then its worth considering the practicalities of fills, slippage, requotes, widened spreads due to news,etc.

I thought it best not to get bogged down in details too soon, unless the flaws are in the details!


I think scose's excel is saying the outcomes are 5 separate trades, and their likely outcomes must be seen individually.

eg, Trade1 has a target of 100, but a SL of 20.
Trade2 has a target of 100, but a SL of 40.
Trade3 has a target of 100, but a SL of 60, etc.

In a random scenario, the 20SL trade is 5 times more likely to get hit than the 100, and thus over time, 5 losses of 20 pips against 1 win of 100, results in a neutral outcome.
I think thats what he was alluding to.

Applying the same idea to the other trades, I have a neutral set-up, which is ultimately negative due to the realities of spreads, comms, etc.
 
It won't work, equal stops and targets plus costs means its negative expectancy.
Nothing will change that.

Its an arbitrary reason for entry, that is all.
Its also not particularly efficient entry wise.
For it to work at all, the key is trade and exit management.

As a learning exercise, or as a basis for automation there is merit.
If you don't fall into either of the above camps, stick to something
like S&R, BRN, eco calendar and H1 trend direction.
 
DionysusToast said:
Yeah - but what they are saying is...

"apart from all that nasty reality stuff.... WHAT IF???"
Click to expand...

:LOL: True, even then it will be breakeven at best.
As you say, I don't see any point disregarding slippage, comms and spread,
not to mention technical issues - rejected orders, order servers going down
with no exit fill etc.
Those are the issues that will, in reality, make or break most methods.

Example, I recently had a rejected order, the SIM account running in parallel turned 89 points on EU.
Live account pulled jack sh1t due to rejected order (order server balls up).
Thats the kind of rare event that must be factored in...
 
BTW, DT, bit of a hijack, but you heard anything on NT8 release?
 
Liquid validity said:
It won't work, equal stops and targets plus costs means its negative expectancy.
Nothing will change that.

Its an arbitrary reason for entry, that is all.
Its also not particularly efficient entry wise.
For it to work at all, the key is trade and exit management.

As a learning exercise, or as a basis for automation there is merit.
If you don't fall into either of the above camps, stick to something
like S&R, BRN, eco calendar and H1 trend direction.
Click to expand...

It is, as you say, a learning exercise. A thought experiment.

In real-life trading, I tend towards identifying the trend on the 1-4hrs, and look for pullbacks / slingshots on a lower TF into the higher TF.

As a skew, I am finding knowing the Euro and US session ranges quite useful, but haven't been able to articulate them into rules. although, I suspect, it's just Sup/Res seen differently.

I stay clear of eo-calendars.
 
trendie said:
I stay clear of eo-calendars.
Click to expand...

Depends how you use them.
I'm not saying trade release, volatility and widened spreads are two good reasons not to.
Always a good idea to know about an upcoming ECB or FOMC statement though
for example.
I'm thinking more in terms of avoidance of ranging, or volatile chop.
 
trendie said:
It is, as you say, a learning exercise. A thought experiment.
Click to expand...

But to what end when you aren't considering how you would get in and out? If you base it on buying the bid and selling the offer, the numbers have no meaning.

You toss the coin, it's a buy, you join the bid, don't get filled. All of a sudden, the only longs you enter are those where people trade against you at ths start.

Or

You toss the coin, it's a buy, you get in at the offer and you give up the spread.

The latter is easier in terms of being able to accurately project the probability.
 
The efficacy or expectancy of the overall strategy or the fact that you need to account for the extra tick for fill isn't the issue here as the net effect is largely immaterial to what's being posited here - whether this type of scaling out will result in asymmetric risk. It won't. Well it will but the resultant will be the inverse of the initially .desired outcome
 
Between t and t1 the initial 50/50 diminishes and skews away from a positive outcome. So says the scose.
 
DionysusToast said:
But to what end when you aren't considering how you would get in and out? If you base it on buying the bid and selling the offer, the numbers have no meaning.

You toss the coin, it's a buy, you join the bid, don't get filled. All of a sudden, the only longs you enter are those where people trade against you at ths start.

Or

You toss the coin, it's a buy, you get in at the offer and you give up the spread.

The latter is easier in terms of being able to accurately project the probability.
Click to expand...

The point is when you're first considering something new, you consider the simplest case. If no benefit there, then no benefit when you add in the drawbacks. If there is a beneift there, then start adding in drawbacks and see if you've still got something.

The alternative is to try to be realistic, so we have a +101 move versus a -99 move, but wait that's not realistic, we need commissions, so we need a +101.25 move versus a -98.75 move. No that's not realistic enough, spread can widen, the spread could be as large as 10 or as low as 1, so we need a move of somewhere between +110 to 101.25 for a win, and a move of -89.75 to -98.75 for a loss. Oh wait, there's also gaps, and then there's slippage or no fills, and why not consider broker problems and internet connection issues...what was the concept we were trying to look at?
 
scose-no-doubt said:
The efficacy or expectancy of the overall strategy or the fact that you need to account for the extra tick for fill isn't the issue here as the net effect is largely immaterial to what's being posited here - whether this type of scaling out will result in asymmetric risk. It won't. Well it will but the resultant will be the inverse of the initially .desired outcome
Click to expand...

It is totally material because it is the amount you will lose overall.

Just like roulette. There is an inbuilt negative edge for the player. You cannot ignore it.
 
Shakone said:
The point is when you're first considering something new, you consider the simplest case. If no benefit there, then no benefit when you add in the drawbacks. If there is a beneift there, then start adding in drawbacks and see if you've still got something.

The alternative is to try to be realistic, so we have a +101 move versus a -99 move, but wait that's not realistic, we need commissions, so we need a +101.25 move versus a -98.75 move. No that's not realistic enough, spread can widen, the spread could be as large as 10 or as low as 1, so we need a move of somewhere between +110 to 101.25 for a win, and a move of -89.75 to -98.75 for a loss. Oh wait, there's also gaps, and then there's slippage or no fills, and why not consider broker problems and internet connection issues...what was the concept we were trying to look at?
Click to expand...

Considering the fact you have to actually get into and out of the market is hardly a drawback.

It seems the concept you are trying to look at is random entry on a fictional market where you don't actually have to place trades.

It's like "OK - let's ignore zero's for a minutes & bet on black on each spin of a roulette wheel" and work back from there.

It's also not dissimilar to saying "ignoring the laws of thermodynamics for a second, let's look at this design for a perpetual motion machine".

Consider the difference in getting a stop and a limit order filled part of the "laws of physics for the markets".
 
The trouble in trying to reduce everything to mathematics is that you finish up with an almost impossible formula if you truly take into account all the variables mentioned by toastie and others. For my money, trading is more art than science.

In any event I doubt the basic starting point of 50/50 chance (on mid-price to avoid the spread aspect) of price reaching x or y in the first place. FTSE currently stands at 6574 so, arguing from absurdum, there is a 50/50 chance that would go to 0 or 13148. In reality, whilst the market exists, there is absolutely no chance whatsoever that it would go to zero - changing its constituents ensures that - so only 13148 remains as the possible outcome between the two.
 
You must log in or register to reply here.

Similar threads

H
The Coin Flip
2 3 4
Replies
79
Views
11K
DionysusToast
DionysusToast
B
Most traders underperform to a coin flip day trading bot
Replies
0
Views
1K
BE Trader
B
V
The Coin System
Replies
14
Views
4K
puregold24carat
P
Sang Froid
Coin Toss Experiment vs Discretionary Trading
2
Replies
30
Views
10K
Jack o'Clubs
Jack o'Clubs
S
Flippin' coins
Replies
4
Views
2K
DionysusToast
DionysusToast
Top