Quote from Random.Capital:
It just occurred to me that I don't, in fact, have any idea what my broker's minimum balance requirements even are. Curious that you do need to know.
Well, maybe not that curious.
Now can we get the thread back to where it was going before? I'm interested in seeing where Whiskey and 'Brot are take this...
I'll clean up the thread later in the week.
I was waiting to see if anybody else contributed something useful and perfectly objective before picking up the pace.
I'm going to suggest a self-evident improvement over the original trading coin-tosser method and comment briefly.
First Modification to coin tosser and Proof:
1-Lets have two accounts, bull account always net long L contracts and bear account always net short S contracts. L and S can change everyday within a range from 0 to MAX.
2-Whatever the coin-tosser outcome is the bull account goes long L contracts at 16:15 and the bear account goes short S contracts at 16:15. The exits remain on next day close at 16:15.
Obviously we don't need the coin tosser services anymore, so we offer him a bath, shave, clothes and we promote him to bucketeer and position manager. At the end of the day he will match sell orders from the bull account with buy orders from the sell account, and also match the new entry orders for both accounts for the next day trade. He will bucket all these orders into a simple buy or sell NET order. He will keep the accounting of both accounts separate and monitor the net position.
So far we have:
1-Diminished overall risk over the original method. (In the case of L=S we arrive to the trivial solution, which we know outperforms the original method, as well as 95% of traders in the world).
2-We diminished somehow the impact of commissions and slippage by bucketing buy and sell orders. (In the case of L=S we have cut the costs to zero for that day).
3-If we force the L=S trivial solution, then we do not need the coin tosser anymore, and neither an entry algo.
Basically, in the case of 3-we have one account entering in the direction of the coin-toss and one in the opposite direction, effectively canceling each other and saving commissions. One could argue that this is exactly the same as the trivial solution, and yes it is...unless we start modifying the EXITS.
So, as food for thought we can start thinking about:
a-How can we increase our edge by exiting long positions and short positions better than at 16:15 on next day close?. Can we use any new information that becomes available DURING the day?.
b-How can we create a more intelligent position manager for the bull and bear accounts?. (Notice that as corollary, the difference between L and S becomes a market direction bias indicator, which could go from MAX long to MAX short in position).
I believe some people may see this better if they think about two games being played simultaneously: The LONG game, and the SHORT game. Like Football perhaps?. Both games can be won everyday in hindsight. Can we bring some of that hindsight to foresight?. Only a relatively small improvement can yield enormous results if it can be compounded over time.
I'm pretty sure oraclewanker will pop a vein thinking about this. Since he drew first blood with his coin up my ass suggestion, I will decide when I let him go.
LMAZO.
