SZ System Flowchart


System One - Original Model (S1)

The original model utilises shots on goal data as a benchmark for assessing the relative strengths of the teams contesting a football

matchand each team builds up a record as the season progresses.

The data model uses fixtures from the 2005-06 season to 2009-10. The outputs were then back-tested with the 2010-11 and 2011-12


Teams must have played at least four home games and four away games before the match can be rated by the S1 algorithm.


System Two - Time Weighted (S2)

The Time Weighted model differs from the S1 model in that recent games are rated more highly by the algorithm than older

games as the name suggests. This introduces an element of 'short-term' form into the equation whilst maintaining a seasons record

for a measure of consistency.


System Three - Recent Form (S3)

The Recent form model concentrates on performance from the last 6 Home or Away games only with each match awarded the

same weighting. It is a well established fact that football teams fortunes can vary throughout the season and it is often the

teams that can show the greatest level of consistency over a season that prove to be the most successful in the long run.

System three is designed to identify the teams that are in form and offer the best opportunity for profit accumulation.

Teams must have played at least six home games and 6 away games before the match can be rated by the S3 algorithm.


System Four - Time Weighted Recent Form (S4)

The Time Weighted Recent Form system is essentially a hybrid of systems 2 and 3. The algorithm identifies teams records 

over the previous 6 matches but more recent games are awarded increased importance over older games. This system

places even greater emphasis on short-term form.


Putting It All Together

Analysis of the outputs from the individual systems indicated that the results were better as the systems were combined

with the best results obtained where selections appeared on all four sub-systems. As expected, the 'least profitable'

results came from single system selections. A Divisional Filter stage was added to 'weed out' the least profitable 

selections and the result is a single 'super-system' presented as a single set of selections as before.


The advantages of the multiple system approach are numerous but I'll outline the main benefits as I see them :-

  • System robustness - multiple systems produce more selections and help minimise short-term variances
  • Four pronged attack - can identify more value selections than a single rating system can do alone
  • More confidence - the majority of selections appear on multiple algorithms, hence more confidence
  • The results - the model and back-tested results outstrip anything I have ever produced so far!

Football Investor System

The Football Investor system utilises the four models above to generate match ratings for the home and away side based on their respective

records. The away sides rating is subtracted from the home sides rating to produce a match rating. A high positive match rating suggests

the home team is superior and vice versa for a high negative score. The match rating is then used to determine the 'fair' odds for each team

based on historic system data. Where the odds on offer are higher than the calculated fair odds, it is possible that the selection will qualify as

a bet (subject to other selection criteria).


Value Ratings

The Value Ratings are an attempt to measure how much value exists in a particular selection. The odds on offer are compared to the 

calculated fair odds and the value is expressed as a % of the fair odds. 

The selections with the highest % value should prove to be the most profitable in the long term but may be subject to

short term swings.