There is a danger with correlations. It's quite trivial to find correlations, an exercise in Excel that any intern should be able to do quite well... what happens, and it has destroyed some people, some very smart people even, is that sometimes instruments become uncorrelated! If you don't have a deep understanding of why they are correlated in the first place then you won't see it coming when they decouple...
There is real edge here. From my experience, an absolute must is a very good time & sales data source - Thompson Reuters or Bloomberg. You have to be absolutely sure that the chronological synchronization between the data sets you are testing is precise. There are at least four trading strategies that can be used to exploit correlations, depending upon the trader's IT and cost infrastructure, and more importantly the lead-lag cointegration characteristics of the instruments being traded.
It would be great to get the list of instruments that are inversely correlated and be hedged quite nicely. But people have done that, things have decoupled and switched to positive correlation and overleveraged people have bought the farm. I knew a guy that suffered a severe drawdown for a number of years.. it was unpleasant to see how much pain he was in, he was depressed from being endorphin deficient to begin with... but he suffered badly. He could not bring himself to turn on a computer and even look at email, that is how much aversion he had to trading... and that was before he talked his dad into investing heavily in K-Mart stock when it was below book... what they did in bk court devastated his dad's retirement account... Some of us cut a wider swath than others I guess Still, hedged is better than not hedged but you have to watch for decoupling, you have to be on top of that with some "decoupling stops" maybe... does IB have "decoupling stops"?? Maybe I should suggest those, LOL...
Eight, with all due respect, you are trivializing and generalizing beyond extreme in this case. If a trader has a crap strategy and poor risk controls, it wouldn't matter if it was a correlation or a technical study that spelled his demise. I've had profitable correlations last for three weeks, and I've had profitable correlations last for seven years (and counting). The fact of the matter is that a statistically significant correlation with a fundamental driver is the most consistent edge in the business, in my opinion. When a correlation falls off below 90%, I stop using it. You test them routinely, and observe and trade them side-by-side in the live market.
Expensive research because you're going to need to have access to ALL the futures exchanges...that's a lot of dollars each month and only merits such if you're doing some sort of correlation trading. However, it's already known what are the inversely correlated markets but it doesn't seen like that's your question. It seems as if the question if about how to "continually measure" changes in inversely correlated futures. Also, I'm not sure if the conversation is about futures that are inversely correlated or about correlated markets that change for example from 75% to 92% because I consider such to be two completely different types of questions/conversations. Thus, there's correlated markets like Gold GC futures and Silver SI futures that change correlations from one trading day to the next. In contrast, there's EuroFX 6E futures that's inverse of U.S. Dollar futures that change correlations from one trading day to the next. I've seen a few websites show some correlation data involving specific futures trading instruments but do not know of one that does correlations for all the known futures markets. Therefore, if you have the money to pay for all the data feeds and you're good at coding...you should be able to build your own correlation data base. Mark
Most people I've met that do correlation trading are not using indicators although I am sure someone out there is using an indicator. There's been a few correlation trading or pairs trading journals here at ET and I haven't seen any indicator discussions as a primary method for such. Mark
I think that there is some misunderstanding in this thread. I believe that Bone is talking about HFT and Lasner is talking about position trading time frames. Here is a list of inversely correlated futures/forex pairs. They have a Rank Correlation of daily return over the last 10 years of at least -90%. Hope this helps. Code: Name Name Linear CorrRank CorrelPair Stabil Danish Krones per 1 U.S. Dollar-#690(59) U.S. Dollars per 1 Euro-#591(60) -99.1_% -99.2_% 1.6 U.K. Sterling per 1 Australian $-#171(37)Australian Dollars per 1 U.K. Sterlin-99.2_% -99.0_% 1.5 Euroyen-3Mth-SGX (SIMEX) Singapore Dollars per 1 U.S. Dollar-#-97.8_% -98.9_% 1.5 Euroyen LIBOR 3 Mth-(SIMEX)-SGX Singapore Dollars per 1 U.S. Dollar-#-97.8_% -98.9_% 1.5 New Zealand Bank Bills(90 Day)-(NZFE)-SFEAustralian Dollars per 1 U.S. Dollar--99.4_% -98.9_% 1.6 Australian Bank Bills(90 Day)-(Floor+ElecAustralian Dollars per 1 U.S. Dollar--99.3_% -98.8_% 1.4 Danish Krones per 1 U.K. Sterling-#590(41U.K. Sterling per 1 Euro-#591(55) -98.5_% -98.7_% 1.8 Hong Kong Dollars per 1 Australian$-#171(Australian Dollars per 1 U.S. Dollar--99.1_% -98.7_% 1.7 Danish Krones per 1 U.S. Dollar-#690(59) Hong Kong Dollars per 1 Euro-#591(44)-98.3_% -98.7_% 1.6 Australian Govt Bond 6%(10Yr)-(Floor+ElecAustralian Dollars per 1 U.S. Dollar--99.2_% -98.6_% 1.3 New Zealand Govt Stock(3 Yr)-(NZFE)-SFE Australian Dollars per 1 U.S. Dollar--99.2_% -98.6_% 1.2 Australian Govt Bond 6%(3Yr)-SFE Australian Dollars per 1 U.S. Dollar--99.2_% -98.5_% 1.3 Australian Govt Bond 6%(3Yr)-(Floor+ElectAustralian Dollars per 1 U.S. Dollar--99.1_% -98.5_% 1.3 SIBOR-3 Mth-(SIMEX)-SGX Singapore Dollars per 1 U.S. Dollar-#-97.3_% -98.1_% 1.7 EURIBOR-3 Mth-EUREX Danish Krones per 1 U.S. Dollar-#690(-98.0_% -98.1_% 1.6 Euro German Schatz-EUREX Danish Krones per 1 U.S. Dollar-#690(-97.3_% -97.2_% 1.4 Euro German Schatz(Last Trade in Close FiDanish Krones per 1 U.S. Dollar-#690(-97.2_% -97.2_% 1.3 Eurodollar(3Mth)-SGX(SIMEX) Singapore Dollars per 1 U.S. Dollar-#-96.5_% -97.1_% 1.5 ICE US Dollar Index at US ICE Euro(Floor Trading Only)-CME -96.4_% -96.1_% 1.3 ICE US Dollar Index at US ICE Euro Currency EUR/USD at CME -96.4_% -96.1_% 1.3 Euro(Floor Trading Only)-CME U.S. Dollar Index(Floor+Electronic Co-96.4_% -96.1_% 1.3 Euro Currency EUR/USD at CME U.S. Dollar Index(Floor+Electronic Co-96.4_% -96.1_% 1.3 ICE US Dollar Index at US ICE Euro E-mini-CME -95.8_% -95.6_% 1.3 Euro E-mini-CME U.S. Dollar Index(Floor+Electronic Co-95.8_% -95.6_% 1.3 Swiss Franc-CME(Floor Trading Only) Swiss Francs per 1 U.S. Dollar-#390(5-95.8_% -94.9_% 1.3 Swiss Franc at CME Swiss Francs per 1 U.S. Dollar-#390(5-95.8_% -94.8_% 1.3 Euro/British Pound-FINEX Danish Krones per 1 U.K. Sterling-#59-95.1_% -94.8_% 1.3 South African Rand-FINEX SA Rand-CME(Floor+ Electronic Combine-95.2_% -94.6_% 1.7 South African Rand-FINEX South African Rand-CME -94.3_% -94.5_% 1.7 Euro(Floor Trading Only)-CME Danish Krones per 1 U.S. Dollar-#690(-94.8_% -94.4_% 1.3 Euro Currency EUR/USD at CME Danish Krones per 1 U.S. Dollar-#690(-94.8_% -94.4_% 1.3 Euro E-mini-CME Danish Krones per 1 U.S. Dollar-#690(-94.2_% -93.9_% 1.4 Japanese Yen-CME(Floor Trading Only) Japanese Yen per 1 U.S. Dollar-#390(4-94.2_% -93.6_% 1.5 Jap Yen Currency JPY/USD at CME Japanese Yen per 1 U.S. Dollar-#390(4-94.2_% -93.6_% 1.5 Euro/British Pound-CME Danish Krones per 1 U.K. Sterling-#59-93.7_% -93.4_% 1.4 ICE US Dollar Index at US ICE U.S. Dollars per 1 Euro-#591(60) -93.2_% -92.9_% 1.3 U.S. Dollar Index(Floor+Electronic CombinU.S. Dollars per 1 Euro-#591(60) -93.2_% -92.9_% 1.3 Euro German Bobl-EUREX Danish Krones per 1 U.S. Dollar-#690(-93.4_% -92.9_% 1.3 Euro German Bobl(Last Trade in Close FielDanish Krones per 1 U.S. Dollar-#690(-93.4_% -92.9_% 1.3 Japanese Yen E-mini-CME Japanese Yen per 1 U.S. Dollar-#390(4-92.9_% -92.5_% 1.6 ICE US Dollar Index at US ICE Hong Kong Dollars per 1 Euro-#591(44)-92.3_% -92.4_% 1.3 U.S. Dollar Index(Floor+Electronic CombinHong Kong Dollars per 1 Euro-#591(44)-92.3_% -92.4_% 1.3 Australian Dollar-FINEX Australian Dollars per 1 U.S. Dollar--91.8_% -92.2_% 1.4 ICE US Dollar Index at US ICE EURIBOR-3 Mth-EUREX -92.2_% -92.0_% 1.3 EURIBOR-3 Mth-EUREX U.S. Dollar Index(Floor+Electronic Co-92.2_% -92.0_% 1.3 Canadian Dollar-CME(Floor Trading Only) Canadian Dollars per 1 U.S. Dollar-#3-92.5_% -91.9_% 1.3 Canadian Dollar at CME Canadian Dollars per 1 U.S. Dollar-#3-92.5_% -91.9_% 1.3 ICE US Dollar Index at US ICE Euro German Schatz-EUREX -91.7_% -91.5_% 1.3 Euro German Schatz-EUREX U.S. Dollar Index(Floor+Electronic Co-91.7_% -91.5_% 1.3 ICE US Dollar Index at US ICE Euro German Schatz(Last Trade in Clos-91.6_% -91.5_% 1.3 Euro German Schatz(Last Trade in Close FiU.S. Dollar Index(Floor+Electronic Co-91.6_% -91.5_% 1.3 Australian Dollar at CME Australian Dollars per 1 U.S. Dollar--89.7_% -91.4_% 1.4 Australian Dollar-CME(Floor Trading Only)Australian Dollars per 1 U.S. Dollar--89.7_% -91.4_% 1.4 New Zealand Dollar-FINEX New Zealand Dollars per 1 U.S. Dollar-89.5_% -90.5_% 1.5 Singapore Dollars per 1 Australian$-#171(Australian Dollars per 1 U.S. Dollar--93.7_% -90.4_% 1.2
Eight, I agree completely. Have had a similiar discussion with Bone regarding old crop/ new crop grain spreads. If you do not understand what it represents and what influences it, then the market can open one morning against you and you can get run over before you can react. e.g. bean spreads after last crop report, moved about 30 cents without having a chance to react. I speak from experience. About 17 years ago I was long 1/2 the open interest of a contract with 2 weeks to go before he first delivery day. The spread was at an inverse. I ended up making money at it, did not have to stand for delivery. The only reason I put that position on was because I had what eight refers to as a "deep understanding" of the fundamentals.