The real problem is that the available data doesn't support such claims.

You don't have independent data points. You don't have a random sample of the population of data points. There is every indication that such things are, in fact, impossible to attain since we are all working on developing strategies to beat the market.

That makes the statistics/probabilities quite a bit more difficult.

Let me be the test to verify that claim. Agree? I can start posting proposed trades for 1 year? Deal? What do i get for having over 60% success rate?

Mine does. The real problem is i cant tell you what it is and nor will anyone else. But even if you look at the strategies of the ill-fated LTCM its quite clear that for even quite large sums of money there are at identifiable times hugely profitable strategies - the problem is what to do when the money is too large and how to avoid getting carried away (especially during the times when there are no great opportunities)

I agree, I think I made this case in the strongest terms. However the strategy I would and do happily bet on makes a apriori type case which all the data supports, even if its an accident its way past the stage where its more reasonable for me to believe it is an accident.

I have a trade (no tricks honest) which is profitable, beats the market significantly 100% of the time and is never at any time losing (except in the degenerate case where the market doesn't move at all in which case there are negligable transaction costs). All the problems are stuff like counter-party risk

I've alluded to it in BFI and wish I could divulge it as I think its fascinating but havent given up on exploiting it yet and am kinda sworn to secracy.

You would get my undying love as far as internet relationships go. I will address you in all posts as "his majesty, masque de Z."

Plus your profits from making such trades. Those would be enough by themselves.

Also, I will (as the most important value add I can offer) do my best to tell you when and why your strategy is unsustainable (i.e. will fail massively) over a lifetime if I can.

In real life, I can place no value on one year's returns. There are loads of people who have made 10,000% yearly returns once or twice.

It is an evolving game, so any wonderful and amazing new insight you have will be crushed eventually by some new idea that makes your idea obsolete.
I'd certainly be willing to put a couple of hundred on your claim though. Probably not worth either of our time though.

It is the addition of claiming 20% geometric returns over a 10 year period that is not supported (outside of lottery play strategies).

And counter-party risk does count as a real life risk. I think you will find that money that you should receive spends considerably differently than money you actually receive.

I'm pretty sure you realize that return percentage on capital per trade is a function of liquidity and not only your skill. And I don't understand why disagreements on yearly returns 20%, 10% or whatnot since they are not relevant, yearly return really shows nothing if you don't know the trade count...

I'm not sure about current liquidity in masque de Z example...being closer to 10k starting capital 1000% returns are not an exception if you have a decent trade count per year for a human trader

and 1mil to 100 mil long term for an algo trader if he can run the strategy across a lot of markets and ramp up the trade count

I profoundly disagree. One thing that is certain about what I'm doing is that that is not remotely an issue. I have to accept that despite all the overwhelming evidence what I do could have some flaw (though it definitely negates your point about what the data tells us) but there is simply no way that if it has no such flaw it will struggle with 20% geometric returns for 10 years, it will barely have scratched the surface of the liquidity pool.

Of course, I was just pointing out a limitation of paper trade examples. However realistic they are they cannot account for the counter-party risk created by the trades actually being implemented.

I'm assuming he is smart enough (and not rich enough) to be making trades at more than 10% of the market daily average volume.

I'm also assuming that he is smart enough to not bother with less than $100k. Not "being an exception" is also not a goal I'd put him on. Loads of people who bet on horses make returns that are not an exception.

And the 1000% (or, 10% above market) returns are precisely what I am arguing the data doesn't support. The available data is not independent and is not a random sample of the total population, therefore any attempts to draw population conclusions are difficult. In practical terms, it is always a case of "it works until it doesn't." And when it stops working, it tends to break people.

Most importantly, the lifetime geometric annualized return is the only thing worth judging. That is a direct measure of whether you won. Being judged by any other measure is just ridiculous.

The important bit is beating the market. If the market gains 10% this year, will you gain 20%?

If not, the lazy path is clearly the correct path.

I'm not talking risk free (heh!) rate of return, I am talking beating the market by 10% consistently.

Such things do happen, but the road is littered with those who have done quite well for what we can call "a while."

There are other risks inherent in all trades. The biggest one is just being wrong about the future following the past. Any expoitable weakness is there for a reason (durr). If you don't know precisely why that weakness is there, you are at risk.

I assumed we were talking about beating the market.

Its very poor logic to deny there are winners because they're are many losers. Most of the losers never had a prayer, the ones who did stick out and there suddenly aren't so many.


Its not obvious you do need to know why in particular, if you can find an arbitrage then who cares why? well i do but its not necessary to know why. I speculate the opportunities exists because you cant comb a hairy ball. I also give thanks to the people who have persuade smart folk that arbitrage is somehow too difficult for amateurs to implement.

The good thing about the future being different is that means an arbitrage opportunity ceasing to present itself which doesn't actually cost you any significant money.

OK, we are on the same page on this discussion. Sometimes just beating the market is a good goal, and sometimes it is a horrible goal.

You have kindly reminded me of that in other conversations. I haven't figured out a work around yet that I am satisfied with. When I have, I will let you know.

Of course, but it is a bigger (and more common) mistake to take the presence of winners as indicating that they are not lucky fools.

History is littered with really smart people winning for a while. History is littered with many more smart people who would have won had they been playing last year's game.

I don't think it is immediately obvious, which is the problem and hence the argument. I think it is quite necessary to understand why an opportunity exists and why it might become a greater opportunity in the future.

(simplifying things by taking out the long position one would obviously take) RIMM has been a crap company for years. Figuring out why it was overvalued is key to figuring out when to short it. Knowing that something is mispriced is only part of the equation. To maximize gains, one must know when it will become correctly priced.

Other price pairs also exist and become mispriced for a reason.

Brent/nymex/ice crude mispricing happens.

Comex/Spot mispricing also happens for a reason.

Stuff getting away from a trend or a pairing happens. If it didn't AAPL and RIMM would be the same price.

Without knowing why, you are at great risk of that "why" becoming an exploitable small difference to becoming very important huge difference and you losing everything. That is the risk of any long-short strategy. Pricing can get quite lumpy and the ugly can get uglier (or maybe uglyer, neither looks right) overnight, stopping you out of positions that are suddenly no-brainers to own.

Combing a hairy ball probably has little or nothing to do with it. If it does, any cowlicks would not be stationary and you've got loads of combers, which would make the math a little beyond me to explore. Predicting tomorrow's cowlick would be necessary. Perhaps such things are slow moving, but that would take all the big gains out of it.

I disagree. When small things that you are taking advantage of in a big way stop working, they cause loss of significant money. It isn't like you immediately change strategy after one loss (or 10) unless you are an idiot.

Sure but we have to move forward in the conversation. We are past that common error and having filtered out the people who cannot get past it the common error is to go to far the other way.

The same problem applies to the rest of your point. You make a true case against pair trades but we need to move past that. Are there arbitrages to which your case doesn't apply? can we find them? can we exploit them? Consider an example at the other extreme to a pairs trade, broker arbitrage. Sometimes two different brokers will have different prices such that ones buy price is above another sells price by enough to make a profit from arbitraging it. Do any of your objections apply?

1) Do we need to know why its happened? No
2) Do we need to be able to predict it happening? No
3) Are we exposed to risk of the gap opening hugely. No (you could argue this but only as a nit)
4) Do we need to know when its correctly priced? Yes but this is clearly not a problem
5) Could it correct too slowly to profit from? No (you could nit again)
6) Does it cost you significantly when it stops working? No, this is the most important no imo, all that happens when it stops working is the price mismatch ceases to happen so there's nothing to trade.

The one risk you do have is counterparty type risk. This particular strategy is a bit like card counting and likely to figuratively get you a trip to the back room.

Arbitage trading is a wide spectrum of exploitability with the naive pairs trade at one end and broker arbitrage nearer the middle. If we can remove the 'backroom kicking' factor and make the broker happy with our business then we are right at the other end where we want to be.

One thing that is well known is that a comination of many pairs trades moves us massively in the right direction. It meets all but one of the above objections (objection 5 remains) including the 'happy broker' problem

maybe a good question to consider is if you made lots of pairs trades then what is the probability you would end up significantly on the wrong side of genuine changes in relative value?

Of course. The possibility that it is all just statistical accident doesn't preclude the possibility that it is not a statistical accident.

Yes. Of course (as long as we are using "arbitrage" in the loose sense). There wouldn't be anyone other than manufacturers and consumers if there were no arbitrage opportunities.

Yes. The presence of successful middle men seems to indicate that such things are discoverable.

My objections certainly don't apply in that case. For us mere mortals, such things are unimportant.

Perhaps not, but if we don't understand why a greater price discrepancy might occur and bring our little endeavor down to its little knees (gnashing of teeth situation), we will be quite sad and have to move into a shed in a seedy part of town.

We could just wing it, an approach I take in nearly all endeavors, but it would be nice to have some prediction of such things happening and (probably more importantly) the magnitude.

It really sucks when you get stopped out of the smart side of a trade because it got smarter after you made it.

Depending on the arbitrage, gaps can get big. If that isn't an issue for a particular arbitrage, it isn't an issue.

It is quite important. Opportunity costs are huge when one is aiming to beat the market.

No. The worst that can happen is that the price mismatch can get much larger than you expect it can.

I don't know what you are specifically doing, so I can't comment...

The broker merely collects rake. I doubt that they have a problem with you playing.

Objection 5 is a big one. Knowing that I could make a profit working as a laborer doesn't mean I should be one.

The other big problem is being stopped out of positions if one is correct, but the magnitude of the mismatch gets bigger than expected. That can (and has been) quite catastrophic.

It depends. The available data suggests that being in the game long is significantly better than taking a paired trading approach. Assuming that the future will at least rhyme with the past, it is difficult to see a benefit in not taking advantage of the 10% or so head start being long gives.

depending on where you are on that scale there are different precautions you need to take. And I'm pretty sure both of you have it right but are just talking from a different point on this scale

they have, if you spam with limit orders that never get hit