Hello folks. I will soon publish several c2 systems. I am having fun deciding which other c2 systems I would like to invest in besides my own. I did my own month-by-month tally of their results based on their c2 monthly gains. Then using the c2 chart zoom, I double-checked my cumulative total by looking at the end results shown on c2 graphs. My figures roughly agree with c2 figures “if” I deduct -1% every month for software and subscription fees. This would probably be correct for my starting balance because I was picking some expensive systems… However… probably this has been answered somewhere already but for me this raises some interesting questions…
1. Am I correct that the c2 linear graphs automatically deduct for monthly overhead, which is not deducted from the c2 month-by-month charts?
2. If so, then the fun question is: does this deduction change according to:
– subscription fee / current balance at that time on the chart…?
– subscription fee / initial balance…?
– does not change with subscription fees, assumes a set figure of about -1% monthly overhead for every system no matter what is the subscription fee and no matter what is the current balance or initial balance…?
At any rate–if my crystal ball had told me to pick those systems several years ago and stick with them–my initial balance soon would have doubled. Thus reducing monthly subscription fee overhead from 1.0% down to 0.5%. And after that getting ever smaller. (Except during the few months when they crashed down temporarily.)
And yet, it seems that I must account an overhead of 1% every month, or else I would get much higher cumulative results than what is shown on the c2 graphs. So it seems to me likely that no matter what method the c2 graphs are using to deduct overhead, the c2 graphs are seriously under-estimating the cumulative totals of successful systems…? (Unless I am overlooking something here.)
Don’t get me wrong. If the c2 graphs are seriously under-estimating cumulative “hypothetical” performance, I agree with this policy. I certainly would not want to base my own decisions on sky-high past results. I would not even want to see the mathematically so-called precise estimates. But, it is still fun to dream vaguely, that just maybe the sky-high past results of crystal ball picks could have been even more sky-high…? (0.5% monthly compounded over several years, or even one year, makes a big difference.)
I would just like to know for the fun of it, is my finding somewhat correct… that the c2 graphs routinely underestimate the past “hypothetical” performance of successful systems…? Thank you anyone.
"I would just like to know for the fun of it, is my finding somewhat correct… that the c2 graphs routinely underestimate the past "hypothetical" performance of successful systems…?"
Only to the extent that they make assumptions about commissions. The numbers change a bit depending on which commission plan you choose but all the plans are higher than what most of us pay. Both the monthly returns and the graphs include commissions and subscription fees. The individual trades include commissions but no fees. The calculated stats – annual return, Sharpe, etc. – include neither so they are way overoptimistic.
Hi Krystof,
Remember that the table of monthly gains is geometric; i.e. percentage gain on current system equity, not the starting equity. So you can’t just add them up to get the total gain.
For example, if a system gains 4% per month like clockwork then you’d see 4%, 4%, 4%…
This would correspond to system equity starting at $10,000 -> 10,400 -> 10,816 -> 11,249… etc. It gains 4% each month which brings it to 60.1% gain after 12 months, not 48%.
Another point you mention is the adjustment for subscription fees, and yes, the effect of subs fee is not adjusted for current balance. That it why the green line on the chart actually makes it more difficult to compare apples to apples, not easier!
Trading costs scale with a trader’s account size so adjusting for them does help comparison. But subs fee stays the same so it skews the chart.
Imagine two people trading a system which charges $100 per month:
Trader A has a $10,000 account and trader B has a $100,000 account. So trader A pays 1% per month in subscription fee while trader B pays 0.1%, a big difference.
If the system shows a balance of $10,000 then the green line is correct for trader A but underestimates the real performance that trader B gets. If the same system showed a balance of $100,000 instead then it would be correct for trader B but trader A would do worse than the green suggests.
Thank you Dennis and Dean. After reading your comments, it seemed I was basically doing the tallies correctly… but… then I realized I was “assuming” that the average of two systems would perform nearly the same as the sum of two system. Not so. I re-did the calculations separately for each system, and then they agreed with the c2 graphs.
Your comments were very helpful. I’d like to summarize here for future reference…
1. Dennis: “”"“Both the monthly returns and the graphs include commissions and subscription fees.”""“
Then my tally based on the charts should agree with the graphs. But they did not. Reason: I was not calculating each system separately.
Dean: “””“Remember that the table of monthly gains is geometric; i.e. percentage gain on current system equity, not the starting equity. So you can’t just add them up to get the total gain.”""“
Yes, pretty sure I am doing this correctly, especially now that I’ve matched the graph results. Each month is calculating the gain x previous month and adding that to the previous month’s result.
Line 4: blank month 1 gain% month 2 gain% month 3 gain% etc.
Line 5: 100% =SUM(A5B4)+A5 =SUM(B5C4)+B5 =SUM(C5*D4)+C5 fill right
Cumulative:
(the final result on line 5 minus 100%) / 100%.
2. “””“The individual trades include commissions but no fees.”""“
I did not use the individual trade figures, only the monthly figures. But this is good to know.
3. “””“The calculated stats – annual return, Sharpe, etc. – include neither so they are way overoptimistic.”"""
That’s interesting. Good to know. So if my own system’s backtested CAGR including overhead seems somewhat less than another system’s c2 CAGR, it might still be in the running.
