Welcome to the new Traders Laboratory! Please bear with us as we finish the migration over the next few days. If you find any issues, want to leave feedback, get in touch with us, or offer suggestions please post to the Support forum here.
dangermouseb
-
Content Count
39 -
Joined
-
Last visited
Posts posted by dangermouseb
-
-
I recommend you take your time. Even with a PhD in stats it'll take you a long while to figure out how to make money trading. There are many people who will happily charge you for the pleasure of listening to them. They sound convincing but until you know how to assess accurately what they say don't believe them.
-- DM
-
Python, pandas, numpy, scipy, matplotlib, r and rpy would be my recommendation now.
-
Would you care to post evidence of your 80% sucess rate?
Best
-- DM
-
This is basic probability / random process theory. Thx for the post. It amazes me that so many people argue with it.
So my questions to you Predictor are 1) given that you're careful in your discussion so far, do you have statistical evidence that your techniques work? 2) if so are you willing to share that evidence - publically or privately (I'm not initially interested in the mechanics of the technique)?
-- DM
-
>> BTW about my actual trading, its way ahead of keeping XL journals, lets drop it altogether.
Apologies - don't mean to preach to the choir.
>> Yes I know, but I mentioned for those who may be interested to track live trading. (Historically everything looks obvious)
I've looked at too many systems where people think they have an edge but don't so now I don't even bother unless I can get a good cross-validated information ratio from their results first, and their method is sound (including record keeping in an electronic form). Saves a lot of time.
Most recent one was a friend who wanted me to implement a wavelet thingy but threw away his broker records on a daily basis!!
Next!!!
;o)
-- DM
-
9 trades isn't enough to make any statistical inference from. I'll have a look when you have 30+.
You might want to keep exact records (e.g. in XL) for your own analysis purposes. (simplest form is entry time, level, size, and same for exit, I also keep a list of all my stop levels and time) additionally I keep notes on my thoughts and feelings as a journal - though that is becoming less relevant as I become more statistically focussed.
-- DM
-
I'm not certain we are talking the same thing. The point I'm putting forward is that an entry or exit can't have an edge by itself. So without a specification of how to exit (which could be as simple as a coin flip) it is not possible to say an entry has an edge.
I'm not sure what you mean by position sizing. Van Tharps is about managing your betsize to be consistent with your objectives. If I consistently betted 10 times the size on the worst performing stock in your portfolio then I'd have a very different outcome compared if I betted 10 times size on the best performing stock. Size of the trade is part of the strategy. Expectation, volatility and covariation with other postions are key. Unconditional probability of reaching a target doesn't mean much.
I'll have a look later at you trades - do you have them in spreadsheet form?
Cheers
-- DM
-
DoOrDie - I think you've missed my point. It's the expectation that counts not the probability. As I said I can easily create a system that has 90% winners but has no edge. I've yet to see a traders account that opens the scale of their edge (not the method) up to analysis. I notice a tendancy that people like to say they are winning but don't like to be accountable.
I only had a brief look at the page you post but it doesn't look like people are posting enough information. It's not possible to analyse "it's going to x", without knowing what action's would be taken if it does something different. So saying it's going to x before it takes out my stop at y (or some other exit measure) is possible to analyse.
-- DM
-
Joshdance thx for the analysis - I did get you mixed up with the original. Apologies. I was hoping to flush out the BS which you did for me.
It strikes me that people say all sorts of nonsense that cannot be backed up empirically. Especially like x happens y% of the time. I can construct a trading system that makes money 90% of the time with no difficulty. The challenge is that when it looses it looses 9x as much as it wins so overall I have no edge!!! ;o)
I've not met a trader (yet) who makes strong claims (50.5% winners would be my cutoff) and is willing to put their trades up for analysis.
ValueTrader - great quote - how do I become a client
ATB
DM
-
Anyway Josh... how did you determine 80% accuracy? would be interested to see your method and results... thx
-
How have you determined that it is 80% correct? Backtest?
-
How do you calculate the value area? top 80%, bottom 80%, 40% either side of the median?
-
Seems pretty straightforward.
Everybody looses the bid-ask spread and commisions on average.
Anybody taking more risk than is sensible for their account will loose because they will hit the point where they are broke. (Basic position sizing).
Whether that corresponds to 90% of traders or not I don't know.
Anyone with zero edge who places appropiate betsizes will gradually drift down at a rate determined by the bid ask spread and commissions with noise superimposed.
I assume then as people blow up their account they then look for things to blame (e.g. wrong indicator etc) rather than just admitting they have no edge yet.
-- DM
-
yup that's the stuff.. kinda the rigourous side of market profile and wychoff... any pointers to where to find a good "quant" forum?
DM
-
I mean Micro Structure Models as in Hasbrouck and O'Hara.
Every now and then someone here postulates a model for microstructure (last I saw was p=1/3 for up, p=1/3 for same, p=1/3 for down) and I'm up for discussion, but can't figure where people talk about this of even if Traders Lab is the place.
-- DM
-
Hi,
Is anyone here interested in microstructure models? Or would anyone be able to recommend a place / forum etc to talk about them?
Thanks in advance
DM
-
@dangermouse ... If you had actually read my posts and had your brain switched on, then you would know this. But this is the crux of the problem. Like so many people in the world Danger, you represent so many, who want something handed to them on a platter, but even then that is not enough, as so many people are simply non-thinking zombies, incapable of using the measgre intelligence they have..You know you're right... I appologize for not having my brain switched on and wanting something handed to me on a platter and being simply a non-thinking zombie... I now realise the error of my ways...
Thank you for your insight and wisdom to my situation...
:doh:
-
@Adrian - The question on this thread is "are markets random or not".
The question is how do you know the market's aren't random.
It's a shame this thread has gone so off thread with people using it as a reason to argue with each other.
I was offering also to generate some random data on the back of the answers to challenge and explore anybodies answer, but so far noone has explained in a non hand wavy manner how they know the market isn't random.
:doh:
-
Goodness me this thread is going nowhere fast.
I'd like to know what thought processes / evidence people use to decide that market's aren't random. I would like to understand their line of reasoning and whay they come to their conclusions.
A question of everyone who's stated a "fact" - would you be willing to provide links to the research that supports your fact?
At least that way we can all evaluated the "evidence".
Examples include "it has been proven...", "it is obvious..", etc.
Personally I'm not interested in peoples opinions per se but in the process by which the come to those opinions.
-
Following on from "The Question of Randomness" thread. How do people know the markets aren't random? I'm up for creating some random data to see if someone can show me how they know it's random or not - just for fun... DM
-
current price at 10, stop loss at 5, stop profit at 20.
assuming arbitrage free random price process then expected P/L = 0 =>
0 = pStopLoss * lossAmount + pStopProfit * profitAmount
0 = pStopLoss * (-5) + pStopProfit * (10)
also pStopLoss + pStopProfit = 1
=> pStopLoss = 2/3, pStopProfit =1/3
also this implies it doesn't matter what your underlying process is
if you're wanting to think in these terms you should start learning some maths ;o)
best
DM
-
My 2p.
I think if the problem is framed as either 0% or 200% then it gets oversimplified. You trade your beliefs about position sizing and correlation. If you are very technical then you'd want to do some Monte-Carlo - a boot strap similar to van's approach (figure out your non-correlated R multiples and your correlated R multiples from backtest results), a permutation analysis (e.g. from Evidence Based Technical Analysis). You should also look at the sharpe ratio as well as Van's System Quality Number. The point is to be familiar to how much volatility is in your system and tuning the risk amount in relation to your capital to meet your objectives (e.g. 50% chance of making 50% per annum with a 10% chance of a 25% peak to trough drawdown).
Happy to provide further pointers.
DM
-
Good thoughts.
Olsen seems to be presenting a stochastic time view of the world - which is fine, it certainly helps with the fat tails seen in the market, but I'm not sure it helps to understand or determine the existence of trends in the price series or not. Also I'm not certain how much would be lost because of the one-way transformation from intrinsic time into clock time. My gut feel is that given the roughness of tools like Hurst Exponent estimation in the first place it wouldn't matter too much.
Not sure what you mean by "BIG problem is the incorporation of various volatility lags"?
--DM
-
I guess I see the idea of fractals like the idea of any stats - a high level summary. The auction process I see as a detail. The fact the auction process can account for trends would explain a fractal timeseries with trends. The auction process (although concretely in the moment by moment order book) happens at different timescales etc and a larger order at a different timescale my not manifest quickly or continuously in the order book. E.g. large orders get worked over hours, days and weeks not just at the smallest level and a volume by price summary is fractal itself depending on the period viewed.
I'm still at the beginning of fractals - I don't expect to use them as a trading strategy other than maybe getting good at Hurst Exponent estimation. The higher the exponent the more trending and I understand different stocks exhibit different exponents (and probably at different times - heresy!!!) so I can use that as a sonar to find fish as it were. Other than that fractals at least explain that there should be money on the table until all the price series have a Hurst of 0.5 or Brownian noise - in which case it'll be time to find another livelihood.
--DM
Introduce Yourself Here - Don't Be Shy!!
in Beginners Forum
Posted
As I said there are many who offer advice. Don't believe any of it until you've figured out how to evaluate advice - and don't believe this because I've said it but believe it because you've yourself understood it to be true. There's a good section on that in Rampaging Bulls.
Bon chance.
-- DM