The oil headline scanning algos have completely lost it. While in the past it used to be just oil "production freeze" headlines that sent crude surging, now any old headline will do. Case in point, the price action since noon today.
Just around 12:37pm Eastern, Bloomberg blasted the following machine-readable headlines:
- IRAQ DEP. MIN. SEES CRUDE PRICE RISING ON INCREASE IN DEMAND
- IRAQ SEES RENEWED TALKS ABOUT OUTPUT FREEZE AT OPEC MEETING
Which as we promptly noted, pushed oil instantly from $46 some 20 cents higher.
Here we go again. Oil bounce on every algo's favorite headline: IRAQ SEES RENEWED TALKS ABOUT OUTPUT FREEZE AT OPEC MEETING
— zerohedge (@zerohedge) May 12, 2016
This was to be expected.
What however caught us and oil traders completely off guard is what happened precisely one hour later, at 1:36pm, when the very same Bloomberg, quoting the very same Iraq, blasted a headline which not only refuted what Iraq had "said" before, but suggested even more oil was about to flood the market. Some 1.2 million barrels in fact.
- IRAQ CRUDE OUTPUT IS 3.8M B/D, SAME AS IN APRIL: DEP. OIL MIN.
- IRAQ PLAN TO BOOST CAPACITY TO 5M B/D DEPENDS ON OIL PRICE
In a world of "logic", Gallium Arsenide or otherwise, one would imagine that the second headline would undo the impact of the first one. Nope.
What happened instead is that as the first headline sent oil - expectedly - higher, so did the second one!
In other words, we just had a "market test" according to which algos buy oil on both "production freeze" and "production boost" headlines, confirming that algos not only don't even scan the headlines, they just look for only one key words, in this case "Iraq", which then proceeds to unleash yet another momentum-igniting buying spree.
Trade accordingly.