The opinions expressed in these forums do not represent those of C2, and any discussion of profit/loss
is not indicative of future performance or success.
There is a substantial risk of loss in trading. You should therefore carefully consider
whether such trading is suitable for you in light of your financial condition. You should read,
understand, and consider the Risk Disclosure Statement that is provided by your broker
before you consider trading. Most people who trade lose money.
I am really confused. Everyone keeps asking me about ‘backtesting’ of my system. Do they want me to make up a nice excel spreadsheet showing my ratios and high powered math equations relating to my winning trades that will guarantee them a future profit? Do my past profits somehow prove my future success? Is everyone a math genius here and thinks that trading is just about running numbers? That this game is broken down into nice neat puzzle pieces that can be fit together perfectly?
I Understand that I am new and my track record isn’t long on here and really have no room to whine as I do have a good amount of subs for my time… But this ‘backtesting’ nonsense is absurd to me. Why can’t people open their eyes and see Whats in front of them without mountains of irrelevant data?
As an aside, do you think anyone views a sub price of $49 and $99 differently? I mean I make more than that with every trade, would the same people not sign up either price? Not sure my theory on the low price beginning was actually very bright on my part, seems all reputable systems charge 100+.
There are plenty of “systems” that only work for three to five years and then break down. By backtesting your system over an extended period of time (15-20 years), you make sure that your system works in a variety of market conditions. Also, you want to know how your system would have performed in extreme market conditions.
While your performance this month may not be indicative of your performance next month, your performance over a 20-year period is the mean to which your future performance is likely to reverse.
Backtesting = Curve fitting . Its meaningless and misleading .
And how the subscriber is going to determine whether these backtesting results are authentic or not ?
I had a subscriber that was so enamored with backtesting (which I refused to provide for the reasons @TSH posted), he chose to leave despite the strategy’s strong C2 results while he was a subscriber. I guess making real money was subordinated to receiving super hypothetical reports. He then went on to write a one star review complaining about the matter.
Let me put a client’s perspective here. I know nothing about system developing so I do not have that bias.
For me, a backtest is the “ideal” circumstances the system could perform at. I assume that the system developers HAS curve fitted, the equity lines are exaggerated/unrealistic and the DD is lower than expected. So if a developer is showing me 25K profit a year with a 25% max historical peak to valley DD, I know the line of thinking the developer has. So if the DD goes to 50% in a non-extreme market scenario and made 10K total per annum, I know that the system has gotten “off track.”
Does this mean that I am out the door no matter what? Absolutely not, it just provides another data point that when aggregated with other data points, can lead to an informed decision. I think the fear from developers and ignorance from the clients is that these curves are taken as ACTUAL trades, even when explicitly stating that is not the case. So if a DD goes 1% above the historical max DD, it is game over for some clients; that is the fear.
I also want to say that I have seen and experienced some quite accurate hypotheticals. Nobody can predict how a system will go but if your backtesting shows profits for every month and then when live trading happens you are often in the red, perhaps you need to reevaluate. Personally I trust a developer much more and will be much more likely to fight on through a DD and stay on as a sub, if the backtesting data shows believable moments of failure.
in conclusion: backtesting data has value to clients as long as they know what they are looking at. I wouldn’t recoil as much as some developers have, nor would I be obsessed about obtaining the data because it is what it is: a “made-up” hypothetical line.
Potential subscriber knows nothing about your trading idea, except some phrases in the description. Why can’t you give him some more data? Good backtest results show that at least in the past your trading idea worked. It is only one conclusion that can be made based on the backtests.
In theory, a 20 year backtest for equities would have more value than, say, a 2 year backtest. With respect to futures, however, finding clean TICK data for 20 years at a reasonable price is a near impossibility especially in markets that are traded electronically.
Personally, I use backtests for one purpose only, and that is to determine whether the system has a chance of making money. If it fails this test, I don’t spend any more time on it.
In theory, a 20 year backtest for equities would have more value than, say, a 2 year backtest.
In practice as well. Charlie Munger joined Warren Buffett after closing down his own firm, which he had to do because he delivered two years of 30%+ losses in a row in the early 1970s and his clients simply took their money and ran.
Personally, I use backtests for one purpose only, and that is to determine whether the system has a chance of making money.
Precisely. The next questions would be, how much money the strategy can make, at what cost in terms of risk, and, in case of alternative asset classes or long-short, what diversification benefits it can offer to investors who already have stock-and-bond portfolios.
I expected something like this. System trader never says that Back testing equal to Curve Fitting, and he’ll be right. Curve Fitting is mainly caused by Inaccurate Optimization of trading system parameters. Back testing itself is harmless procedure.
In and of itself, this is okay, as long as you disclose it to clients properly. Clients need to understand that they face a possibility of catastrophic loss in case you have a heart attack or get hit by the proverbial bus. But this is not the only possible approach to the investment process; other people may choose to do it differently.
Was Charlie trading an algorithmic system for stocks?
Essentially. Value investing is the original algorithmic system. The first investor known to systematically employ quantitative screening was none other than Benjamin Graham all the way back in 1920s.
Are you confusing a track record with a backtest?
For the purposes of this discussion, the difference is immaterial. Any performance history, whether actual or honestly simulated (as opposed to shamelessly manufactured), is likely to have outlier periods, uncharacteristic of the strategy’s long-term performance. Charlie Munger’s actual performance looks a lot like backtests of Benjamin Graham’s screens that Henry Oppenheimer ran for his 1986 paper, “Ben Graham’s Net Current Asset Values: A Performance Update”; both performed abysmally in 1973 and 1974. Munger’s actuals were -32% and -32% again, Oppenheimer’s backtest came back with -43% and -19%. This is why you need long backtests; it helps to have an educated guess about the duration and magnitude of your worst case…
It’s clear you know your value investing history, and I appreciate your point of view, but with all due respect, I find the comparison of fundamental analysis on stocks circa Graham, et. al. with today’s algorithmic systems that can crunch a humongous amount of data (much more than a simple balance sheet and ratios), like comparing apples to oranges.
But I am curious why Buffet would hire Munger given the latters lousy performance.
I find the comparison of fundamental analysis on stocks circa Graham, et. al. with today’s algorithmic systems that can crunch a humongous amount of data (much more than a simple balance sheet and ratios), like comparing apples to oranges.
You are entitled to your opinion, but it doesn’t change the fact that Graham invented quantitative screening and did so at a time when “computer” was a job title for a human being proficient in the dark art of slide rule (that’s the weapon geeks carried before pocket calculator was invented). Nor was Graham’s analysis strictly fundamental; it incorporated market measures (specifically, AAA corporate bond yield) as well.
I am curious why Buffet would hire Munger given the latters lousy performance.
First of all, Buffett didn’t “hire” Munger; he took him on as a partner. As to why, please recall why we are having this conversation. We are debating relative merits of short(er) and long(er) performance histories. Munger’s performance history at the time was 13 years long, the last two being a complete disaster. Buffett, who was likely experiencing a similar misfortune, believed that the last two years must have been outliers, which, while hard to stomach, were not representative of the long-term performance.
I must say, rather valid points by the naysayers and the proponents- HOWEVER
My main complaints are still:
I could create a backtest ‘analysis’ based on what I think you want to see… By using hind sight I could create any backtest that fits the criteria I think you are looking for (sharpe ratio, alpha, calmar, etc). Would you like overall gains? Or risk adjusted gains?
backtesting is backwards looking and IRELLEVANT. All that matters is what you believe you are RECEIVING IN THE FUTURE in exchange for your hard earned money!
Why would you pay for past performance? We can all manifest a ‘perfect system’ if we have all the concrete data in the world at our disposal.
Google “reverse engineering”
Subscribe to my system, (Collective2.com/system92195189) before I raise the price to $99.99 from an artificially depressed price of $49.00. (Free 7 day trial)
If you would like to purchase unlimited ‘theoretical’ backtests, I do require a one time fee of $99.00, and will supply any of a multitude of backtesting algorithms at your request. I am 100% certain you will be confident with my system after viewing my ‘time-tested’ results!
