Fact check: Does FUT have an unnatural amount of even results?

Share

In this article, we drive the final nail in the coffin for the idea that EA has implemented some kind of logic to level the playing field.

It’s a widespread belief that FUT uses a rubber banding mechanism in online matches, which for some reason have been dubbed “momentum” in parts of the community. The alleged logic behind this perception is that EA shifts the momentum around in order to keep matches tight. If it exists, a match leveling feature will leave a fingerprint on match results by increasing the probability of equalizing / gap narrowing goals. This will, all other things equal, increase the occurrence of ties and narrow wins in general.

Given the above, we should be able to test whether a rubber banding mechanism exists by testing whether FUT’s actual draw ratio (i.e. the percentage of matches ending as a draw) exceeds it’s natural draw ratio (i.e. the percentage of matches which will end as a draw due to natural reasons).

We already know the actual draw ratio for FUT 16 seasons. During FIFA 16, we collected match data for more than 2,000 players through the web app’s game data section. On average, the players in or sample had a draw ratio of 18 %. Determining FUT 16 seasons’ natural draw ratio is of course are trickier affair, but we have come up with an educated guess.

What controls the draw ratio?

How could we determine FUT’s natural goal ratio? A place to start may be to look at what factors determine the draw ratio. We know from real life, that different sports have different draw ratios. Football has a relatively high draw ratio (26 % draws) compared to sports like handball and basketball.

A reasonable assumption would be that this in part is explained by football having a lower goal ratio. If an average of 2.7 goals per match produces a draw in 26 % of the matches, it’s hardly surprising that only 8 % of handball matches end as a draw, given that handball has a goal ratio of 57 goals per match and therefore a much lower probability of the involved teams scoring the exact same number of goals. The same pattern can be found for sports like rugby and basketball.

If FUT was a real league

How strong is the connection between goal ratio and draw ratio? As it turns out, it’s pretty strong. The draw ratio in football varies considerably from league to league, and if our assumption is right – that the goal ratio determines the draw ratio – we would expect leagues with more draws to contain fewer goals.

In the chart below, we plotted draw ratios and a product of the goal ratio for 230 different football leagues. Statistical tests confirm the immediate impression that leagues with more goals generally have fewer draws. We can even draw a trend line, which works as a prediction of the natural draw ratio at various goal ratios based on the actual sample. According to our trend line formula, [draw ratio] = 64,192 [goal ratio]^-0,944. In other words, if FUT has a normal draw ratio, we should expect to see it fall onto that pattern. If, on the other hand, FUT has a “momentum feature” leveling out the results, we would expect it to have a considerably higher draw ratio.

goal_ratio_vs_draw_ratio_incl_fut
Data for the 2016/17 season. Each blue dot is a league. Only leagues, where minimum 50 matches had been played at the sampling time are included.

Now, as you might have guesses, we already added FUT to chart: It’s the red dot. What is clear is that it lands pretty close to the predicted line. In fact, the red dot is located slightly below the curve, meaning that FUT in fact has a slightly lower draw ratio that we would expect, given it’s goal ratio. If FUT had a momentum feature, we would have seen the red dot hovering well above the red line. It clearly doesn’t.

Using the trend line formula (y = 64,192 x^-0,944), we predicts FUT’s natural draw ratio as 19%, which isn’t far from the actual result:

FUT 16 actual FUT 16 predicted
Draw ratio 18% 19%

(Actual and predicted draw ratio) 

FUT’s actual draw ratio is nowhere near the 30-40’ish % we would expect to find, if EA had put a match leveling mechanism into the code.

What this tells about match leveling

The results above do not by any means support the assertion that FUT matches are made even. The fact is that they contain considerably fewer draws than real matches, and that the draw ratio corresponds almost exactly with the expected draw ratio.

Based on these calculations, we basically can reject the match leveling theory: If there was any match leveling going on, we would have found a residual of draws that we couldn’t explain by natural causes.