When does the market actually react to news releases?

It seems some days the market will coincide with a news release and move along with it for a session or even for a day following into the next sessions around the world and then other times it's basically ignored. Sometimes it's followed for a brief period and sometimes it's followed for a long time as I said. How do you know when the market will actually follow the fundamenal or numbered data?

Don't follow fundamental. Honestly, it sounds crazy and I never used to believe it and would fight it a lot but...
All you are trying to do is create a positive process and "positive coin flipping" game: IE in simple terms- win you get 1.25, lose you lose .50 and your odds are 50/50. IE you NEVER control the outcome and should not try to explain it outside of using price statistics and basic price.

These are my personal processes:

1. run models on raw price
2. Have a "signal" for whether to be short side only or long side only
3. Have a computer manage your trades (or your worst worst case risk. For ex: I look at 2k as the worst. It almost never gets hit, but that is my worst case stop there since RISK MGT foremost right?)
4. USE the news releases as an opportunity:

IE if you have your bearings and a one side "short only/long only" based on your filter, then you can usually use a news release to get near your point of entry (market influential point in your model). Do like 80/20 position size: based on your systems stats, then take the 80% off when it hits the median expected (hopefully positive. It should be if you have a good system) value then let the rest ride (with a stop at your systems "influential point"... NOTE this is NOT the filter (long/short) point. I prefer to not stop, but to "hedge" with an opposite point (so I really need a foreign broker since it is not allowed in the US).

Also note: I'm not sure the specific entrance and exit matters all that much... mostly your worst-case risk management and always having that in place 110% of the time. But that is just my opinion.

I'm also new to actually collecting results, and I'm getting results on demo account now so I'm not sure how much you want to listen to me yet. 5 years of education in finance/computers/ etc. About 2 focused on risk. And I'm a programmer (I deal with market data) by trade.

--------------------------
Also look at it like this:
*If all events in life (or your trading career let's say) are random flips anyway that you cannot control, why would you limit yourself to making decisions on flips that are by nature limited in number?

*I do use them to setup my opportunities (they push price close to my "influential market points"), but I'm talking about this: why would you want to use a system that only made a decision off of these few decision points?

1.) If you are wrong you validate slower. IE if you drawdown, you do not know if it is mostly by chance or if something is broken.
2.) If you have positive expectancy with a process, you want to hit it hard many times as possible.
3.) Compounding works better with many events. Making many (still positive but smaller expectancy) trades should always be favored up to the point where commissions get crazy. But if you have good systems, I'd almost not worry as much about the commissions as well until later.
*I usually review a trade report and my percentage of commissions per my P/L at the end of each day.

Just my .10 though.

Think about who are the players in the FX market. Most of them are not speculators, they are not reacting to every news event. The trends are formed by actions of multi-national corporations, financial institutions, central banks -- the actions they take are fairly stable over time and don't change on a day to day or week to week basis. Find the medium term news headline -- i.e. Euro QE, Japan QE, US interest rate rise -- and then trade the trends created by those using technical analysis. They are stable trends.