What's your options success rate?
 |
 |
What's your options success rate?
In the event that you studied the picture above, I would be happy to answer queries. Took effort for this.
Require a small IS sample size, 7k bars, optimize an algo (with 5-6 requirements for it, like iRSI iBands etc.) for it, do so x amount of times with a neural network with no pre-concieved trading thoughts (what's random), making sure at least 64% or something. Then run all algo's generated with this % (we wish to unite them for more signs - but watch itsignal bars may double up and thus % could fall!) Via a primary OOS sample of true broker data, of let us say 100k bars (I just have 76k bars of specific broker data so that is what I use). Some won't hold up. Them, drop. Then run consequent algo's through 2 million bars orso. Again fall any which do not hold up (threashold 59 percent for example). Combine all algo's and filter our those that drag down the entire %. Ensure total% someplace 58-59%. 55.5% is needed for 1 minute 80 percent BO's.
$1/trade, ~2.5 weeks of data. $193 matches 30K bars example from drawing above
Same algo, with over 5 decades of data. $1 per transaction, 60.77%. Green bar is right, red bar is wrong. Red arrow reveals trades beign. All from algo compiled on 30K bars. Similar effects on less bars. See above. Ideal appears ~7k bars. 7200==one week of data. If the algo has a number of great combinations (try iRSI iMFI Close[I] Compare Close[I x] iADX iBands UPPER vs LOWER diameter Barsize) forget about overfitting - let the NN to create as accurate a formula as you can, and several of them, at a certain percent, with enough occurence, then combine them and finetune. Thougt; per-hour or per-day optimization may make sense also.
These are 1 minute transactions. 10 Minute transactions have lower payout, although higher %. 1 minute is preferred by me. Trading on pub close (minute =59). Obtaining timing directly (millisecond timer, 100 millisec per check) seems significant. Also broker oddities, like my broker disables trading if less then 7 price fluctuations occurred on this candle - that has to be contained in formulation (Volume[I]gt;=7) and they disable trading for several hours ( TimeHour(Time[I])! =0 etc.) All this should maintain one's algo, otherwise you cannot optimize correctly.
Hmm. Also need to incorporate draws better. I.e.% check and balance increase/decreas on gains/losses only.
Thanks for the image. I had a lot of joy recreating the specific chart within JavaScript that I will use in the future to better rate my systems. :--RRB-.
Only 1 thing, can you tell me the way you defined poor systems? This seems like an arbitrary announcement.
Hi roel13. .
Great set of recent articles... tracking your evaluation frame believing. .
And the use of programming skills to churn information to pinpoint odds.
IMHO Binaries give them selves for this type of analysis much more readily than Position.
Nice task.
But my ignorance of programming skills (depended for many years on an exIBM partner. . Regrettably no longer with us),
has left me dropped about some of your thoughts. .
Can U post a pdf or something in greater detail... logic-wise rather than program-wise?
For C challenged individuals like me. . ??
C
Reply With Quote