Puntz,
I am not exactly a mathemetician but I have studied statistics at university level (ANU) and I am also somewhat of an expert with Excel (having worked for Microsoft in the past). Having tested Berts system yesterday and studied the underlying formulas, etc, in his spreadsheet I have the following opinion/conclusions/analysis (call it what you will).

The "target betting" is really no different to a loss-chasing staking plan. All Berts spreadsheet does is look at your daily expenses and losses so far and then sets the next bet amount to recover that amount plus your target ($5). eg. If you have $5 expenses and have bet $90 so far today then the next bet would try to win $100 (90+5+5) - so if dividend is $5 the bet would be $21. No great mathematical genius required for this - anybody could work it out with or without a spreadsheet.

The selection method - I have no real idea on how he came up with his selections but am sure it was through some formula (you couldn't do form analysis on that many races in the time available). Suffice it to say that based on the selections yesterday and some of his previous results there is no real magic here - the strike rate multiplied by average dividend is well under the value required to break even at level stakes.

So how has he won every day for 9 months?

Well the first point is that yesterdays results make this claim somewhat questionable - maybe he has but other people (ie. Me!) following HIS RULES EXACTLY with the information available made a loss yesterday. There are also other days in his results that show a loss following his rules (one day last month on dogs/trots and another day when there was a dead-heat).

In my opinion if he has won for 9 months straight (ignoring the losses listed on the website) this is nothing more than pure luck. The "secret" to his success is getting at least one $4 winner each day - if he does that he makes a small profit - if he has a day with no winner then he loses big time. In the long run can anyone guarantee at least one $4 or better winning selection? Of course not - sooner or later you are going to have days where you miss.

The simplist way to look at this statistically is to compare it to the old "doubling" scheme in roulette. Let's say I want to make $5 a day. Each day I go to the Casino and put $5 on red at roulette - if it wins I put the $5 profit in my pocket and walk away - if it loses I double it ($10) and keep doubling until I win and walk away with my $5 profit. My bet sequence is therefore 5, 10, 20, 40, 80, 160, 320, 640 - if the 640 bet loses then I blow my bank. The chance of this happening is roughly 0.5% each day - so 99.5% of the time I will win my $5, 0.5% of the time I will lose $1275. Now the chance of me being able to win every day for 9 months is around 26% (.995^270). Now the maths for Berts system will be a lot more complicated as it depends on his strike rate and the number of bets per day (ie. he will be more likely to survive a day with 50 bets than one with only 20 - but will lose a lot more if the 50 bet day fails) but the principle is exactly the same - if his selection system loses at even stakes (and he doesn't deny that it does) then in the LONG run (and 9 months is not really the long run) it will lose.

In summary if you want $5 a day then go to the casino and play roulette - if you're lucky it will work for a year and you will only waste 10 minutes of your time each day instead of many hours.