I don’t understand why a signals realism factor doesn’t seem to get much attention? When I put my signals out there I tried to attain the highest realism factor I could. Isn’t that the most important thing after profits? It just seems like there are so many signals that can’t be duplicated with real money. Are people really subscribing to signals that can’t be duplicated? Thanks
Richard,
The realism factor is misleading in some cases. A good example is Weekend Trader, with a RF of 26.7.
This system only trades NYSE listed stocks Mondays at the open. Because of the opening auction any market order submitted before the open will be filled at the open guaranteed. Because C2 also fills at the open, the system has a RF of 100% by definition. Of course, profits might deteriorate if more and more people start trading the system, but they should still get the same fills as shown on C2.
Another example is Entropia, with a RF of 99%. I auto traded this system for a while using different brokers and found that both slippage and transaction cost led to quite different results across the different forex accounts I had. IOW the RF depends on the kind of broker.
In other cases, when the RF is based on real life fills, the RF might be accurate for the average trader. But nobody knows what “average” means. If you’re above (below) the average volume your true RF might be lower (higher).
So, it’s a statistic that’s very difficult to interpret.
> The realism factor is misleading in some cases. A good example is Weekend Trader, with a RF of 26.7.
> This system only trades NYSE listed stocks Mondays at the open. Because of the opening auction any market order submitted before the open will be filled at the open guaranteed.
Are you sure about this? Is it the open price or the opening range? Moreover, wouldn’t actual large orders (submitted before the open) move the “open” price anyway?
> Because C2 also fills at the open, the system has a RF of 100% by definition.
Here is an actual trade:
"BTO 3,974 KOR 9.53"
KOR often has volume at the open way less than 3,974 (often the entire first hour’s volume is a small fraction of 3,974). Assume 10 subscribers and the order could be a market order for 39,740 (more than the total daily volume for KOR most days). How can the RF be “100% by definition”?
Of course if the vendor is paper trading and there are no subscribers there will be no real time impact on the opening price. So, the RF of 100 would apply to paper trades, but it might not apply to “real” trades.
Isn’t the concept of reality implicit in the RF?
Sam -
What Science Trader said is that on the NYSE you will get the opening price if you put in a market order before the market opens. Guaranteed. And this is correct. He stated in his post that if the volume is too high then the opening price will be affected. Should that affect the RF? It depends. If the profit per trade is much higher than how much the market would be affected then the RF why would the RF be penalized?
The problem that I have is that there are many systems with high RF (approaching 100) that:
- can’t be traded at all
- can be traded but only by persons who can watch the markets / positions all day long. (i.e. babysitting). For many people this is not real life.
- traded with such high leverage that the results are so misleading that one could consider the results not real.
In any case, it seems to me that in most cases RF is attempting to deal with a slippage issue and not necessarily doing a great job. If C2 is attempting to judge slippage then it should be called a slippage factor. Calling it a realism factor implies that the system can’t really be traded. This biases potential subscribers away from what could be a good system…
see http://www.nyse.com/productservices/nysearcaequities/1157623605117.html for an explanation of the auction process. I think it’s actually the market order auction rather than the opening order auction. My understanding is that at 9.30 the market order auction will execute at a single price, which is the price at which the largest volume can be executed. This is consistent with my own experiences trading this system.
Of course a large number of subscribers could affect the auction price, and hence future profits of the system could deteriorate as the number of subscribers grows. But for the current subscriber base this effect is already present in the opening prices. As long as you consider your own volume as marginal compared to the combined volume of all current subscribers, you can expect the system to be 100% realistic in the short run.
It all depends on how you define the RF. If it’s defined as “did the C2 reported fills reflect real-life fills at the time of the trade”, it should be 100% for Weekend Trader. If it’s defined as “will the current C2 reported fills reflect future fills”, you need to make additional assumptions on the speed at which the subscriber base and its combined trading volume grows.
> This system only trades NYSE listed stocks Mondays at the open. Because of the opening auction any market order submitted before the open will be filled at the open guaranteed.
Weekend Trader holds stocks an average of six weeks, so price movement over a few minutes or hours is usually insignificant. The C2 realism factor is flawed because it doesn’t take into account trading frequency. Price slippage that will destroy a day trading system can be no problem at all for a longer term system.
Although Weekend Trader officially trades Monday at open, the system results are at least as good if trades are placed at other times such as Monday at close, Tuesday at open or Wednesday at open. There is a wide window of opportunity to enter and exit trades at favorable prices. This is because Weekend Trader is a fundamentals based system and does not need speedy entries and exits which are characteristic of technical systems.
You don’t need to buy/sell Monday at open unless you want to exactly match the C2 results. In my own trading, I take more time to enter and exit at favorable prices.
If you place market orders before the NYSE open, then you will almost always get the NYSE official open price. Very rarely you might get something different if trading is briefly suspended at open due to breaking news. This is very rare and I’ve seen it happen only once. Over the long term, breaking news should be neutral - sometimes helping, sometimes hurting performance.
Regards,
Anthony
> What Science Trader said is that on the NYSE you will get the opening price if you put in a market order before the market opens. Guaranteed.
Does this apply to odd lots? I was reading the NYSE odd lot rules
(weekend trader trades mostly odd lots) and it reads that odd lots get
filled on the “next” round lot trade price. So, look at SPM. It often has few or no trades in the first 30 minutes. Weekend Trader is trading 8491 shares. Where does he get filled? When the daily volume is 100 (like yesterday) does 8491 shares move the open a little or a lot? And that’s with zero subscribers. Is a RF of 100 (as ST suggests) fair? It doesn’t matter if you call it slippage or RF or whatever. And it really doesn’t matter what the NYSE rule is either: some of these trades are pure fantasy. Please just look and tell me in “reality” you think these
trades are “realistic”. Let’s not talk theory.
(I am NOT picking on WT. ST picked him out to prove his point).
>The problem that I have is that there are many systems with high RF (approaching 100) that:
> - can’t be traded at all
They should have lower RF’s then. I agree.
>- can be traded but only by persons who can watch the markets / positions all day long. (i.e. babysitting). For many people this is not real life.
That is another issue. C2 can’t be expected to know what is “real life” for
you, me, whoever.
> - traded with such high leverage that the results are so misleading that one could consider the results not real.
That is again another issue. This has been a long time suggestion to have C2 show a realistic P/L per unit or some alternative Cumu$ based on a single lot or similar equivalent.
Right now a losing system can not only have a positive P/L per unit, but a superior one. One system vendor who has a losing Cumu$ was touting his “highest on C2” P/L per unit.
> Weekend Trader holds stocks an average of six weeks, so price movement over a few minutes or hours is usually insignificant.
Not so. Look at the SPM trade. The total daily volume on those
days would barely handle your trade. An order of 8491 shares against
daily volume of 6,000 on 10.30.06 could have pushed the opening
price several cents or more. Same thing on the exit. The total
profit was .07. While you may have a complaint on other specific
trades, it’s hard to see how the SPM trade could have possible.
>The C2 realism factor is flawed because it doesn’t take into account trading frequency.
Sometimes. Sometimes frequency isn’t important. Let’s say I buy
10 esm07 on the open today. I exit two hours later. I could have perfect fills, and 10 subscribers could have the same fills. No problem at all.
Try the same thing with 8,000 shares of SPM with entries and exits six months apart. Unlikely the fills would look anything like what they do on
paper. A winning paper trade could lose in reality.
With few exceptions, NYSE stocks have surprisingly good liquidity - more so than trading volume suggests. I’ve found that trading on the NYSE is far better than NASDAQ when it comes to quality of execution. So far, concerns about Weekend Trader fills have been more theoretical than real. I suspect that many subscribers don’t necessarily trade at Monday open, but look to get better prices later on Monday or even Tuesday and Wednesday.
The important question is: does the system make you money consistent with C2 reported results? If it does, then the realism factor doesn’t matter a whole lot to you personally. Nevertheless, I will keep an eye on actual fills and make adjustments to filter out smaller stocks if liquidity becomes more than just a theoretical issue.
Regards,
Anthony
> Nevertheless, I will keep an eye on actual fills …
>…more than just a theoretical issue.
>>Look at the SPM trade.
What were the actual fills on the 9,000 SPM on the entry side? Since
the total volume that day was 6,000 I just wonder what the fill was on
the opening tick.
> I suspect that many subscribers don’t necessarily trade at Monday open, but look to get better prices later on Monday or even Tuesday and Wednesday.
Then we are getting into the complaint Steve had about other
systems: i.e. you need to manage the trades yourself. He suggested
those other systems should somehow be penalized on the RF because you can’t just auto-trade them or need to watch the trades
intraday. Indeed, a major selling point, and the name itself, imply you
don’t need to watch the trades:
"The “Weekend Trader” system is designed for the busy person who places trades on a weekly basis over the weekend …"
Also, I find it ironic that we are discussing the stats on this system
and no one noted this:
Cumu $ $61,997
after typical commission $59,131
and real-life slippage $739,279
or this:
P/L per unit $0.62
after typical commission $0.61
after real-life slippage $7.30
Actual trading results in your own trading account are what matter. C2 has flaws in realism calculations because with very few exceptions actual subscriber results with Weekend Trader do match C2 reported results.
I won’t engage further in this hypothetical back and forth discussion.
> and real-life slippage $739,279
All but 3 of the closed trades have Avg slippage = n/a, and these 3 have Avg slippage = 0. So I suspect that the error in the quoted result happens in the data of one of the open trades. Well, I hope it does, otherwise the algorithm to compute the "Cumu $ after commissions and real-life slippage" seems to have an error. The RF itself can still be right (or wrong) if it is not based directly on that $739,279.
"Look at the SPM trade.
What were the actual fills on the 9,000 SPM on the entry side? Since
the total volume that day was 6,000 I just wonder what the fill was on
the opening tick."
I have to agree with Sam. If you are trading even a fraction of the total daily volume then there is a problem. You should be filtering out such trades.
I had an EMAIL communication with Matthew last year (I’m not sure if he has changed the algorithm since then) and I believe the RF assumed that 5% of total daily volume would have an effect on the stock price. 1% would be insignificant.
That being said, last year I was concerned about my Realism Factor on the first system I traded here. My RF was 50ish. I modified my trading to only buy stocks with 20 day average daily dollar amount of $10M minimum. I was buying $10K worth of stocks at market at the opening (100 lots). So in general I was buying less than 0.1% of stocks daily dollar amount. I felt that I could support 10 maybe 20 subscribers with $100K accounts before slippage would start to be a factor. Yet my RF only went up to 0.70. At that point I was ready to give up because I could see many systems with very high RFs that were clearly not tradeable.
My opinion is that there should be the RF should be split into three: a Manual Trade RF, AutoTrade RF and a slippage factor.
Manual trading is for the masses (the overwhelming majority of investors). Amateur investors who want to put in their broker’s instructions and forget about them. There should be an RF number to guide them.
There should be an Auto-Trade RF which tells auto-trade addicts how effective autotrading is with a particular system. Systems with multiple orders and cancellations during the day should be heavily penalized since they need to be babysit all day long (auto-trading gets tripped up).
Leverage should impact RF for both manual and autotrading. The higher the leverage, the lower the RF.
Slippage factor should be separated from RF and consider limit orders converted to market, liquidity (as discussed in this thread), etc. Although you can graph the equity with realism and get slippage, it is still under the label “Realism”. I think people uniderstand the concept of slippage more than realism.
Steve
I agree that it is better to split the RF into distinct factors. This distinction is made in the trade details, but not in the overall RF. Imho it would be better to make these distinctions in the overall RF too.
This is similar to what we discussed in another thread about ratings and performance statistics. I think it is generally better to present conceptually different statistics apart instead of blending them together into one mixture by an – often somewhat arbitrary – formula. The latter will always result in a loss of information. Especially when the blending formula is a black box.
> Leverage should impact RF for both manual and autotrading. The higher the leverage, the lower the RF.
I don’t have a problem with much of what you say, and I agree abuse
of leverage should have consequences, but:
“higher the leverage = lower RF” is just wrong. Should a trader be penalized for trading 10 ES futures ($700,000 value) on the open compared to a trader trading 10,000 SPM ($40,000 value)? The emini trade is likely very realistic, while the SPM is a cloud of fluff. Your formula somehow would turn things upside down without any rational reason.
Once again lets deal with reality here. The emini has volume in the first half hour of 110,000 and SPM has volume of 100. Your formula would reward the SPM trade and penalize the ES. Why should the 20X leverage in the ES have any connection to the RF?
Prefer apples to apples? Would a 1 lot ES vs a 10 lot make any difference? A 5 lot vs a 30? Not a whole lot in the context of the ES
opening range volume.
"The emini trade is likely very realistic, while the SPM is a cloud of fluff"
I agree somewhat with the SPM argument Sam. However you can still buy SPM. It is just that the premium payed would be substantial. This translates into massive slippage.
The problem I see is that providers use leverage to generate spectacular returns. Without some Foolishness Factor (FF) to deflate these systems, potential subscribers will be continually drawn to recklessly traded systems with high returns. Hawk-FX is an example. Don’t be surprised when you start seeing 5 star reviews again for Hawk-FX followed by 1 star reviews a couple of months later 
Many providers don’t trade in real life with anywhere near the leverage that they do with their C2 offered systems. And the reason is trading on paper is quite different than real life. With paper trading you can blow out an account and walk away unscathed. In real life, losses are potentially unlimited. The trader may not only lose his trading capital but his house and the shirt on his back.
So you can call overleveraged abuse. I call it unrealistic. Or maybe it is too realistic for the poor subscriber who will one day suffer through a (what is it called?) sigma 6 event. And money management stops may not protect the trader.
> I agree somewhat with the SPM argument Sam. However you can still buy SPM. It is just that the premium payed would be substantial. This translates into massive slippage.
That’s the point. The equity curve is not real. It assumes he bought 8,000+ SPM @ the opening tick price. I.e. the RF.
>So you can call overleveraged abuse. I call it unrealistic.
Once again buying 1,10, 100, even 1000 eminis when the opening range volume is 110,000 is more realistic than 10,000 SPM when the opening range volume is 100. Let’s not get sidetracked with trying to
feed the poor and protect every dumbass newbie that rolls into town
> Or maybe it is too realistic for the poor subscriber who will one day suffer through a (what is it called?) sigma 6 event. And money management stops may not protect the trader.
… sorry, this is another topic. I don’t disagree, but it is not what the RF is, or should be about.
