Kiki worked all weekend on her FloBot automated entry system. I call it an entry system because we have decided that we will use it an entry system, so we can catch trades in a number of markets and then manually manage them in accordance with the philosophy I have mentioned in the blog.
We are starting with the MultiCharts platform and Easylanguage as we have a production application we can deal with. We will do a MarketDelta version once MD finishes it's beta. We continue to use MD, as usual, for trading as per the chart below.
Kiki and I have been talking about the FloBot for a while now and decided to build it in modules. The modules are the different trade entry setups we use. The criteria for us is that whatever else we do, the testing of the entry must result in a win rate of at least 60% with price reaching the first scale point. When we run the testing module in MC, we are only looking at the %Wins when writing the trading script. I expect to both add and delete nuances into the script to get the win rate we want.
The interesting thing about MultiCharts is that there is a button on the chart to switch the autotrading engine off. In this way we can switch it off once an entry has been made and then go to our DOM and manage the trade from there on our usual discretionary basis. I also expect to better the %win rate in real time as we will exit trades before the testing stop loss if we see the context warrants it.
I'm running the FlowBot with our first version and its quite exciting
Maybe Kiki will win me over with this automated entry, as it will ensure that Kiki does enter all the trades.(a problem for her) And I must admit I wouldn't mind not having to concentrate so hard for so many hours, while be able to watch more markets. What happens afterwards is another thing. I'll show a picture later this week.
Now for Today's Trades
I started trading the ES RTH just before the bell. The first trade was a loser of 0.75 per contract where I cut it when I felt, what seems to have become, the dreaded horizontal wait. The chop continued with the 2 EMAs almost superimposed. Trade 2 had the 45CCI pull over the zero line and then fail with order flow selling. We had opened on "the zipper" of single/double prints in the Profile and I was looking for the lows of Friday to be tested so was an anxious seller at every opportunity.
Today is a perfect example of "context". This single/double print part of the profile means that price can run up or down the zipper very easily. The first trade was a loser because I sold with nothing to lean on. I had a downward bias after seeing that the London morning trade peaked and failed at just over the 1103.00 area, just under Friday's VAL. The VB was red but the CVD had turned green and I could have been a bit less aggressive I suppose, but I would rather lose 0.75 than not be in the trade when it ran down the zipper. I had a very controlled risk. As soon as there was a HH (higher high) the selling game was over and I had to re-think what was next. This visualization of context is important. I get asked a lot of questions about indicators but not many about context. The problem for a trader trying to become successful is that you want to make everything mechanical without breaking down your mechanical rules into sub-categories of context. A zipper is a significant context for me and I trade taking that very much into account.
The HH came after Trade 5 and I have finished for the day. We are at the 99EMA and the market could fail again. If it doesn't I would look for sales against the developing VAH (DVAH) or a bit higher if it turns into a Neutral day in Profile terms. There are also still single prints of the zipper above todays RTH highs which may be tested.
Click to enlarge
Do you have any intentions of making a version of the flowbot available to us EL blog readers?
ReplyDeleteWithout giving away the shop, I would find it helpful to know what conditions the FlowBot is looking for in order to make an entry.
ReplyDeleteEg:
33 slope and not near the 99?
6CCI crossing the 45 ?
X-ticks from the 33
VB in the direction of the trade.
Any form of Trend detection?
As an side, I got a little chopped up this AM in the 930 to 10 EST chop zone ... so it was helpful to read that your test for chop is a 33-99 overlap.
As well, you were able to wait for the right entry (after your small stop) while watching the CCI45 work its way above zero and then rollover. That is helpful to know.
Thanks for the best blog on the web.
Hello EL,
ReplyDeleteYou have always said that trading is art, not science. How do you reconcile this with the idea of an automated entry system? I suppose the question is how much context can be built into such a system -- could your FlowBot factor in a "zipper", for example?
Hello Boss,
ReplyDeleteYou remind me of a new show on Sci-Fi called Caprica. It's about a dad who turns his daughter into a robot. Lol
May I ask what gave you the confidence to take the second trade? After the open the long term
CCI was under 0 as well as times that the CV was red before then. Was it the second candle closing red under the EMA’s? I understand that you wanted to take advantage of ridding the zipper. Was this more of a “put the trade on and manage it” then something else? (this question is in consideration of context).
Thanks,
Rino
Ps Kiki, do you know about Daleks? http://tinyurl.com/yaps7yp
Rino, EMAs were opening up, candle rejected high (wick), position of candle, flow was red enough and, yes, get the trade on and manage it.
ReplyDeleteAnon, FloBot will only put the trades on. Kiki and I will then look at the context and make a decision on management. Its part of my "get the trade on and manage it phylosophy.
Guys, I don't know yet what will be done with the FloBot. Its Kiki's project. Let's see how it looks in testing. Getting the winrate high enough for the first scale is the first goal.
Tom,
ReplyDeleteWhat questions do you ask yourself or what principals are you looking for when determining context? You've mentioned the zipper, going with the direction the market is attempting to go, and breaking the profile into distributions, but there seems to be much more to it. I know context is something that is really learned by practice, but can you point us in the right direction?
Aaron
Aaron, see post 26 Jan.
ReplyDelete