quandl : seems interesting and doable. Now that we have custom series maybe we can add functions that suck data in. Placed a call with them. For example
durandus : we’ve gotten away from writing our own language by creating more powerful functions like FRank(). But I don’t think we need a full blown language. We would need a language to maybe do some pre-sim setup, like creating series. Some setup before relancing. And maybe some programming to do position sizing. It would help if you give us some examples on how you’d use a programming language
macroaxis : (was afraid to move the mouse with all those things moving…) P123 can generate a correlation matrix with the screener. Not as pretty , a bit kludgy, but more flexible. You can specify the period and the range. Please see this screen: http://www.portfolio123.com/app/screen/summary/137783?mt=1&st=0
Friends, I told you before, P123 is the best www.bloodhoundsystem.com seems to be the biggest competitor, still not good enough, since earnings estimations make 50% of
my Performance in most of my modells and they do not have it!
I’m curious: How about price data for a lot of commodities – presented on a continuous basis (i.e. without having to deal with the gyrations that occur as contracts roll over)? Might folks here see that as a potentially useful thing for timing/economic models?
I’m not sure about continuous commodity data but one thing I would like to see is Commitment of Trader Reports for futures and futures + options. I haven’t looked at this stuff for years but commercial traders for S&P 500 used to be a very good prediction for the stock market. Quandl doesn’t seem to have it but https://www.pinnacledata.com/ has both continuous commodity contracts and COT Reports.
P123’s core business is very solid and likely to stay that way.
The big business issues for P123 growth are the following:
a) they are a ‘generalist’ platform (which I love) offering a wide and ever growing set of tools for a relatively small, narrow niche of hardcore users (smaller retail investors with quant / engineering / science / finance backgrounds and time on their hands and passion for markets and investing and less then huge port’s typically).
b) Companies that P123 is competing with for ‘big growth markets’ will be much more narrow and ‘niche focused’ and the ones to worry about will be thosed focused on these ‘thin slices’ that can scale most rapidly have much larger core niche audiences and narrow focus, and get huge investment backing, i.e. Wealthfront.
It’s also key to recognize that companies that raise lots of money and are in the same general space are huge threats. If a company like Quantopian (and I don’t know them at all) raises $30 million, they are likely looking to invest that in under 3 years to get to huge user growth and revenue milestones.
That means $10 MM a year (or more) on their goals. So… what could they do with that:
a) $6MM per year on user base growth
b) $2MM on developing core, proprietary data streams that no one else has and are PIT and super high quality - data licenses and talent (IT and finance) to manage
c) $1MM on core user interface simplification
d) $1MM on attracting talent to go out and land institutional investors for a hedge fund
That’s every year. And that’s in their early years. To not view that type of investment as a huge threat, is a mistake.
I would love to see P123 add new, proprietary, hi-quality data sources - even if they are ‘add on’ subscriptions. New / better data is one very unique source of potential return drivers. Making the ‘autotrading’ better, safer - with more order types and customization, so I can autotrade lower liquidity systems (but not letting R2G sub’s do the same - or I am competing with them in these spaces).
However, growing into a bigger business makes P123 less valuable to me as a private investor using this platform.
But P123’s existing business will likely be fine, because it’s offering a lot to a very narrow audience and is the first in the market. That’s a huge advantage. It’s only if you want to grow a lot that you need to think about other ideas (narrowing focus on new launches - and uniquely branding them).
I don’t know www.bloodhoundsystem.com but they seem to have data that extends much further back in time than 1999. Does anybody think it would be worthwhile for P123 to extend the data back in time to make backtesting of strategies more robust?
Seven,
Backtesting earlier than 1999 might make a model more robust but not necessarily. After all it may be LESS robust. There is simply no way to know.
When you looked at COT, did you actually try modeling with it? If so, how did it fare? I worked with COT a bit in the mod-2000s, mainly for oil and copper, but haven’t drawn much conclusion one way or the other about its efficacy. I never tried it with equity index futures.
In terms of specific data:
a) Options related data - this has been posted on in the past (put/call ratios, changes in put call volumes vs. trailing averages, actual vol of the stock vs. vol predicted by options prices, etc), and is likely to contain a lot of ‘smart money’ information inside.
b) Better earnings data. (crowdsourced estimates and the like are ‘all the rage.’)
On the futures side, I do know one CTA trading macromodels using oil futures as in input. I haven’t worked with this, but know someone who has - and he uses it as a core input on market direction forecasts - and has found it robust across many economies.
EquitiesLab is a newer P123 competitor. I haven’t used it, but from talking to the founder it seems like they mainly focus on fundamentals oriented backtesting. The downside seems to be that they use an inferior database (it’s not Compustat, I think it’s Thompson Reuter). The upside is that it’s cheaper than P123.
Steve Briese at www.InsiderCapital.com has several newsletters centered on interpreting the COT reports. I think he’s been very good with things like gold and oil and not so good with the S&P500. Perhaps the Fed intervention has distorted the markets too much. There appears to be some valuable information there (in the COT) but it takes some experienced to dig it out.
[size=1]“Zack research wizard - 1990’s windows front end, “lagged data”. Could be a serious competitor if they improve it. Be ready for lots of emails when you signup.” [/size]
I used Zacks Research Wizard for a bit around 2007. They had easy and a decent rule building functionality. But at the time, I believe their data had surviorship bias and actual stock picks didn’t match to the backtested results 3-4 months later. Of course, that was a while back and things maybe different now.
[size=1]“Quantopian - geared for intra-day, super hard to use. Dabbling in fundamentals but what’s the point? You are either intra-day or you use fundamentals. We’ll keep an eye on them.”[/size]
I realize that p123 is geared more towards longer time frame, but I am curious if anybody is using any of P123 features for intraday trading or intraday entries? I started using p123 to try to find a decent entry intraday and hold for a few days (both long and short). So far only experimental trading for me and some moderate upside, though it is still too early to tell.
We discussed some stuff on the phone a while ago. My strategic approach for you would be:
Maintain and further develop the platform so that designers and users can REALLY produce above market returns. This includes relevant and MUST-HAVE features like position sizing, simulating order types (such as LMT orders), broadening data to EU (and later Asia), providing daily ranking etc.). All these should increase returns and capacity in the first place. It will be hard for other platforms to compete. Actually Wealthfront used to be KaChing (failed approach), and also CoVestor really never succeeded with user-provided models, and also Marketocracy struggled with performance. Reason was that the models/strategies they promoted were often crap. So performance and capacity will be key.
Develop and promote a “marketplace” like Motif, Covestor or Wikifolio. Offer trade following, auto-trading, etc. but also extend the platform. Add new brokers, heavily raise VC money and market this marketplace product. Only select and present a small sub-set of great strategies (from P123’s R2G) for different market sub-sets (sectors, market caps, trading, investing, market neutral, long/short/neutral market vola payoff profiles, ETF allocation systems, …) otherwise investors are going to be lost. I would probably do this on a different website and platform to make it easier for marketplace subscribers. Basically it is a totally different user base than the P123 designers. Just leave P123 R2G as is and crowd-source investment talent from P123. I have many ideas, also funding wise as discussed previously, so get in touch with me directly if you like.
Again, as mentioned many times. P123 is great now and in the lead, but Motif and Quantopian have so much VC money, and their development efforts and capital grows hyper-exponential. That means they may develop certain features today already 4x faster than P123 and next year may-be even 16x faster due to heavy investment in R&D.
But again, you have to have a solid 1) otherwise the whole marketplace will fail if models are weak. So don’t waste your time with new series, etc. Focus on the big picture stuff NOW like introducing new EU data, position sizing formulas, increasing liquidity and capacity (LMT orders, daily ranking/rebalancing, etc.)
While I am less inclined to offer specific recommendations, P123 is in the lead and has the opportunity to stay there.
I really just did not understand Marc on my previous posts. Skewness is closely related to the discounted future expected cash flow? I’m trying out weird ideas that might just be lucky?
I really don’t need to understand everything. Marc has helped create something beautiful. My fear P123 will break something has subsided a bit.
I mostly agree with whotookmynickname. I think there are no leader on this marketplace so suggest P123 to go it’s own way. I was dreaming about P123 strategy a while ago.
And I have a feeling P123 is developing “a bit slow” :(. There is no VC for more resources, that is OK. Other way is to optimize own workload. There are a lot of request and I do believe most of them needed. Maybe it would be useful to P123 with Community to list all of the requests, range them (I know there is voting but I see not too much voters) and targeting auditory (balance between P123, designers and investors), combine requests falling into the same developing strategy etc.
Probably one or another way this is done inside P123. Could it be more useful not only to collect requests and report results to community publicly, but actually to some extent involve community into decision making process. Actually this is what is this thread about, just a little bit more structured and formalized workflow?
Many of your ideas are sound business strategy. However…
“heavily raise VC money and market this marketplace product.”
This will be the best thing for putting money in the pockets of the P123 owners, but the worst for current users and ordinary traders. As more sophisticated VC’s and more money flow in, the average P123 user will experience much lower returns and be crowded out of trades. Except at the very lowest ends of liquidity, this is the history of all these platforms (most recently Peer-to-peer, like Lending club).
