The financial crisis has become a good test for the correlation and cointegration of pairs. For example, pairs like V*200-MA*100, AMLP*700-XOM*100, retained their character behavior and did not fall apart.
But there are many pairs whose character has changed a lot. Volatility increased and delta chart changed. For example, in pair XEL*300-AEP*200 formally, the convergence has been preserved, but you need to be very confident in this pair and have money to take such divergence. You need enough money to withstand losses.
Unfortunately, many pairs were broken. For example, XLK*100-QUAL*130, which for the last few years kept the delta chart horizontal, drew this chart:
If you look at the two charts, XLK and QUAL separately, you will see that in the fall QUAL crossed XLK from top to bottom:
Therefore, the delta chart has a trend. This is one of the reasons why I do not trade all delta deviations. First of all, I am interested in the moments when the two graphs diverged, and I take them for convergence. If you take the delta graph in early February and see what happened at that moment with two XLK and QUAL graphs, you will see that this is the point where the two graphs converged.
We would have to take this pair on divergence. I do not like such points, they have an increased degree of uncertainty. Instead of starting to diverge, QUAL, on the contrary, struck XLK from top to bottom. Although I think that even such a discrepancy, as it is now, will converge, but I don’t know how to take such losses, I never had to do this in pairs.
In general, it will be interesting to watch the pairs that showed sharp movements on the delta chart. It will be useful to return, for example, to the pairs IWB*100-RSP*200, VTI*200-QUAL*300, DRH*200-HST*100 in a couple of months and see what happens to the graphs.
In a crisis, the selection of pairs should be even more stringent. In pairs that have maintained their stability when volatility increased, expect deviations to further levels. Do not rush to enter highs or lows, wait until a U-turn is formed. Look at the graph of the delta, how reversals were formed earlier, look at the moments when the volatility jumped on the history.
Use technical analysis on the delta graph. Levels, trend channels on this type of chart also work. In pairs you can also meet long-term trends. For example, take NIE*700-VOT*100. On it, you can trade all bursts upwards in the expectation that they will go back down:
You can try to trade “tight”, horizontal charts, something similar to AMJ*100-AMU*160, ITOT*300-IWV*115.
But in practice there are many difficulties. The pair must be balanced very precisely in volume, which requires the broker to support the trading of fractional lots. Deviations can be quick, for example, at the opening, and it is very difficult to take them in manual mode. Or such a horizontal chart can be drawn because it is trading in the spread between bid and ask in two instruments. Perhaps there is something you can earn, but obviously not with your hands, here you need a robot.
There are pairs in which you can catch the closure of emerging geps. For instance, FUTY*340-VPU*100, IWB*100-SPTM*450, IVE*100-SPYV*400.
Even in such a difficult situation as it is now, you can find opportunities for deals. There were pairs who, although they had increased their volatility, remained predictable. For example, look at a pair of BRX*100-KIM*100. In the screenshot, I marked the working deviations with green lines:
Yes, good pairs are few now, but you can find. Observe, for example, the pairs DGX*200-LH*100, XLI*200-VIS*100, ZG*100-Z*100, PAGP*100-PAA*100, XLU*200-ATO*150. Perhaps, you will be able to earn something on them.