Possession can be observed on the match report, at the end of the first half (minute 45) and the end of the second half (minute 90). That number is calculated using a formula which takes on consideration the midfield level of both the teams at that particular minute.
Possession(your_team) = 100 * (Midfield_(your_team) / (Midfield_(your_team) + Midfield(opponents_team)) )
- The scale of the midfields values is: 1=disastrous (very low), 2=disastrous (low) ...
- It's also possible to use values: 0,00=disastrous (very low); 0,25=disastrous (low); 0,50=disastrous (high); 0,75=disastrous (very high); 1,00=wretched (very low) ...
- You can test an implementation of the formula at The HattriX-Ray
One consequence of the formula (in terms of gameplay) is that as the average midfield level gets higher, possession differences get smaller for the same difference in midfield levels. For example:
- disastrous (very low) vs wretched (very low) (1 level) = 71% possession
- solid (very low) vs 'excellent (very low) (1 level) = 54%
Also, the midfield values used in the formula are the ones on that particular minute, while the midfield rating that is displayed at the end of the match is the average of the midfield during the 90 minutes. This explains the cases where discrepancies between the possession percentages at the 45th and 90th minute and the midfield rating seem to appear. Events such as:
- red cards (to players affecting the midfield)
- injuries (to players affecting the midfield)
- weather non-goal events (to players affecting the midfield)
affect the midfield from the time the event is reported, hence affecting the possession as well.
Player stamina may alter possession minute-to-minute, if a player affecting the midfield is below excellent stamina, but this will not produce an event in the match report.
Additional factors that affect midfield ratings uniformly throughout the match include:
Possession vs goal opportunities nonlinearity
Statistical analysis over the goal opportunities distribution taking into consideration the possession levels suggest that the relation between the possession and the number of chances your team has is far from being 1:1. The numbers suggest that you will most likely get 9 or 10 chances in the game if you have 65% possession.
- When you have 54% possession, you will most likely get 6 out of 10 chances.
- When you have 57% possession, you will most likely get 7 out of 10 chances.
- When you have 62% possession, you will most likely get 8 out of 10 chances.
The formula is believed to be (for the % of chances for Team A): midfield_rating_A^e/(midfield_rating_A^e+midfield_rating_B^e), where e=2.72 (e is the base of the natural logarithm, http://en.wikipedia.org/wiki/Euler's_formula)
It is possible to see a a table which shows the percentage of chances that a team would get depending on the midfield of both teams. For example, if team A's midfield rating is wretched low (2.25) and team B's midfield rating is poor high (3.5), then team A will get the 23% of the chances.
Although this nonlinearity is widely believed, HT-Hasse is on record with a conference post stating that the chance a team will get a scoring chance at a particular point in the match is exactly equal to their possession percentage. This statement, however, has later been dismissed by HT-Tjecken, saying "If that's true, then Hasse was most probably hallucinating" in post 7831771.115.