Transfers

Start-of-season transfer market guide

This is a 2017-18 version of a basic guide I wrote at the start of the 2015-16 season to answer questions about player price changes in Fantasy Premier League (FPL). For many visitors to this site the information will be familiar, but hopefully the guide will be useful for newcomers and, maybe, a reminder for others. Some seasoned readers will undoubtedly be aware that some situations are more complicated than explained here, but I have intentionally tried to keep things simple for newcomers.

What we know for sure

All we know for sure are a few details written in the game rules and FAQs:

  1. Player prices change during the season dependent on the popularity of the player in the transfer market. Player prices do not change until the season starts.
  2. Once the season has started, player prices may change by £0.1m a day, based on activity in the transfer market.
  3. The price shown on your transfers page is a player’s selling price. This selling price may be less than the player’s current purchase price as a sell-on fee of 50 per cent (rounded up to the nearest 0.1m) will be applied on any profits made on that player.

To be clear, it is the sell on fee which is rounded up, not the profit. The difference between 7.0m and 7.3m is 0.3m. The 50 per cent sell on fee is 0.15m but as this has to be to the closest 0.1m, it is rounded up to 0.2m. Take the rounded up sell on fee from your sale price of 7.3m and you are left with 7.1m.

The following table demonstrates how this works:

Price change example

The price paid for a player when you bought them, the current market price for player, and the sale price of player for each player in a team is available by clicking on the List View tab above the team in the transfer page in FPL.

What we think we know

Beyond this information, we can only start seeing how the transfer market starts working each season once the first price changes have occurred. There have been several consistent features of the transfer market in recent seasons and the first 12 days of the 2017-18 season have yet to suggest any major changes.

Price changes are based on net transfers in (NTI). Put simply, if Player X had 6,000 transfers in and 5,000 transfers out, his NTI is 1,000. If he had 5,000 transfers in and 6,000 transfers out his NTI is -1,000. This figure keeps going up or down as a player receives more transfers in or out. Once a day the FPL checks if the player had hit his NTI threshold for a price rise or a price fall. If they have, they rise or fall 0.1m, the NTI counter is then reset to zero and the count begins again as the player is transferred in and out by more Fantasy managers. If a player hasn’t hit the threshold, nothing happens; the NTI counter carries on from whatever NTI level it had reached.

The price change has usually occurred in the middle of the night UK time. Some of the more popular price change prediction sites (more on them later) will carry a summary of the changes by the time the UK wakes up.

It is thought the threshold for a price rise is based on the number of people actively playing FPL (i.e. making fairly regular transfers), while the threshold for a price fall is based on the number of people who own a player. The distinction is important as it affects the number of transfers required for a player to move in price. Every player has the same rise threshold, but each player would have an individual price drop threshold. Let’s use an example where Player A had 10,000 owners and Player B had 1,000,000 owners. If the price rise threshold was, say, two per cent of all active players and there were 2,000,000 active players, both players would require 40,000 net transfers in before their price could rise. But if the drop threshold was ten per cent of each player’s ownership, Player A would require an NTI of -1,000 to drop in price, while Player B would require an NTI of -100,000 to drop in price. After a price change, the threshold counters would reset.

However, different price-drop rules apply to players who are above their start-of-season price because their falls happen more quickly. One theory is that the drop threshold for a player above their start-of-season price is based on a negative version of their NTI threshold for a rise. If this theory is correct, Player Y, who started the season at 9.5m and now costs 9.7m with 500,000 owners and a price rise threshold of 40,000 NTI, would have a price fall threshold of -40,000 rather than the -50,000 it would be if his current price was 9.4m with 500,000 owners.

Price changes have historically been limited to a maximum rise or fall of 0.3m in a gameweek. In 2013-14, price rises of 0.3m in one gameweek were quite possible for players who everyone was rushing to bring in as the same threshold applied to each price change, but in 2014-15 the FPL appeared to change the rules around the price change thresholds for the second rise in a gameweek so that it was higher than that of the first rise (once a new gameweek began the higher threshold limit returned to normal). The rate for a third rise within one gameweek was even higher still, so a triple price change in one gameweek became much rarer. Until we have more dramatic gameweek transfer activity we can’t be absolutely sure what will happen this season, but the early evidence suggests there has been no return to the 2013-14 approach.

There were several other factors that altered NTI and price thresholds in recent years and these included a player being flagged – i.e. having a yellow, orange or red “i” (for injury/suspensions) triangle beside their name on the transfer screen. The price drop threshold for a player with a yellow flag was higher than for an unflagged player. It was higher still for a red flagged player.

Players have been temporarily locked, or protected from price movement at times in past seasons. The NTI counter for locked players also resets to zero. This usually happened in one of two scenarios:

  1. When a new player was added to the game, their price was locked for eight days or until the start of a new gameweek (where the gap between gameweeks was longer than eight days). Much like pre-season, this allowed the player to build up an ownership without their price changing. In 2015-16, at least one player appeared to be locked for more than eight days and this should continue to be monitored.
  2. After the status of a flagged player changed to so they were no longer flagged.

Transfers made by managers with an active wildcard were also believed to affect NTI figures, with one popular theory being that they don’t count at all towards NTI. If this is correct, a player who had 10,000 transfers in, including 5,000 on a wildcard, plus 1,000 non-wildcard transfers out, would only have had an NTI of 4,000. Wildcard transfers had the potential to skew how quickly a player rose or fell in price in a gameweek.

New to FPL for 2017-18 is the Free Hit chip, which acts like a one week wildcard, before the old team is restored the following week. I’ve not seen any discussion about the impact on NTI of transfers made using the chip, but if wildcard transfers don’t count towards NTI it would seem logical that Free Hit transfers wouldn’t count either.

Blank and double gameweeks also appear to slow the speed of price falls and price rises for the players involved. This phenomenon has only recently been spotted and the specifics are not well understood yet.

Predicting the price changes

FPL participants can get a very rough idea of which players might be heading for a rise or fall in a particular gameweek by selecting transfers in (round) and transfers out (round) from the “Sorted by” drop-down menu on the Transfers page of FPL. However, remember that a player’s NTI counter only resets to zero if they change in price or are locked. If Player B in our earlier example had an NTI of -80,000 between Gameweeks 1 and 2, then an NTI of 60,000 the following week, he would neither rise nor fall because his NTI thresholds were -100,000 for a fall and 40,000 for a rise and his counter at the Gameweek 3 deadline is at -20,000 (-80,000 + 60,000).

Several price change prediction websites operate by monitoring the number of player transfers in and out, adjusting for the various factors that can affect NTI, and matching those figures against estimated price rise and fall thresholds to predict when a player is likely to increase or decrease in price. These sites display this as a percentage, with a player being predicted to rise when they reach or exceed 100 per cent, or fall when they reach or exceed -100 per cent. My favourite price change prediction site is FPL Statistics.

At the start of the season these prediction calculations are largely based on experience of how the price changes worked in previous years. Therefore they can be imprecise, particularly if, as in 2014-15, FPL changes the price rise and fall formula. Once the first price changes happen, the price change prediction websites will start to refine their threshold estimates and NTI calculations. As the season progresses they usually become more accurate.

During the first few days of the 2017-18 season the prediction sites have done fairly well at forecasting which players will be rising and falling in price each night, which suggests there haven’t been major changes in the price change formula this year. However, we will need more data to gain a clearer picture.