Well, I purchased it through e-signal, $150 per month, but the billing cycle is on the 9th, so I only have to pay pro-rated though then, and have this time to analyse it Looking through the backtesting tool on e-signal, the results are AMAZING! But, under closer examination, the backtesting tool gives you credit for the opening price of the bar that creates the signal Changed (a copy) of the code to use opening price of the NEXT bar, and results for all last year are mildly negative, with signifigant drawdowns I haven't spent a lot of time on this, so if anyone from e-signal has an explanation, you are more than welcome to counter this. But it does bug me that promotional efs are being released that still have this blatent slippage "feature". When the backtesting efs were first released last year, I was content to call it a new release bug. Why are new efs still doing this?!?
One possible explanation, that could mitigate the previous post, is that eMesa DOES give a message of "sell at tomorrow's opening -X points", or "Buy at tomorrow's opening + X points" So that COULD make the tool profitable, even with a negative backtesting result (using next bar), as the actual signals could be given "mid bar", thus making a realistic backtesting result meaningless So I guess what I am suggesting is that it may be impossible to use the backtesting tool for anything more than a visual of when the signals were given
A trading system should only measure based on the close of the bar. Sounds like crap... Keep me updated! LOL. "All results mildly negative, with significant drawdowns"... Love it! I don't think it's in their best interest to do so... I reckon it's because they're doing fine the way they're doing now... The slippage "feature" is probably a sales feature to make the system look better. LOL - Why is MSFT still selling Windows??? Keep us updated on eMESA ! - The system is either fly or die! ~Scientist
One other thing, in fairness to eMesa, is that their claims in the promotion were NOT equal to he backtested result that e-signal's tool gives you. although they did mention that slippage and commissions were not counted. I think it is safe to say that slippage would not be equal to an entire bar, as changing the tool to "Nextbar" would suggest. So time will tell
I don't see how this stuff could ever work. Most of these transforms assume a stationary time series. All the cajoling needed to make them stationary greatly diminishes the very information that is important to trading. Their real application may be as a "better" moving average for the DISCRETIONARY trader. nitro
Maybe we should ask Alex Elder? From reading his last book I think he's probably a credible enough sort of trader. On his site ( www.elder.com ) he endorses the MESA software, saying: MESA Software John Ehlers The best software for finding cycles in market data. The author is recognized as the worldâs leading expert in cycles-based trading. I just emailed him and invited him to ET. See what he says. ~Scientist
I fully agree with nitro. As a control engineer I have been following for years the mesa business. I would refer to it as scientism: i.e. the application of scientific looking tools to areas where they do not make sense. I cannot understand why TASC allowed this guy to publish his articles for years without any requirement of proof of validity. I do not recall having ever seen an article where serious backtesting is published. As to its applicability in discretionary hands, I would classify this with astrology.
OK, I'm going to eat some crow here, the efs provided by emesa was not in the backtesting directory, and I used btEMA by mistake btEMA is actually exp moving avg, and it used the THISBAR, but it was dated y 2002 The eMeas efs DOES use the NEXTBAR, so my remarks about e-signal not fixing this on new efs was wrong. I havn't had the chance to backtest using the correct efs, but felt I needed to get this correction out pronto, and my apologies to e-signal/emesa
well, did a few backtests going back to 5/2/2, results are +$8487.14 on one contract, whick isnt bad considering "NEXTBAR" gives you the penalty of opening price on nextbar, whereas the signals are generated somehere midbar nearly everything I have previously backtested using e-signal's backtesting tool gives negative results when using nextbar, so this isnt bad. eMesa claims $15,771 for the last year, and this could be possible with trades placed somewhere "midbar" should be interesting...time will tell