How do YOU come up with trading ideas?

[Handle123]

Looking for traders to share how they come up with hypotheses about the market to backtest. What is your process (if there is one)? How do you look for an edge to exploit?

As you study Price, eventually you find how it works, and it is mostly where to screw the little trader, somebody has to feed us the profits. First couple years you waste trying to get best entries and hardly anything on money management, five ways in and one way out. NOW, I spend time finding 3 ways in and 45 money management rules.

Lots of books on trading, best are on Price Action, maybe one indicator like moving ave, you will find price patterns, patterns with moving ave, volume patterns, patterns too much volatility to trade. After awhile you have 90 ideas to test and then goes the weekend.

 

[d08]

altucher teaches how to get ideas,

http://www.jamesaltucher.com/2014/05/the-ultimate-guide-for-becoming-an-idea-machine/

His recommendations are probably worth as much as his books, nothing.

On the topic of ideas...
A great source are blogs, while most are useless there are some really interesting ones out there that make you think and test.
Books are full of ideas, none of them worth much. I've concluded that most gifted traders don't write overly long books, they just present the numbers with a few lines for explanation.
Listening to good music helps creative thinking.
Taking breaks from obsessive pattern seeking is necessary to allow the subconscious to function, it's often much better than the conscious mind.
It still takes years of price watching to realize the full potential of any idea.

 

[Aileron]

I bull**** around until something pops out at me

 

[arbs-r-us]

Did Nitro leave trading?:eek:

Likely a "forced" exit. He was short spooz for nearly 700 handles.

 

[tom_czr]

"How do YOU come up with trading ideas?"

Lurking on forums, watching markets, live trading, improving existing strategies, statistical research of everything possible and unusal approach to everything.

What have I discovered and what is working?
1. I am not trading most liquid futures contracts, because my system on CL (for example) has two times better performance on CL five months to expiration, where there is not too much intraday guys fighting for each dollar.
2. Strategies, that are difficult to test and trade, because there is no publicly available software for them - like trading pairs, triplets, different baskets of stocks. For newbies it is much more difficult than simple long and short speculation so in this area there is lot less competition = easier money for me.
3. etc...

 

[NoDoji]

There’s a problem solving method that elementary school children often use called “Guess & Check”. It involves making an educated guess based on what is observed or known, then applying the guessed answer to the problem to see if it checks out.

This can be a useful method of putting a trade idea to the test, and possibly ending up with an entire trading plan.

Suppose you observe something in the market that seems to repeat a lot. You can describe this pattern or signal or event in precise terms, then apply the description to every instance of it over, say, 50 to 100 trades and log the results (maximum favorable and adverse price moves from entry) for analysis. You’re guessing you’ve found a high odds idea, then you’re checking it out through careful backtesting, followed by forward testing (sim or replay).

Here are several things I noticed as a beginner that I ended up trading off of, starting with the event that sparked my first trade ever:

1. When a fundamentally sound and strong stock is dragged significantly lower by news that doesn’t affect its fundamentals, it’s likely to bounce nicely.

2. When a company raises guidance and its stock jumps higher, it’s likely to continue higher for at least a couple days.

3. When a stock hits a 52-week high and trades higher for a few days, it’s likely to pull back nicely.

4. When a heavily shorted stock announces any kind of news that’s better than expected (even a huge loss that’s less bad than expected), it’s likely to continue higher during the day of the news.

5. When a stock hits the hi ticker or low ticker more than 80 times, it’s likely to end up hitting it more than 100 times that day.

What happened with these ideas is I noticed a pattern, then decided to trade the next time I saw a similar signal.

I had amazing successes with each of these ideas; unfortunately, I never developed a robust plan, meaning I skipped the careful analysis and risk management end of things, and eventually ended up trading a lot based on my opinion instead of a specific plan. Needless to say, I went from profitable to quite unprofitable in no time at all.

When you see something happen in the market and think to yourself, “That seems to happen a lot” or “I see that all the time”, define exactly what that is, and check it out thoroughly by analyzing a series of 50 to 100 appearances of that. If you come up with something that works more often than not and nets you a profit after accounting for slippage and commissions, you likely have the beginning of a viable trading plan.

 

[eurusdzn]

"How do YOU come up with trading ideas?"

Lurking on forums, watching markets, live trading, improving existing strategies, statistical research of everything possible and unusal approach to everything.

What have I discovered and what is working?
1. I am not trading most liquid futures contracts, because my system on CL (for example) has two times better performance on CL five months to expiration, where there is not too much intraday guys fighting for each dollar.
2. Strategies, that are difficult to test and trade, because there is no publicly available software for them - like trading pairs, triplets, different baskets of stocks. For newbies it is much more difficult than simple long and short speculation so in this area there is lot less competition = easier money for me.
3. etc...

I would like to hear your opinion on "2."
I think the only exclusions would be the "Joe the plumber next door who trades as well". If there are many thousands of equity spreads or triplets i would think that
any firm or actual trader has the tools to find,test and trade these prices.
I as well see ranging spreads such as BNO/FXI and dont get what causes this to behave as such. Yeah, cointegration, fundamentals, statistics etc...
So, when i see a well behaving spread my question is "Are there big market players at these extremes and s/r levels effecting price or is it just bullshit random price oscillations.
Seriously, will traders effect the price of brent crude and China 25 index due to
BNO/FXI touching support? I doubt it.
Problem is that my testing software tells me it does and i have a great system here. I think not.
What am i missing here. Thanks.

 

[tom_czr]

I would like to hear your opinion on "2."
I think the only exclusions would be the "Joe the plumber next door who trades as well". If there are many thousands of equity spreads or triplets i would think that
any firm or actual trader has the tools to find,test and trade these prices.
I as well see ranging spreads such as BNO/FXI and dont get what causes this to behave as such. Yeah, cointegration, fundamentals, statistics etc...
So, when i see a well behaving spread my question is "Are there big market players at these extremes and s/r levels effecting price or is it just bullshit random price oscillations.
Seriously, will traders effect the price of brent crude and China 25 index due to
BNO/FXI touching support? I doubt it.
Problem is that my testing software tells me it does and i have a great system here. I think not.
What am i missing here. Thanks.

a) BNO-FXI - hmmm, I have never had this pair in my screening as a good one.
b) "Are there big market players at these extremes and s/r levels effecting price or is it just bullshit random price oscillations" - who knows, maybe both together
c) "will traders effect the price of brent crude and China 25 index due to
BNO/FXI touching support? I doubt it." - I don't
d) "Problem is that my testing software tells me it does and i have a great system here. I think not." - Why do you think so?

BTW Trading pairs/triplets has lot of pitfalls, like using simple ratio model, correlation instead of cointegration. If cointegration is used (for example Johansen test), then people do not know what level of significance is good and by the way, there is no easy way for rookies to calculate with what probability pair/triplet is/is not cointegrated.
Then how to manage whole portfolio? What about earnings? What about Beta? What is proper selection method for pairs we plan to trade? Most of people are not successful even if I tell them about some tricky things here... and it is good for others, who are able to find the way

 

[tom_czr]

When you see something happen in the market and think to yourself, “That seems to happen a lot” or “I see that all the time”, define exactly what that is, and check it out thoroughly by analyzing a series of 50 to 100 appearances of that. If you come up with something that works more often than not and nets you a profit after accounting for slippage and commissions, you likely have the beginning of a viable trading plan.

Very well written.
Maybe we can highlight word LIKELY in your post.

One of my friends has presented system based on FAS/FAZ and was explaining that this is working because of time decay of leveraged ETFs.
I have done some pretty crazy simulations and tests which proved, that his sytem, though profitable in tests and live, is in fact just matter of luck and that there is no time decay in those ETFs... so it was only question of time when he will suffer some serious losses.
So it is very important to understand why exactly are our systems working, what is the reason. Sometimes only simple backtest is not enough.

 

[NoDoji]

I have done some pretty crazy simulations and tests which proved, that his sytem, though profitable in tests and live, is in fact just matter of luck and that there is no time decay in those ETFs... so it was only question of time when he will suffer some serious losses.
So it is very important to understand why exactly are our systems working, what is the reason. Sometimes only simple backtest is not enough.

The real danger comes when something just seems to work again and again and again and we decide there's no need for risk management. That was my undoing following an amazing run of beginner's luck.