There's been some recent discussion in this thread about playing biases. So I did some research with my lottery software.

Here's some stats about high/low and odd/even biases.

Since July 16th 2001, there have been 7,283 FF5 drawings which means a total of 36,415 balls (numbers) have been drawn. 18,266 of those numbers were odd numbers and 18,149 were even numbers. (Almost a 50/50 split which is considered "normal") So the odd/even bias over the life of FF5 is negligible. As far as a high/low bias is concerned, the split is 18,173 low numbers vs 18,242 high numbers. That's also negligible. Both high/low and odd/even numbers are in balance over the entire life of the game, which is to be expected. I guess that's probably why players say "Everything evens out over time with regard to lottery numbers". Yet many players do look for a bias and act them when they see one.

So that begs the question of; When does a bias become "actionable" or playable? How far skewed out of balance do the numbers have to be, one way or the other, for players to start picking the opposite side of the unbalanced bias? For example, if low numbers have been drawn a lot lately, and are skewed out of balance, at what point should players start playing a lot of high numbers? Some biases are so small they aren't worth bothering with.

The answer to the above question boils down to an individual players preference. I like to use the last 10 draws (the last fifty numbers drawn). A normal bias for ten drawings would be a 25 low/25 high number split which is a zero bias. Because there isn't a bias, you cant act on it. But if I saw 35 lows and 15 highs (10 more low numbers were drawn in the last 10 drawings than "normal" or the expected number which is 25) then it's probably time to start picking more high numbers than low numbers. I'll usually pick at least three high numbers for every line I play. If a lot of low numbers continue to be drawn, the higher the number of lows will become, which simply means lows are getting more out of balance, and you should continue picking high numbers. That out of whack bias you saw will eventually come down. (It will "even out over time") When it does, stop picking predominantly high numbers.

I've taken a look at using the last 100 drawings (the last 500 balls/numbers drawn) but I cant figure out at what point a bias becomes playable using 100 draws/500 numbers. I'm not too sure if using the last 100 draws is a good sampling to use or not. **G5**