Everyone has their own strategy to come up with Key digits, pairs, sums, positions, singles, doubles, roots, etc......
I say do what works for you after you've thoroughly backtested it. All we are doing is essentially going the scientific method and basically making observations, testing our hypothesis, experimenting and posting our results. I stress backtesting because that's the main way you can be the most confident in your strategy before putting money into it.
There are many ways to come up with Key digits. Some people use the top 2 hot digits within the previous 10 - 50 drawings.
I focus more on predicting things that are extremely rare/cold conditions. One way that I come up with key digits is to find out which state or combination of states have two digits that have "not hit" within the previous 11 drawings. Why 11 drawings? Because I base my predictions on the records of all classifications like sums, roots, positions, pairs, digits, etc.... Each classification has a record. According to my records over the years two digits have always hit within 18 drawings. Since I usually make less than a 7 drawing forecast I wait until I see a state that has two digits that have not hit within the previous 11 drawings. 18 drawings - 11 drawings = 7 drawings. It's best to keep things as simple as possible.....
I say do what is most accurate.
I only make multi-state forecast in order to win within one day since most states only have two drawings a day and it may take longer than two drawings to win. Instead of waiting all week for one forecast to win I include other states. My key digits for the 1/13 NC day drawings are a part of a list of states in a multi-state forecast that I made. My key digits and pairs are for the following states...
Key digits: 6, 8, 9
Key pairs: 68, 69, 89
States, Dates & Drawings included in this forecast
Drawing #1: 1/13 Virginia day drawing:
Drawing #2: 1/13 North Carolina day drawing:
Drawing #3: 1/13 Maryland evening drawing:
Drawing #4: 1/13 Georgia night drawing:
Drawing #5: 1/14 Texas morning drawing:
Drawing #6: 1/14 New York day drawing:
Drawing #7: 1/14 Virginia day drawing:
I wouldn't invest in my own or anyone else's predictions unless I see a track record of success. Most people do not make predictions that are specific enough. You don't know whether to play in Exact order or Any order or when to expect to hit so whatever you do, be specific. I've found that it's best to make multi-drawing forecast in order to win within a range of drawings instead of just trying to only predict one drawing at time. Try starting out practicing to win within the next 10 - 15 drawing forecast and then make shorter and shorter forecast as you improve.
No matter what make sure you backtest, backTEST, BACKTEST your strategy and give it a grade/accuracy rating before putting money into it. My minimum is a 100 drawing backtest to verify where my accuracy is before investing especially you rookies. Start out working you way up to winning 80 out of 100 times until you reach winning 97 out of 100 times.
It's frustrating to any predictor to give any effort to help someone learn when they are lazy about doing the work involved to increase their accuracy. You can't complain about sucking as a predictor if you don't put in the work. Over time (maybe years depending on how talented you are) you're intuition and accuracy will increase. It takes being obsessed to master being consistently accurate throughout the whole year and not just hot for a day or a week.
You're either going to hustle to learn or beg accurate predictors for predictions like most people do.
I'd rather be blunt and tell you like it is. I've been doing this a long time and doing pretty well. Predicting is not easy and will take a long time to master your skills like most worthy endeavors but you can make a consistent income and even replace a job doing this, similar to being a stock market day trader.
One more thing, it's best to use software to help analyze conditions and backtest faster. You might want to consider being a platinum here and buying more software.
100I hunt pairs & Infinite Combos (IC's)United States
Better look from now on!
) 
