On the NTT20 podcast this week, George Elek suggested our underlying numbers were a little concerning, and he didn’t see a top two assault continuing into 2026.
Top six, yes. Automatic promotion? No. He’s backing Stockport, based in part on xG data. As a staunch believer in xG, and given George’s arguments around Plymouth in 2018/19 were where my xG love started, I should be worried, right?
Ali Maxwell, suddenly the more popular NTT20 presenter among some Imps, disagrees. He feels the numbers don’t tell the full story, and in our Discord channel, the debate raged on. It’s all good-natured, by the way, but there was a split opinion.
So which is right? Should we be worried about our underlying numbers? Or, like a library book with 100 pages missing, does it only tell half the story?
Fan Opinion
For the argument
Those more positive about Lincoln’s outlook felt the discussion was balanced and fair, with our Chris noting that Ali’s optimism had substance behind it. The core defence was that xG figures are heavily distorted by game state, something Ali highlighted and Chris expanded on in detail.
Lincoln score first more often than anyone else, lead at half time far more frequently, and spend longer ahead in matches than their rivals, all of which naturally suppress attacking xG while protecting leads. Paulus and PhilB424 added that form, confidence, and momentum are difficult to quantify, while Drew H argued there is no statistical law that says overperformance must regress if a team is simply good.
Several posters welcomed deeper analysis (this article), especially game state-specific xG when level, ahead, or behind, with Chris offering to produce data to give a fuller, fairer picture.
Against the argument
Sceptics accepted the logic but questioned whether game state alone could justify the scale of Lincoln’s xG overperformance. DridImp argued that other teams spend similar amounts of time ahead without showing the same statistical gap, suggesting tactical choices when leading may be more significant than narrative explanations.
George Elek’s original caution centred on sustainability, with the view that outperforming both xG and xGA over a long run is unlikely to continue unchanged. Some contributors were uneasy with how xG is interpreted at all, debating whether summing probabilities meaningfully reflects reality, particularly in scrambles or phased attacks.
While most agreed data has value, the concern remained that relying too heavily on any single metric risks overstating confidence and underplaying future variance, especially if goals were to dry up.
Context
Before we go further, it’s also worth mentioning the validity of xG. I loved xG, I was a massive fan, and to a degree, I still am. However, I do have a few issues; for instance, a cross that just evades a striker’s boot is not included in the xG data. When Jack Moylan went one-on-one with a keeper a few weeks ago (I forget the game), he was tackled without getting a shot away. That’s zero xG.
They’re both chances, but the data doesn’t show them as an attacking threat in the xG number. There isn’t a stats model available to us that does show that, but the club uses a different metric. They have models which take into account the attacking threat, passes that break lines and chances where a shot isn’t taken. I have seen the data, but obviously can’t put it out here as it is something the club (rightly) sees as a competitive advantage.
It has made me think differently about how we view our attacking threat. There was one game recently (again, I’m playing it coy as the data can’t really be in the public domain) where the Wyscout numbers suggested a draw, 0-0. However, the club’s numbers for raw xG suggested a 1-0 win for City would have been right, and the numbers for bypassing defenders and creating chances, but not necessarily shots, showed us to be twice as good as our opponents.
Of course, there is then the argument about taking those chances, but that’s a debate I’m happier to have when we’re second, rather than trying to defend what we’re doing when we’re 17th. Anyway, you’re not here for my slow turn from xG fan to critic, are you?
The Data
So, what does the data look like. Presented without comment (at first, obviously, otherwise there would be no point) are the tables, researched by Chris, which show game state.
Lead durations (overall) in percentage |
| Team | Leading | Level | Trailing |
|---|---|---|---|
| AFC Wimbledon | 21.70% | 56.70% | 21.60% |
| Barnsley | 32.90% | 40.70% | 26.30% |
| Blackpool | 27.20% | 42.90% | 29.90% |
| Bolton | 26.00% | 46.00% | 28.10% |
| Bradford | 31.80% | 52.90% | 15.40% |
| Burton Albion | 23.50% | 49.90% | 26.70% |
| Cardiff City | 29.10% | 58.30% | 12.60% |
| Doncaster | 18.20% | 54.60% | 27.20% |
| Exeter City | 21.50% | 49.30% | 29.30% |
| Huddersfield | 30.40% | 47.00% | 22.60% |
| Leyton Orient | 24.80% | 45.90% | 29.30% |
| Lincoln City | 47.20% | 41.40% | 11.40% |
| Luton Town | 25.10% | 45.40% | 29.50% |
| Mansfield | 27.60% | 44.20% | 28.30% |
| Northampton | 15.60% | 57.90% | 26.40% |
| Peterborough | 25.40% | 42.10% | 32.50% |
| Plymouth | 22.20% | 40.20% | 37.60% |
| Port Vale | 12.10% | 59.50% | 28.40% |
| Reading | 26.70% | 46.10% | 27.20% |
| Rotherham | 21.50% | 44.90% | 33.60% |
| Stevenage | 21.20% | 65.60% | 13.20% |
| Stockport | 32.50% | 47.80% | 19.70% |
| Wigan Athletic | 21.70% | 49.00% | 29.40% |
| Wycombe | 30.70% | 37.90% | 31.40% |
| League average | 25.70% | 48.50% | 25.70% |
Data sourced from soccerstats.com. Percentages based on 90 minutes per game.
xG Levels And Game State
| Team | Level | Ahead | Behind | ||||||
|---|---|---|---|---|---|---|---|---|---|
| XG | XGA | R | XG | XGA | R | XG | XGA | R | |
| AFC Wimbledon | 0.65 | 0.69 | 48.5% | 0.37 | 0.61 | 37.8% | 0.41 | 0.55 | 42.7% |
| Barnsley | 0.47 | 0.78 | 37.6% | 0.79 | 0.82 | 49.1% | 0.60 | 0.51 | 54.1% |
| Blackpool | 0.43 | 0.66 | 39.4% | 0.62 | 1.01 | 38.0% | 0.56 | 0.80 | 41.2% |
| Bolton | 0.73 | 0.43 | 62.9% | 0.63 | 0.56 | 52.9% | 1.10 | 0.50 | 68.8% |
| Bradford | 0.85 | 0.56 | 60.3% | 0.56 | 0.54 | 50.9% | 0.97 | 0.86 | 53.0% |
| Burton | 0.65 | 0.77 | 45.8% | 0.78 | 0.76 | 50.6% | 0.53 | 0.65 | 44.9% |
| Cardiff City | 0.79 | 0.62 | 56.0% | 0.73 | 0.53 | 57.9% | 0.66 | 0.54 | 55.0% |
| Doncaster | 0.74 | 0.60 | 55.2% | 0.26 | 0.55 | 32.1% | 0.71 | 0.78 | 47.7% |
| Exeter City | 0.42 | 0.75 | 35.9% | 0.65 | 0.66 | 49.6% | 0.55 | 0.47 | 53.9% |
| Huddersfield | 0.65 | 0.57 | 53.3% | 0.86 | 0.47 | 64.7% | 0.66 | 0.49 | 57.4% |
| Leyton Orient | 0.51 | 0.53 | 49.0% | 0.46 | 0.54 | 46.0% | 0.84 | 0.83 | 50.3% |
| Lincoln City | 0.69 | 0.46 | 60.0% | 0.60 | 1.02 | 37.0% | 0.53 | 0.48 | 52.5% |
| Luton Town | 0.55 | 0.51 | 51.9% | 0.87 | 0.56 | 60.8% | 0.77 | 0.61 | 55.8% |
| Mansfield | 0.56 | 0.74 | 43.1% | 0.62 | 0.51 | 54.9% | 0.56 | 0.68 | 45.2% |
| Northampton | 0.46 | 0.83 | 35.7% | 0.61 | 0.39 | 61.0% | 0.59 | 0.78 | 43.1% |
| Peterborough | 0.71 | 0.50 | 58.7% | 0.66 | 0.75 | 46.8% | 0.64 | 0.63 | 50.4% |
| Plymouth | 0.38 | 0.47 | 44.7% | 0.68 | 0.98 | 41.0% | 0.74 | 0.96 | 43.5% |
| Port Vale | 0.77 | 0.60 | 56.2% | 0.53 | 0.38 | 58.2% | 0.69 | 0.47 | 59.5% |
| Reading | 0.48 | 0.61 | 44.0% | 0.74 | 0.77 | 49.0% | 0.61 | 0.46 | 57.0% |
| Rotherham | 0.39 | 0.61 | 39.0% | 0.39 | 0.88 | 30.7% | 0.63 | 0.70 | 47.4% |
| Stevenage | 0.59 | 0.64 | 48.0% | 0.29 | 0.56 | 34.1% | 0.29 | 0.18 | 61.7% |
| Stockport | 0.65 | 0.55 | 54.2% | 0.50 | 0.58 | 46.3% | 0.77 | 0.55 | 58.3% |
| Wigan Athletic | 0.74 | 0.54 | 57.8% | 0.65 | 0.65 | 50.0% | 0.64 | 0.40 | 61.5% |
| Wycombe | 0.54 | 0.45 | 54.5% | 0.93 | 0.74 | 55.7% | 0.78 | 0.59 | 56.9% |
XG ratios colour coded by match state. Green indicates positive control, amber neutral, red negative.
Conclusion
There is some important context behind these numbers that is worth spelling out. We spend by far the longest amount of time ahead in League One, while also spending the shortest time behind and the fourth shortest period drawing.
That matters when looking at raw expected goals figures. Our total xG conceded when leading sits at a high 1.02, but that is partly a function of volume rather than pure weakness, because opposition sides simply have far longer to build attacks against us once we go in front. The more telling number is the ratio. When ahead, we concede 42% more xG than our opponents, with an xG ratio of 37%, which is the fourth worst in the division. Compared to the rest of League One, that does point towards a vulnerability in how we manage games once we have the advantage, even allowing for how long we are in winning positions.
Elsewhere, the picture is more encouraging. When matches are level, our xG ratio is the third best in the league, which suggests we are generally effective and balanced when the game is still there to be won. When behind, although we rank 14th for xG ratio, it is worth stressing that we are only chasing games for 11.4% of total minutes played, the lowest figure in League One.
In that context, creating 0.53 xG when behind is actually a positive and notable return. Overall, the data shows a side that is ahead more than anyone else, rarely falls behind, and performs well in drawn situations, but also one that could tighten its grip on matches after going in front.
There is plenty more that could be unpicked within this, and doubtless the Discord debate will do that, but it does at least frame where the Imps are strong and where there is room for improvement across the season.
Was George Elek right? Was Ali Maxwell right? No. Neither were, and I’m not, nor is Chris, and nor are you. Football can be broken down into stats, and it is a good model for predictions, but football is also opinion, it is also art, and therefore nobody can accurately predict the future. When one stat comes out that I feel I can believe in, there are reasons to dispel that as well. Stats are a shield to hide behind, a stick to beat you with and an acid-based liquid to blind you permanently.
They’re not the future; they’re only a prediction model to help build opinion, and while they won’t tell you Maccabi Tel Aviv played West Ham when they didn’t, they will occasionally tell a story that might not be entirely aligned with reality. Unless you find a number that fits your story.
Fun though, aren’t they?