G-Bot Trading Methodology: order cloud overlay
(a neutral layered scalping/hedging trading procedure)
by Tom Gastaldi
G-BOT is essentially based on the concept of overlaying "order clouds" at different price levels. An "order cloud" is a (dynamic) set of orders which yields a (statistically) profitable scalping action.
Hedging is essentially achieved by starting multiple order clouds on diffferent layers ("superposition" or "overlay"), as the price moves, in time, through its range.
An "order cloud" is formed through a set of algorithmic rules
- called a "game" - which govern the
creation of the orders within the order cloud.
A game and a random price process give rise to a random order cloud: as the price moves, we will
clearly get 1
realization among the infinite order clouds which are possible with
different prices curves.
(Note that we reject the concept of "single trade" (which is instead
the foundation of most systems), as the notion will generally lead to unprofitable approaches.)
A game can have infinite definitions, however we will use only the ones which prove to be statistically effective as to scalping action in a price corridor. The simulator included in G-BOT can be used to study the statistical properties of a game and how the corresponding order cloud is formed.
As the price sweeps its range, new games have to be started to harvest new scalps which serve the crucial function to compensate for unfavorable vertical movements of the prices (which will cause drawdown of the order clouds which accumulate an excess of positions in an unfavorable direction). Thus multiple order clouds will be defined at different price levels, which we call "corridors".
The order cloud overlay can be achieved
essentially through 2 ways:
- "cloning" instruments (for use on different price levels) within a
folio,
- starting multiple instances of the folio.
These 2 ways can be mixed at will, according to the trader's preferences. Thus,
we can have multiple folio instances, each one possibly containing cloned
instruments:
Creation of a new layer
The games used on the different layers need not to be equal: they can be
different.
They can possibly be forced to be "unidirectional" on the CT
side (if we feel
to be near a price range end, or if we have other layers with an excess of
position on one side (which is imbalancing the folio) and we want to hedge. The packet size can be different ,
etc.). See instrument panel to force entry direction.
The profitability of the whole
trading procedure arises in time essentially from the following
factors: practical boundedness of price range, overlay, reversion and absence of
vast permanent losses within the order clouds. (Drawdown will take the form of a relative excess
of unfavorable vertical movement with respect to the scalping action. In time, the
continuous scalping action and the various reversions will compensate for the
unfavorable vertical movements occurring inside the price range.)
It is clear that the sustainability of the procedure depends on capital, price range and order size. It will be advisable to start initially with relatively small order sizes and gradually increasing when proceeding with the overlays, as it becomes clearer how much money is necessary to sustain the procedure through each instrument price range, and given the observed volatility. The folio should be large with relatively small packet size in order to have a massive scalping action, while distributing risk over diversified situations.
Diversification and folio balance (CRUCIAL)
The most important concept, to understand, when trading with G-Bot is
diversification and folio balance.
This is more important than any other consideration, even the scalping
efficiency of the games. In fact, there is no performance level which will save
you if you make a wrong allocation of risk.
Each instrument should NOT absorb more 2% of your capital.
So if you are trading a game which takes at most 3 packets, make sure that the
order size is no bigger than about 0.6-07 of capital%.
So
check carefully the value of the packet of each intrument and make sure you are
allocating an even and small % of your risk capital.
Check also volatility, and
you might prefer instruments which do better at scalping (e.g. ETFs and not
stocks of single companies). Failing to do so will potentially expose to large imbalance which could
be hard to recover with the folio logic. Do not allow an instrument to imbalance
the folio. In case open a new layer forcing (strategic entry or forced
direction) an opposite first entry.
As to instruments, trade ETFs (any capital > 25K), Futures (if capital, really
large, in the seven digits) and forex (any capital). Try to absolutely avoid
stocks or any instrument which is related to 1 single company (or use really
tiny packets for it, if you want to give it a try). Don't be ashamed to trade
even just a couple of shares if necessary to diversify respect to the available
capital!
Oscillator instruments (such as volatility indices) may be preferred choices for
relatively small capital.
Games - Order cloud formation
I have defined a few example games, which can be used as an inspiration
or a starting point to
create the preferred ways to form the order clouds.
Every time a new game is
created, it should be fully understood and tested
accurately through the provided simulator. I strive to provide statistically good
games, but it's not excluded at all that you can find other games which
"dominate" (that is are uniformly better) the examples included with G-BOT.
(Further, some instruments could benefit of specific adjustments (e.g., scalp
sizes), based on their volatility and price range).
The simulator will provide an idea of the average performances of the scalping action which will be expected by the game. As we don't know the precise realization of the future, the best we can do is to select games which are statistically "strong", that is which performs well, in average, in the universe of all possible price curves. The simulator purpose is to assess average performances based on samples drawn from that universe.
The simulator will sweep through, and beyond, the toughest conditions ever possible for a price curve, and will easily bring into light any possible weaknesses of a game. The simulator purpose is to verify your intuition, and possibly push it beyond the current comprehension and insight.
Discretionary: surrounding order clouds with options / game anticipation
Depending on how one or multiple order clouds develop, you may also want to accompany the games with options.
G-BOT supports in full this activity, allowing trading and following closely the options. This will require some practice with G-BOT and some individual smartness, but massively writing covered options around the games, can yield an additional good and steady income.
Here are some hints. Feel free to suggest more. A (rather prudential) idea is using a "collar":
Case1: Order cloud unbalanced short
---- CALL options (long, to cap margins - optional, not suggested with oscillators)
Cloud with short position excess (or bullish expectation)
---- PUT options (covered options: short)
Case 2: Order cloud unbalanced long
---- CALL options (covered options: short)
Cloud with long position excess (or bearish expectation)
---- PUT options (long to cap margins - optional, not suggested with oscillators)
(I suggest never be theta negative with options, which might eat up all the work
of the scalper, so the short sale should always make up for the options bought.)
_____________________
Another trading approach is the following, where a "short
strangle", rather than the above "collar":
---- CALL options (short)
Either scalp the options if both in profit, or transfer the loser to the underlying instrument at expiration (as "strategic CT entry" to be automatically traded normally according the layered approach)
---- PUT options ( short)
In any case, when a short option expires in loss, we could "convert" it into a position for the underlying and continue trading, just "as if" we had been holding that position on the underlying all the time. [In other words, you should short the options (possibly on one side only), when you are willing to accept holding the corresponding position beyond the strike price.]
So just liquidate the loser and "transfer" the option position on the underlying, by opening an "injected entry" (see the sphere icon in the trading window) and continue trading either automatically or manually (through the order enqueuer) letting it to the care of G-BOT. (This way you have just "anticipated" the automated game and grabbed some profit from the options, which might even expire both in profit).
You will appreciate with practice the advantages of this approach, as most often you will be able to scalp multiple times the short options before expiration, and in any case you are getting the premiums. The "loser" would just turn into a position normally managed by G-BOT (thus possibly turning "profitable" in the near future within G-BOT trading scheme), which could have been open by G-BOT (but in the meantime we have been profiting from time decay and possible option scalping.)
Hint: when selling options, it comes very handy to use the Order Enqueuer. In fact due to fast movement of the bid you can capture favorable spikes by planning the orders in advance at a much more favorable price (even a few points % higher):
I have also defined a new experimental game to automatically scalp short strangles of options until expiration, and more updates will come about this point.