Sooner or later it happens to almost everyone. For weeks the target looks better, the arrows land closer together, and at the end of the session the scorecard shows much the same number as before. Or the opposite happens: the score goes up by a few points and the group, when you look at it, is as wide as ever. Archers often ask me the question that follows from this, which is why the score does not tell them whether they are grouping better.
The answer lies in how the score is built. Each arrow is worth whatever its distance from the centre of the target earns it, and that distance depends on two separate things: how wide the group is and where its centre sits. The scorecard adds them up into a single number, so progress in one can be swallowed by a step back in the other. The studies that recorded where each arrow landed, rather than counting arrows by ring, show whether this happens and how far the research has got.
The two measures the score combines
Coaches and researchers use different names for them, precision and accuracy, grouping and centring, but the meaning is the same. The first tells you how close the arrows are to one another, in other words how well the shot repeats; the second tells you how close the centre of the group is to the centre of the target, and it depends on the shot but also on the sight, the wind and how the bow is tuned. They are corrected in different ways: moving the sight moves the centre of the group without tightening it, while a tighter group does not move towards the centre on its own.
The score does not tell them apart. A tight group sitting entirely to one side and a wide but centred group can give the same total, and the same applies to changes over time: if over a month the group tightens and over the same month its centre drifts sideways, the scorecard can stay flat. In competition, centring weighs more than you might think. A study that photographed the target faces at two major international events in 2013 recalculated the points lost to off-centre groups, and concludes that on average archers centred their group to within about 60% of its size, with the exception of the best men's compound archers, who did markedly better. The offset of the centre, then, is not small compared with the width of the group, and it counts towards the score alongside it. I went into what an off-centre group costs in the article on arrow impact patterns, and in the registry the claim that at major events many archers lose points to a badly adjusted sight is reported.
The study that measured score and group on the same shots
That two archers on the same score can group differently follows from the geometry of the target. Seeing it happen, though, takes a study that records both the score and the position of every arrow on the same shots, and the registry holds one, published in 2026. Twelve young Indonesian archers, aged 12 to 15 with two to ten years of practice, shot fifteen arrows each with an electronic bow at a virtual 122 cm target shown on a television screen 1.5 metres away. For each archer the authors calculated the score, out of a maximum of 150, and the spread, meaning the average distance of the arrows from the centre of their group.
Scores ran from 104 to 144 and spread from 5.2 to 22.3 cm. At the extremes the two measures agreed, since the tightest group belonged to one of the two archers on 144, the top score in the sample, and the widest to the archer with the lowest score. In between, they came apart. In the per-athlete chart, two archers on about 125 points, one point apart, had spreads of about 10.5 and about 19 cm, nearly double. The pair the authors cite in the text, on 135 and 136 points, has almost identical groups, 9.2 and 9.6 cm: the archer with one point more groups slightly worse, but the difference is four millimetres.
This is why the registry reports the claim that two archers with the same score can group very differently, and goes no further. The study is small, the decisive case is read from a bar chart rather than a table, and above all the conditions are a long way from a competition field: an electronic bow, a target on a screen, fifteen shots each. It shows that a score can hide very different groups, not how often that happens at 18 or 70 metres with your own bow in your hand.
Between distant levels, the group follows the level
Move from the gap between two archers close in score to groups of archers at different levels and the picture changes, because between a beginner and a national squad archer the distance is so great that the higher level and the tighter group go together. The registry holds seven independent samples, all pointing the same way. The oldest dates from 1990: at eighteen metres the group outline of three Olympic archers ran from 7.8 to 10 cm, that of six Belgian national-level archers from 11.9 to 26.7. In 2017, at thirty metres, ten Olympic-level archers from the British squad were more precise than ten club beginners while taking the same time per shot.
The study that separates the two measures best is a 2014 congress abstract on twenty-nine Turkish recurve archers split into three levels, with seventy-two arrows each at eighteen metres, photographed after every end and plotted on a coordinate grid. The high-level archers had the tightest group and the centre closest to the centre of the target, and the beginners' spread was five to seven times theirs. The middle group, though, was the most off-centre of the three, with a much smaller spread than the beginners' and a centre further from the middle of the target. The author sums up the three levels in just these terms: high level accurate and precise, middle level precise but not very accurate, beginners accurate but not very precise.
The figure comes with two caveats. The values are for each level taken as a whole, not for individual archers, and seven beginners whose groups are off in different directions can have an average close to the centre even if none of the seven is centred; the abstract, moreover, does not give the unit of measurement. That said, group width falls in step with level and centring does not, which is the same split between the two measures that showed in the two archers on 125 points, this time across groups of people.
In competition the gap shows on the target faces. In the photographs from the 2013 Shanghai World Cup the top twenty men's compound archers lost an average of 0.30 points over 36 arrows to off-centre groups, against 1.35 for their division as a whole. Within a national squad, over ninety arrows at eighteen metres, the best archer shot no six and no seven and the worst one six and fourteen sevens, so level showed in the tail of the worst shots more than in the average. Taken together, these studies leave the claim that the higher the archer's level, the tighter and better centred the group at reported: they are all comparisons between different archers, two are congress abstracts, and in competition the better centring of the leading archers shows clearly only in men's compound.
The two readings, the young Indonesian archers and the levels, fit together. When the difference between two archers is large, the tighter group goes with the higher score, as at the two ends of the Indonesian study; when it is small, centring can weigh as much as width, as in the pair on 125 points. Progress over a few weeks is almost always small, and that is exactly where the scorecard is likely to miss it.
Your teammates' scores as a yardstick
Many archers judge whether they are improving by comparing themselves with clubmates or with the average for their category, and this yardstick is less steady than it looks. Whenever a study reports individual scores, the spread within one group is wide: twelve Indonesian club archers ranged from 188 to 325 points over 36 arrows at 30 m; four Russian masters of sport, over sixty arrows each, shot 31, 28, 27 and 17 tens; nine competitive Korean high-school archers, from the same sports school, averaged between 7.3 and 9.1 per arrow at 50 m. With around ten independent samples pointing the same way, from beginners to masters of sport, the claim that within one group of archers, scores differ widely from one archer to another is corroborated.
As the level rises the spread narrows, without closing. In the Chinese national squad the average per arrow at seventy metres was 8.7 with a standard deviation between athletes of 0.5, in the Beijing squad 6.8 with 1.4, and even in the national squad individuals ranged from 7.7 to 9.4. Two small cross-sectional studies show this, and the claim that as the level rises, scores within a group of archers sit closer together is reported. All these figures share one limit: they are snapshots of a single day, and they do not say how much of the gap between two clubmates is a stable difference and how much is each one's day. If your total matches a clubmate's, then, you still do not know whether you group the same way, nor whether you will still be close in a month's time.
Where the research stops
Lined up, these studies do not all say the same thing with the same force. That scores differ widely within a group of archers is the most solid part. That higher levels have tighter, better centred groups rests on seven samples, all cross-sectional. That two archers on the same score can group very differently rests on a single small study, carried out in conditions far from competition.
The question that matters to anyone looking at their own scorecard, whether in the same archer, week after week, score and group can move in different directions and how often that happens, has not been put to the test: nobody has followed the same archers over time recording score, group width and the position of the group's centre together. What can be said comes from the geometry of the target and from comparisons between people, which is why what follows is a way of measuring, not a promise of results.
What I take to the shooting line
The first thing I take from this is that if I want to know how I am grouping, I have to record where the arrows land and not only what they score. The site's tap target is built for this: you mark each arrow with a finger on a face drawn to scale and get, besides the score, the centre of the group and its spread, which are exactly the two measures the scorecard folds together. A photograph of the target face after each end does the same job. That arrow positions can be recorded by hand from a photograph comes from a 2012 study in which ten people with no archery experience marked, from photographs, the 72 arrows one archer shot at 70 m: the positions they marked matched the measured ones almost exactly, and the error weighed less on the position of the centre (0.57%) than on the size of the group (1.53% on the mean radius). Nobody has yet tried it with archers and coaches on photographs taken at competitions, so in the registry the claim that arrow positions can be recorded by hand from a photograph of the target face is transferred.
The second is to look at the two measures separately, and over enough arrows. If over a few weeks the spread comes down but the centre of the group moves, the score can stay flat even though the shot is repeating better, and the thing to watch is the centre, which can move because of the sight as much as because of the shot; if the score goes up and the spread does not, the improvement has probably come from centring. The centre of three or six arrows, though, wanders a good deal around that of the whole group, so it pays to read it over several ends before touching the sight, as I explained when discussing what an off-centre group costs.
The third concerns comparisons. A one-point difference between two sessions, or between you and a clubmate, says little, both because of the weight of centring and because the score of a round has its own natural fluctuation; how many points it takes to speak of an improvement is covered in the article on how many points it takes to say one archer is better. With the group measured, the comparison is made on the thing you are actually training. No study shows that tracking spread makes the score rise sooner, but it lets you see progress in the shot that the scorecard, by the way it is built, can keep hidden.