While I am certainly working on improving my C2 score, I am also trying to best understand it too. My score jumped from 60 to approx 160 in a day and then back down to 60 today.
I found the Topaz system where the C2 score is 1000 still after the system has LOST 50%.
Ques 1: what effects the variability of this C2 Score?
Hi Eli,
if you search through the C2 forums you’ll find many posts about the C2 score.
Some simple facts:
It is a score about a vendor, not individual systems.
There is a large influnece from the number of systems a vendor has published and the age of these systems.
thank you for the comment.
seems "age" contributes significantly as we live in a world of data and stats. nevertheless, it appears, God willing and given no major deviations from my methods, I can only go up from here then.
Off the topic, if the market continues to decline, you still going to be long only? what is the aversion to going short or using cash. From personal experience, one cannot let 50% losses like that get away from you. The decline in performance is feedback that your perspective and methods to the market are currently not appropriate. A trader learns a lot from getting broken, however, I would not have it any other way, from pain comes gain. Just my ice berg observation.
Good observation. Performance ought to be the king, just as cash is king. I guess C2 score is in continuous improvement but it may not look credible enough (along with the C2 site), when it is not highly correlated to the latest performance as some system vendors with high C2 scores have systems in current stinky trading lists on the home page. Just pointing out a fact.
There is a lengthy discussion of C2 Score in this thread:
http://www.collective2.com/forum?call-for-c2-score-algorithm
I do agree that such a score should highly correlate with performance.
But what does this mean however?
In fact it depends heavily on the timeframe used to measure performance.
With a short timeframe of several weeks you’ll find systems with spectacular results which quite often lose all their profits after some more weeks.
With a longer timeframe of 1-2 years you find mature, more reliable systems. Profits are less spectacular, such a list does not change very fast. In the long term it will be more useful (if your goal is overall profit instead of excitement) to concentrate on these “slow but steady” systems however.
A site like collective faces a strong conflict of interests here: To attract new customers all results must look as spectacular as possible and all these newbies will most certainly lose some money with shooting star systems.
If C2 advocats “old-faithful” systems (like it did until a few months before) then the Top-10 list (“Most looked at” - with highest C2 score) does not move fast enough, the site tends to look boring.
… we live in strange times. Things must look exciting, no matter how valuable they are …
>I do agree that such a score should highly correlate with performance.
Thanks.
>With a short timeframe of several weeks you’ll find systems with spectacular results which quite often lose all their profits after some more weeks.
Disagree because you are generalizing.
>With a longer timeframe of 1-2 years you find mature, more reliable systems.
Disagree, because you are again, generalizing.
>A site like collective faces a strong conflict of interests here: To attract new customers all results must look as spectacular as possible and all these newbies will most certainly lose some money with shooting star systems.
Disagree. There are no conflicts of interest between two rational parties willing to work towards a common goal.
>If C2 advocats “old-faithful” systems (like it did until a few months before) then the Top-10 list (“Most looked at” - with highest C2 score) does not move fast enough, the site tends to look boring.
Disagree. Performance should be the only criteria in including systems in the Top-10 list.
>we live in strange times. Things must look exciting, no matter how valuable they are
Sure it is strange, because according to your statements in your posts it is apparent to me that performance means very little to you despite your initial agreement that it does.
The most important "number" for gauging a trading system is
statistical significant risk reward ratio
This means:
* Measure profits
* measure risk
* do both over a long enough time interval.
My only point is: There is much debate about what "long enough" means.
For me it is "years"
For most young excited vendors and subscribers it is "weeks"
There is a huge difference between "weeks" and "years"…
Sharpe ratio is 1.035 for your system Topaz. I’m sure that is what you meant.
Does C2 score takes into account Sharpe Ratio at all? I dont get the impression it does. Maybe the C2 score algorithm needs to be modified to give more importance to this number as even the vendor with the highest C2 score agrees with it. Well, anyway, it does not matter to me. Thanks.
better yet, the C2 algo should consider heaviest weight with the SORTINO.
There are two environments to consider.
1. Mental environment: where beliefs, associations, and memories are stored. This environment is not as subject to time, as the past does not change.
2. The Physical environment: where attractions, curiosities, and the NOW reside. This environment is always changing and never constant.
Whenever the both of these environments are in disharmony, in conflict, and or OUT of SYNCH.
ILLUSION manifests.
This might explain why C2 score may not reflect truth because the C2 Score is too heavily concentrated on calculations, data, algo, etc… in the past where there is little change.
Sortino measures downside risk and may be a helpful stat to move scoring away from stats synonomous with constant less changing environments.
Skill should also be measured.
Any negative large performance drawdown should signigicantly change the C2 score.
in my opinion, Topaz should be 500 or less not 1000 losing 50%. This means his method may have stayed the same, but did not adapt to the market feedback. The market is always right and no one should have 1000, only the market should.
It would force him to reconsider mental flexibility or the spending more than 20 min a day on system inputs.
Any measurement that synchs the physical environment with the mental environment will represent the best perfection.
1000 trades done over a day, week, month, year represent skill in one form or another regardless of duration. Too much weight is applied to duration of time, which I propose is not the best approach.
playing the “time” card only serves to increase the C2 subscription revenue and SLOW PLAY subs out of their money. TOPAZ may be a black swan and may not even know it. His score gives a false impression he will do well, when in fact he may be riskier than most, especially because subs will be less willing to quickly adapt and get out losing more money than other systems.
Another approach is needed, until then…
the C2 score should be removed forcing subscribers to make their own decisions and not consider a shortcut stat that is an illusion and only risking Matt’s entire platform which he rightfully is entitled to protect.
Interesting that you considered Sortino as the measure of risk. But what about consistency, only Sharpe measures that. Why not a combination of both of these measures (Sharpe + Sortino) which is the measure I prefer.
Good explanation of what represents skill. that it involves number of trades, not time. Time card is the favourite card used by mediocre long time C2 system vendors as that is the only thing (in their own mind) that can undermine an up and coming system vendor.
In the context of the high C2 score, Topaz does look like a Black swan which broke the C2 score of the vendor. It is time that C2 score is removed and let the subscribers decide for themselves which system vendor to follow.
Black Swans highlight the dangers of using a single universal measure to rate system vendors.
sharpe ratio is a measure of return relative total risk, and return consistency is something i never considered regarding the utility of this ratio. furthermore, treynor is return relative only systematic risk. both measures are measures of return relative risk.
Sortino is similar to sharpe yet more of a downside risk measure, yet focused on RISK ADJUSTED return.
all these are return relative risk measures and do not assess skill. therefore none of these measures should be used to measure skill. these ratios are more of a risk measure comparison tool.
Jenson’s alpha measures added value and is the more applicable measurement together with other measures to consider when analyzing skill.
Return attribution relative proper benchmarks are also better approaches in measuring manager skill.
Consistency (or Integrity) of performance reflects skill in managing a portfolio assuming that one is not compromising on productivity or performance.
Consistently producing .000001% above risk-free rate of return is not skill, however high the Sharpe Ratio is (it could be near infinity.)
