Sunday, 7 February 2010

Kiki's FloBot V1

I have had quite a few emails about Kiki's FloBot so I indicated on Friday that I would provide some details of where we are so far. Below is a screen shot of Friday's trades and below that are some statistics. To make sense of what I am talking about, I need to tell you what you are looking at.The stats chart covers about a month of trades starting January 2009 and a few days of February 2010 and shows the results doing trades in two different ways, we were running both. The first line shows the results for an "average" good trade exit under current volatility conditions, which for us is a single target of $375/contract profit. The second line shows the results based on exiting the trade at what is my usual first scale: 2 points or $100/contract profit.


Click each chart to enlarge

  These are the results that would have been achieved letting Kiki's FloBot enter and exit the trades, assuming no slippage on a 1 contract basis.

What is interesting is that there was a 74% chance that each trade made the first scale out of $100 profit with a $200 stop loss. This gave a profit for about a month of trades of $2392 after commissions with a maximum draw down of $886.50. That means you could have earned $2392 with a minimum capital of $500 day trade margin plus $886.50 set aside for losses. Using my usual conservative approach, it would be earning $2392 for every $5,000.

Had you opted for the "average" exit of $375 and tightened your stops to $125, then your win rate drops to 39% but the profitability jumps to $3227 with the lower draw down of $756.50.

OK, so which is the best way to trade? Well, it depends on how you actually want to use an automated system.

My idea when I agreed to Kiki's idea to get this project going was to use the FloBot only to enter trades. I didn't, and I must say that at the time of writing, still don't believe that it is possible to program all the "context" requirements into the model. Maybe over time we can get closer but at the moment we are far away from that.

So the idea of using the FloBot to only enter trades will make me more profitable by:
  • allowing us to trade more markets at the same time or to trade one market without needing the high level of focus as we need now, and,
  • ensuring that trades are entered and not missed for both discipline and focus reasons
  • allow the FloBot to exit the first 1/3 of my position at the scale out point - I get this 74% of the time, and I manage the balance as I do now.
The big plus point is that I can then have the best of both worlds by manually managing my trade after the FloBot's exiting the first 1/3. I can then decide whether to exit the rest of the position at the same 2 point profit or whether to run it and scale out progressively. Taking this route, I hope to maintain the 74% win rate and keep the average profits at the level that I am at now. Overall, this should translate into a larger monthly payday as I can cover more markets for less concentration.

I see this as a step forward for a trader as it makes the job easier. You don't need to work at the 100% level 100% of the time, which is impossible anyway.

So how should Kiki use it? We have spoken about this at length. We have several options:
  1. Let the FloBot do all the work and just hang around and make sure that the internet connection stays up and that the FloBot doesn't get out of synch with the broker, or,
  2. Let the FloBot only enter trades and manage the exits, or,
  3. Use the FloBot as a mentor to make sure that Kiki's trades meet her trading plan
I think that what will happen for Kiki as a mixture of all three, depending upon how she feels on different days.


  1. Based on the chart, FloBot needs the mentor. Most of the losing trades were Outside-Out trades. You have mentioned that one consideration for entry is that there be enough range left to the Edge for the first target to be hit and in those instances when you have traded Outside-Out in the past, it is as a trend continuation trade. Seems prudent to require FloBot to take a profit on the high probability Inside-Out trade as a prerequisite for a same direction follow up Outside-Out trade. A hedge against (at least on this day) the lower probability outcome.

  2. If they both use the same entry rules, why does one approach have 62 trades and the other have 74? Shouldn't they both have the same number of trades?

    Many thanks for your excellent trading blog.

  3. Tom,

    Thanks for sharing the flobot results and giving your thoughts on how you might use it. A couple of questions in general ( not related to flobot) - are most of your trades in the direction of trend? Most of the examples and the trades you post don't seem to consider market profile levels. Do you use them in a different context?


  4. Hi Tom,

    I have to ask two kind of basic question in regards to your plan and MP setup - sorry if you answered them before.

    First: From what I destilled out of your posts it seems that you seem to have a rule do the second scale after a counter candle at S/R. Is that true?

    Second: In terms of your MP chart setup (or even in general): These are set to the session times? This has to sound stupid, but I ask because most of the contracts are traded more or less round the clock.

    One more thing: How do you decide when to split a profile? I'm not really experienced using MP at the moment and it's still tough for me, to see structure apart from the obvious levels.


  5. In my experience of system trading the ES, the market won't give you anything like the entries and exits that the program assumes, whether you use limit or market orders. The reason for the seemingly good result is down to the excellent liquidity that the program assumes. Now you have to add a lot more of your trading experience to the program to get the profit factor up to at least 2.5 after an assumed slippage of at least $12.5 on all entries and exits in the back test. Good luck, I am following with great interest, and thank you for the excellent blog.

  6. Hi, I have entered on the 5th tick only to see the bar close in the wrong direction many times lately. Do you exit the trade right away when that happens?

  7. Anon 17:04, FloBot is to enter the trades and for us to then manage them.
    tickvix, there are levels and levels in the Profile, both splits, volume and balances.
    timokrates, nothing is rules out as it depends on the context. For ES I look at RTH and 24hr but RTH S&R has an enormous effect.
    Paul, I'm just looking for entries so I can manage them. Getting exits programmed is way too hard if you want to maximise profits.
    Anon 15:46, I exit if I called the momentum and order flow wrong