I've spoken about trading once a day and then stopping and I have spoken about trading in HFT mode. Today I'd like to take the next step and provide another alternative to consider.
My dad loved casinos. He wasn't a big gambler but he did love to gamble. His thing was roulette. He spent a lot of time in his retirement working on a roulette system. Ultimately, he did come up with a system that regularly made money. Of course it only made money because of his money management rules. The roulette table has either a 1 in 37 or 2 in 38 edge in favour of the casino so the only way you can compete is controlling how you bet - the only thing you can control.
I can use this same money management system in trading.
One of the components of his money management system was to stop playing when he had made a certain number of times his perceived draw down.
What we are doing with Kiki in parallel with the trial of FloBot using a demo account is to data mine our statistics to determine:
- Is it more profitable to stop trading when a certain profit or loss level is hit every day?
- Are there times of the day when we shouldn't trade?
The outcome of this could be quite useful. We are thinking that the solution for an electroniclocal is to trade a few markets simultaneously on the basis that we stop trading when a target amount for the day is reached. We have noticed that even on a losing day in our HFT trading model, there is a point during the day that the system was in profit.
The maths could look something like:
We trade 3 different markets with a daily profit target for each of only, say, $200 per contract. This looks as if it's easily doable on our current statistics. If all three markets reach this target then trading each with just a 5 lot would yield a profit of $3,000 a day. Not too shabby. We will need to further study the best daily stop loss for the system so that if a market hit a certain level of loss for the day it would stop trading.
An extra benefit of auto trading is that in these days of 24 hour markets, the volatility and moves in the market I follow may take place during a time zone when I'm asleep or stopped trading. While I am not comfortable leaving FloBot completely unattended, I have no issue with it trading while I am asleep as long as I have both internet and broker disconnects alarmed so I can awake and sort out any issues.
Kiki is a huge fan of this way of trading. She is a huge Nipponophile and would love to spend a year in Japan and she sees this as a way of achieving that. I must admit I am becoming a bigger and bigger fan of FloBot trading as we get deeper into it.
I'd love to get some feedback and ideas from anyone who is interested in going down this road. I think it opens up a trading career for people who may not like or be suited to discretionary trading.
I am definitevely thinking in automation as a part of my trading career, as a way to diversify from trading 100% depending on my self and be able to spend more time with family.
ReplyDeleteAs far as I know there are brokers that focus on trading automated systems, in their own servers, so one should not worry about internet connection issues, nor alarms, etc.
Sorry, Tom, but there is no roulette system which makes anyone money because the casino has the edge. Money Mgmt. rules are fine and if you have good ones you will lose slowly against the casino. Or do you really believe in a roulette system???
ReplyDeleteCheers,
M
Do you really think it makes sense to set a profit stop for the day? I cut your most profitable winning days short. This does not apply to a daily stop loss. The latter one helps to avoid being not hurt to much during times when the system works terribly.
ReplyDeleteCheers,
M.
would be great to get rid of the weakest link (me...)i have managed to define 2 setups which are great but in real time i can't make it work. ( i have a small account and dont give it ehough room, premature profit taking, etc.)
ReplyDeletetoday on the es i count 6 setups (from 15 pm my time (im in the netherlands) till now 20 pm = 5 trading hours.)
all setups were winners , im shooting for 8 ticks with a stop at 8 ticks as well.
today was good for easy 48 ticks....but i dont have them in my account.....
im trying for seven years but im still not CP.
I need a bot to take the trades for me...
I have MultiCharts but am a lousy programmer...
thanks
Otto
I think everyone will be interested. lol. Not sure if your profile work is used in the strategy but I think that could help give areas of interest. Maybe that's one of the discretionary parts. For me it's important to keep a reference of what the range has been, open to high, open to low etc. a possible outlook so to speak. Depending on when the system is traded those numbers with be dramatically different. I wish you guys the best of luck!
ReplyDeleteHi Kiki,
ReplyDeleteHi El,
Do you consider the correlation among the markets you are going to trade? There is a high correlation between equities and currencies, equities and fixed income (negative correlation), etc... I think that the signals for each markets will be according to the correlation... What do you think?
Regards,
Sam
Count me in the mix Tom and Kiki! I've been here from close to the beginning and would love to see "Flo" come out to play!!
ReplyDeleteHi Tom,
ReplyDeleteI am a long-time CP trader who has followed your blog with great interest. Believe it or not, you've taught this old dog a few tricks. Yours is the only blog I bother to look at anymore.
Regarding trading to a profit target each day: I noticed early on in my career that this worked well, but I was never sure why. One day it became clear to me. In the makets I trade (ES and CL), there are a limited number of meaningful prices moves in any given session. Once those moves have played out, staying in the game leads to losses. Therefore, once I cashed in on a good move or two and hit my goals, stopping trading meant keeping profits in my account rather than trading them away in a sideways market.
Best regards,
TD
Hello Boss,
ReplyDeleteWill comment later on the building of EL Auto nation.
I am sure you have commented on it before however, how often do you check the range bar specifics? If one trades today, they know
with the tight ranges and Super X meets Rigger (Citigroups revamp algo http://tinyurl.com/yjfamco) there is a fair amount of chop. I have been using 1.5 which seems to help. Seems like a little, But does make a difference.
Long live EL Nation. (Even if you turn me into a Stepford.)
Rino
Would EL have taken these trades? 2 for the day: 1 short and 1 cover/ reverse long.
ReplyDeleteI stripped off other indicators but did use Keltner/ Momdots + EMA's.
http://bit.ly/a8UqZp
Hi Tom, Kiki:
ReplyDeleteThis would be of interest to me as well since it would allow more time to do things with the family and others.
In regards to a profit target each day, I would propose that one would be good as noted by TD but would also add that the trading plan allows for additional trades but that there is a specific amount of "give back" or number of losses that would trigger stopping for the day.
If you are on a "hot streak" this would allow for an "outlier" day as Don Miller talks about which can make the week, month, or year in terms of profitability.
One does want to avoid however, the familiar scenario of being profitable in the morning only to give it all back (and sometimes, more) in the afternoon.
Best regards,
Milt
Developing an automated trading system that performs well in backtesting and in sim/demo trading is easy. Developing a HFT trading system that does well is even easier. As I understand it you are assuming fills on all your limit entry orders and you are assuming no slippage on your stop exits. This is going to come back to bite you when you go to trading real money in real time. With HFT for small gains it's all about percentages. Your 65% figure seems to be based on "perfect world" trading and with real world automated trading, as unforgiving as it is, that 65% figure will slip away.
ReplyDeleteI am very much on board with daily profit targets.
ReplyDeleteHaving a daily profit target reduces market exposure and lots of commissions. Also I am not a greedy person, I just want to be CP.
Thanks for a great blog. I agree with a previous poster that yours is the only blog I bother to look at anymore, except for Dr Brett's blog...
Thanks
John
Hi Tom & Kiki,
ReplyDeleteI am very much looking forward to seeing the progress of the Shock & Awe Bot...looks like you could be on to something quite special!
Aloha,
Kaloko
I also have noticed that MP rarely enters the conversation anymore in your posts. Is it written in to the flobot?
ReplyDeleteAs said above, my biggest obstacle to successful trading is myself....programming the trades is a brilliant idea that is becoming more accessible to all of us. thanks for sharing
Hi EL, thanks for generating some ideas on this end. While my preference is to rely on discretion, I can see how flobot opens up the realm of possibility for more opportunity. I imagine that I'd be able to use my read on various markets to filter the bot's signals making it more profitable while adding more trades to my arsenal and therefore making me more profitable. However, if you're trading a sufficiently liquid instrument like ES or the euro, why trade multiple markets when you could simply increase lot size in a single market? Thanks.
ReplyDeleteTicha, if you do that and you have something good, you may be sharing it with the world.
ReplyDeleteRoachApproach, I believe in Money Management which can provide an edge where no edge exists.
RoachApproach, If I optimize a years trading and see that over the year a profit of, say, $1000, and stopping is better than not stopping, it works for me.
Otto, there's a lot on the web that can help. Try building a template into which you can add and change entry and exit conditiond then you don't have to be such a good programmer.
Sam, correlation doesn't worry me because each market has to stand up on its own. I assume I am fully at risk in each market and my drawdown analysis provides me with risk management info.
the haul, MP is hard to integrate except where the VA is.
TD, yes, that makes a lot of sense.
Rino, I am wary of making changes to fit current markets as I don't want to curve fit. I would rather get an "average" long term performance metrics that have predictable outcomes than make a change where I don't know what my possible drawdown may be.
Milt, I'm not sure I accept that outlier concept where there is a high win rate. Its more appropriate to a low win rate with winning trades much larger than losing ones. I accept that it works for Don due to his trading style. It's not a way of trading that works for many and seems to work better in very volatile markets.
Anon 23:36, we're testing against a pretty good SIM engine. At the moment we're getting the same win rate as theoretical but only capturing half the profits. Due to the very small targets and entering and exiting on limit orders, slippage is not a concern. 3 years ago, my answer would have been different but the technology has moved forward so far that autotrading is a better deal now.
Aloha, Kaloko.
Brian, As I'm getting older, the more sense autotrading makes.
Anon, 03:31, Markets get very quiet from time to time and I want to keep earning.
Tom / Kiki,
ReplyDeleteI am a believer in auto trading. Isn't 70% of the daily ES volume generated by computers?
I am happy to volunteer as a beta tester on a live account :-)
"...Due to the very small targets and entering and exiting on limit orders, slippage is not a concern."
ReplyDeleteNo it is not, but limit orders always getting a fill is. To be realistic and stringent in developing automation with limit orders you have to assume that all of your losing trades will get filled and perhaps only 80% of your potential winners will actually get filled for an entry and of those, less than 100% will get filled at your limit exit. Depending on how tight your profit target is and your plan "B" exit logic, a few of those may come back to breakeven. You are going to incur "statistical slippage". If out of 10 opportunities you only get 8 entries the win/loss ratio is going to change. Add at least 1 tick slippage per round turn and if the numbers still look good then perhaps you've got something.
Re: anon's limit order fills.
ReplyDeleteI agree.I have programmed my version of the system and run it daily with very good results on the ES using Tradestation. (It is not yet possible on Tradestation to program a strategy using OrderIntrabarGeneration on Momentum bars)
I have had to enter using an order 1 tick beyond the entry bar in order to assure a fill. This forces a larger stop loss .. but still produces good results.
I found that using volume to confirm an entry was also a prerequisite to getting reliable trade entries.
I was referred to your blog by my friend Glenn and I've been a big fan for many months now.
ReplyDeleteAs a former NYMEX local, your methodology has allowed me to regain the edge I enjoyed in the ring, and I thank you and Kiki for your selfless instruction.
Unleashing Flo, I agree, would indeed be unanimously and enthusiastically received by your many followers (myself included).
Do you think widespread use could skew the results in the future if a few hundred practitioners all jump in the same market at the same time, or do you believe the markets are large, liquid and diverse enough to shrug that off?
Thank you both, again, for your hard (and outstanding) work.
DNO
p.s.
I've since cancelled my subscription to IBD,
between you guys and Glenn, i have all the artillery I need for a successful trading day;)