Pre-emptive training is using the available data to figure out whether a player's form is likely to drop after the next match, and initiating training ahead of that so as to reduce the form drop. The risk in pre-emptive training used to be the risk of overtrains, and the upside was training a player after just 1 red arrow and preventing a large form drop.
In Oct 2013 Xpert11 player overtraining was removed which removed takes away part of the advantage of preemptive training, but hass created something of a culture of "if in doubt... train" which will be successful enough - but very expensive.
Players have a form trend, the change since the last form notation, and this is represented by a green, black or red arrow. In a simplistic world you would train on a red arrow and only on black/green if your form is low. The problem is this green arrow and change in form since the last notation (as shown on the VIP graph) is made from a combination of factors and isn't perfect for judging when to train.
Players also have the form tendency. This is a hidden stat and is how the form is naturally changing and is what we want to base training decisions on. Problem is we can't see it. The tendency is combined with other factors to create the new form value, but if the tendency is negative then it can get more negative after the next game and drop very sharply.
After a match there are several factors that are added to the tendency to calculate the current form:
Player performance in the match
If a player performed over his skill level during the match his form will be adjusted upwards accordingly. if he underperforms his skill level then his form will be adjusted down.
Team part performance during the match
If the team part performs well then there is a positive form adjustment to the player and if the part performed badly then there is a negative adjustment. Team parts have the following roles:
Goalkeeper - stop goals being scored. Typically conceding around 1 goal in 3 chances seems to be the baseline so better/worse than this will accect form accordingly.
Defence - their role is to limit chances. From personal experience I would say 4-5 chances is the benchmark so conceding more/less than this will affect the defenders' form accordingly.
Midfield - their role is to create chances. Similar to defence, around 5 chances seems to be the baseline
Forwards - their role is scoring chances. Scoring around 1 in 3 chances seems to be the benchmark here.
Deciding when to train
There are no defined rules how much player or team part performance affect form so it really comes down to experience and guesswork. You look at how much form changed since the last match, add or subtract maybe 2-3 points for player/team performance and if the form should be higher it implies a negative tendency and you train.
The following examples may help:
In this example our player has basically flat form but that's partly due to his performance of 8 from the last match. Without the boost from a good performance his form would have dropped 1-2 bars. Check the team performance but likely train.
This is a case where the player probably does not need form training. His form is again pretty much unchanged but his last match performance was poor giving him a negative form adjustment after the match.
Similar to the above example there is a slight drop in form here but the poor match performance will have caused the form to dip 1 or 2 bars. Look at the match performance and base your decision on that.
All the above examples are close but in the bottom 2 it is probably not necessary to train. A lot will also depend on the player. If it's a big investment then you might want to aim higher and train more often to help guarantee good form. If it's an older player or money is more of an issue I definitely would leave the last 2 players to see what happened after the next match.
TIP
As the player's performance in a match affects his form slightly many players always have a 50k win bonus even in friendlies. This is a cheap way of slightly affecting the form on 11+ players ina positive way.
