That's one of the ways they're using to extend the life of an O2 sensor.
They don't even install the sensor on the exhaust pipe, but rather a remote location with an air pump that sucks some of the exhaust out to sample.
Once you pull that exhaust sample out, you can run it through an air/oil separator and over a heated O2 sensor.
It's a continuous flow system, meaning the pump is flowing exhaust over the O2 sensor at all times and nearly as fast as if it were inside the pipe.
Another benefit is that you can hook the line up to the stinger if you wanted to avoid that bit of fresh charge that gets pushed into the exhaust, preventing erroneous O2 readings.

Wonder how they keep the O2 Sensor from getting oiled up?

Just want to say, lot of stupid in here.

Just want to say, you don't need to dyno tune it for every mod.
You don't have to retune it for every mod.
It will make more usable power.
O2 sensors can work.

My experience with running two strokes in closed loop and using O2 sensors has not be very rewarding.   The life span of the O2 sensos are very short when you use leaded racing fuel.  Even when using unleaded racing fuel, keeping the O2 sensor hot enough to keep the O2 sensor from fouling is just one of the many problems I encountered running a two stroke in closed loop mode.  Large single cylinder engines seem to give more O2 sensor problems because most riders do not run the engines at wide open throttle as much as the smaller engines.  Keeping the exhaust system real hot and the average exhaust temperature inside the pipe high is helpful in extending the life of the sensor.  Two stroke racing engines need a lot of oil in the fuel to meed the lubrication demands and this also causes premature failure of the sensors.  The LSU heated Wide Band sensors that Bosch makes seem to last the longest but the life span is still not acceptable by my standards. 

I started using EFI on two strokes in 1993.  The controllers we were using were supplied by EFI Technologies.  At the time they were supplying the engine management systems for the Cosworth and Ilmor Chevy Indy car engines.  We had the same systems and software as the Indy Cars were using but we were only using 2 cylinders of its full capability.  The system we were using could control up to 12 cylinders with 2 injectors per cylinder and each injector had its own fuel map.  These systems were sequential, allowing the tuner to start or stop the injector at any desired crankshaft position. 

The above system also had a lot of options as far as the different types of ignition systems that it could control.  We used the system to control our ignition and had a different spark map for each cylinder.

We were most successful running the Alpha-N type system which used RPM and throttle position as the primary variables for fuel and spark mapping.  The optimum air/ fuel curves for high performance two strokes are not nice and flat like the majority of the four strokes.  Four strokes are a natural candidate for using an O2 sensor to control the engine's fuel needs but I have never been happy with the quality of the fuel control that an O2 sensor provides on a high performance two stroke.

Efi 2 stroke may work at a constant rpm, like aircraft. Just not so well on a Zilla

Why is everyone concern about running fi in closed loop mode? You can tune and use fi in open loop. Many fi atv and dirt bikes don't use an o2 and Many of the sport bike aftermarket fi tuners eliminate the o2 and you just tune the engine on a dyno as would with a carb.

so if one could maintain a consistent temperature across the exhaust system, would that minimize the effects of air flow fluctuation due to the changes in exhaust pipe temperature?
i agree that if you can make more power with a carb system over an fi and keep the reduce complexity of the fuel system that fi would not be so impressive.

side note or way out in "left field"
ive contemplated if you could somehow combine the reed and throttle system in to a single operating unit and use a fuel injector or multiple injectors to directly inject pre mix into the crankcase, similar to the idea of direct injection on automobiles. would this allow a designer more freedom to play with port design rather than to fall into the constraints of using the same combination of carb to reeds to crankcase to cylinder...
 
feel free to elaborate motogeek

Yamahas H.P.D.I. (High Pressure Direct Injection) & Evinrudes E-Tec both utilize the injector in the combustion chamber of the cylinder head. Lubrication is via separate pressure feed oil lines to vital areas. According to an engineer I spoke with, the advantage of this on a 2 stroke is the intake tract can be optimized for dry flow as opposed to a fuel laden wet flow design.The problem is physical size of the injector pumps, they are big & bulky.  If technology were to shrink these, is there any future with this in the powersports industry?

Yamahas H.P.D.I. (High Pressure Direct Injection) & Evinrudes E-Tec both utilize the injector in the combustion chamber of the cylinder head. Lubrication is via separate pressure feed oil lines to vital areas. According to an engineer I spoke with, the advantage of this on a 2 stroke is the intake tract can be optimized for dry flow as opposed to a fuel laden wet flow design.The problem is physical size of the injector pumps, they are big & bulky.  If technology were to shrink these, is there any future with this in the powersports industry?

There is a future if they can find a way to make the cost of EFI and or direct injected systems about the same as conventional carburation and CDI ignitions.  The reason for the direct injection in two strokes was to comply with new emission standards and decrease fuel consumption, not to increase the power.  There is not any power advantage to direct injection.  The boat, jet ski, and snowmobile engines that use direct injection make a lot of power because they are large displacement engines that are not in a real high state of tune. 

The direct injected engines are expensive to manufacture, repair and expensive equipment is needed to diagnose and repair the electrical systems on these engines.  I do not think that the power sports industry is going to move in this direction unless high emission standards are placed on all of the off road vehicles.

so if one could maintain a consistent temperature across the exhaust system, would that minimize the effects of air flow fluctuation due to the changes in exhaust pipe temperature?
i agree that if you can make more power with a carb system over an fi and keep the reduce complexity of the fuel system that fi would not be so impressive.

side note or way out in "left field"
ive contemplated if you could somehow combine the reed and throttle system in to a single operating unit and use a fuel injector or multiple injectors to directly inject pre mix into the crankcase, similar to the idea of direct injection on automobiles. would this allow a designer more freedom to play with port design rather than to fall into the constraints of using the same combination of carb to reeds to crankcase to cylinder...
 
feel free to elaborate motorgeek

Keeping the exhaust temperature inside the pipe from one end of the expansion chamber to the other constant and repeatable throughout the whole length, would help the repeat ability of the air consumption and fuel demands.

I do not think that there would be much of an advantage of injecting fuel into the crank case over a well tuned carb. 

Suzuki has had doing R&D on some variable lift-variable duration cam engines for about 5 to 7 years but has not been released for production.  These engines do not have a throttle on the intake.  All of the air flow into the engine is controlled by the variable lift /duration cams.  At an idle there is very little lift on the valves and very little duration.  At high RPM and max power mode the valve lift and duration of valve opening is more radical than a conventional cam setup.
   

I am with PCS on this, and I've been saying it for a long time as well. Someone just needs to nut up and actually give it a try. As soon as one person does it and figures it all out, or most of it, then a second person will. Then a third, fourth, so on and so forth. After the first few people, majority of the kinks will be worked out.

If I weren't broke I would give it a try on my 250. Just because it's different, mainly. But I'm broke, and know the building an EFI set up will make me even more broke.

how could you be broke when you have an army of atvs in your back yard