Are the tester & demo trading both BS?
backtesting is for debugging your code.
-nowadays mostly all brokers have variable spread, which isn't on tester.
-ticks are not real in tester, they are calculated based on M1
-no requotes no offquotes, you can send 100 Orders at the same price
-tester offers you the possibility to optimize.
enought? all these will tested result look better than forward trading.
-
-
Demo vs Live
on most bokers the spread differs extremely. best example is alpari.uk. in demo account you have the "same" spread as on pro accounts, but you don't pay commision.
other brokers have better spread in demo mode.
if your strategy's profitness depend on 2 pips per trader this faktors are substantially.
Why does it take so much digging to find out about anapparently fundamental differences between using the two accounts?
it isn't you have just to read the answers we already have given you. this is now the third thread on the same subject!
-
-Z
IMHO, backtester has its strength and weakness. Use it for what it is good at, not what it's bad at. It is essential for testing functionality... did it calculate all your variables correctly, enter as you expected, trail as expected and exit as expected. Did it send all the notifications it was meant to. You can't code effectively without it.
It's also good at strategy testing in the true sence of the word. over thousands of iterations, if x, y, z happens what is the probability price behaves in a desired / percieved manner. Then what can we learn to exploit that.
What it's not good at, imho, is strategy refinement. The backtest process, can't replicate real time. Variable spreads, latency, connection loss changing markets etc. so it will never produce drawdown or profit factor or... (pick your stat) that will be replicated in a live environment. And thinking of variable spreads, I don't think these variances are unique to scalpers. I normally watch tester in visual mode (looking at functionality). Many times, say on a short, I watch Bid cross the TP line but Ask miss it and eventualy close on the stop 50 pips or so worse off. If the spread was 1.5 pips instead of 2, My latest draft is 50 pips better or if I rounded the TP to 4 digits rather than 5, a superior EA. You can't use tester to determin the quality of refinements around a base strategy. All you are doing is curve fitting and tester moves from being a useful tool into being a hope generator.
Just because I missed a target by a fraction of a pip, does that mean my strategy is fundamentally flawed? Should I change my strategy based on one report to the next? The answer, imho, is don't know. There is nothing in the backtester report that can help you. You need to look at your strategy more fundamentally than the net profit. So is tester worthless? Only if you are using it for something it can't do!
V
Viffer: .... And thinking of variable spreads, I don't think these variances are unique to scalpers ....
That is very true. We make allot of fuss about scalper,scalper,spread,spread. Truth is any strategy can fall victim to the I'm off by 1 point. Just yesterday, I tried to place some manual orders and 3 times in a row I was off by either the Bid or the Ask. 1st was the sell_limit and price went up to miss it within a point before plunging down_ward. But I was patient and price came back to fill my order. Then next was the take_profit all I was asking for was 12.pips again came within a point to miss and spike back up. Then came the stop_loss, I was like nah-eh I'm moving this baby to 48.pips away n good distance from any Candle body or wick_endings. Wanna take a guess what happened? Yep, price went up to meet it exactly before plunging back down again.
It this point I'm thinking the brokers are surly after me. Not with standing, this was on a Demo account with Fixed Spreads. Yeah, they were after my monopoly money all right ;). So all this talk about strategy tester being bad or deceiving etc. Meta_Trader is a free program and much more professional looking and features including than any proprietary broker software I've seen. What should we do? Call the Forex police because the free software we were given doesn't let us simulate Real_Ticks, Broker_Spreads, Server_Latency, Broker Re_quoits and Oh by the way give me a feature called "Change my Mind" in the back_tester. Because I would have definitely changed my mind when my strategy missed by one Point in reality.
All this conversation about the back_tester is all good to help one person understand what we all had to learn over time. But I bet you some Demo money, within the next 48 hours someone else bust in here with "OMG the back_tester is Un_Real" and expect everyone to be excited about that topic. And when we say search, they get offended.
Are the tester & demo trading both BS?
in my opinion also most "MM" in the world r BS 2
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
I hope I'm not offending anyone's search before posting sensibilities, as it seems to be a provacative subject here. Proceeding in the hope that this is more than just an archived site, please note that I've searched and I've posted to a relevant thread, and I've learned a bit about this subject. But I still don't have an answer to two concise, basic questions.
...
Background: backtesting results vary profoundly when logged into a live vs a demo, account. Not even close, totally different curves, winner in one vs loser in the other. I ask the following in reference to previous threads on this site (please search the site before you answer)
...
Q1: Do the anomalies of backtesting while logged into a live vs a demo account, necessarily apply when trading forward in both cases? Will the same EA trading forward live alongside a forward demo, have the same results? Or do the differences in tick stream, market info or other still apply?
q1: If the demo result is still different going forward, then what is the point of using a demo account? The demo account is pitched as a way to test things out before opening a live account. If it produces profoundly different results than live trading, then this pitch seems fraudulent. Forex.com told me to start using the demo, that it was the same as live trading, only on paper.
...
Q2: Does backtesting while logged in live, utilize substantially the same data as trading forward live? As close as can practicably be?
q2: If not, what is the point of the tester?
...
Q3 Extra credit : (for conspiracy theorists): I have several EA's that are backtest winners in demo and losers live and none I believe, the other way around. Is it possible there's bias somehow applied to make for better results in demo? The answer notwithstanding, why, with such anomaliies in the results and so many others asking about it, does MetaQuotes or brokers or someone not make a point of clarifying or indeed, at least pointing out, the issue? Why does it take so much digging to find out about apparently fundamental differences between using the two accounts?
...
I see that some speak almost religiously about the importance of backtesting. But this tester has so far given me very little confidence to do other than test forward with live micro accounts.
...
I would really like to know what others think about this.