I’ve had a lot of questions about Munster’s last two games, and whether it represents a turnaround after a difficult few months. My answer to that is…. yes, but also no.

Isn’t that annoying?

Anyway, here’s the graph.

The For vs Against part of this data is interesting.

That view splits the single net line into the two halves it’s hiding, and what’s striking about Munster’s season is how independently the two sides of the ball move. The For line (green) and the Against line (brick red) aren’t tracking each other — they’re doing their own thing, and the net line is just the gap between them in any given week.

The For trend sits flat across the season at around 2.4 points per entry, which is a respectable but unspectacular attacking baseline. We reliably get into the 22 a healthy number of times per game, but their conversion rate of those entries into points is middle-of-the-road. The rolling line shows the attack actually has a clear three-act shape: a stable opening block in the 2.5–3.0 range, a dip through midwinter where January and February see us stuck below 2.5 for several games running (the Ulster Ravenhill defeat and the Leinster loss in Thomond Park are the worst at 1.5 and 1.0 respectively), and then a clear lift through the closing stretch — Glasgow, Benetton and Ulster home all produce For numbers above 3.5, which is why the green rolling line climbs steeply through the final five matches. The attack, in other words, is genuinely improving.

The Against line is where the season actually breaks down. Through the first nine games, it tracks below 2.5 with a couple of obvious exceptions — Connacht’s 3.0 at Thomond, and the Bath outlier at 5.7, which sits well off the chart and pulls the early rolling line up. Then, around match thirteen, something shifts: Castres put up 4.0 per entry, Dragons 4.2, Sharks 5.0, Exeter 3.5 — four of the five worst defensive numbers of the season come in a six-game window.

The brick red rolling line, which had been bobbing around 2.0–2.5 for months, spikes to nearly 4.0 and, annoyingly, stays there. That’s the defensive collapse the net view shows as a trough, but in the For-vs-Against view, you can see it’s not a collapse on both sides — our attack didn’t disappear in those games, the defence just started leaking at unprecedented rates. The closing two fixtures pull the Against rolling line back down sharply, but the trend line for Against still tilts upward across the season, which means defensively we have ended in a worse place than we started.

So the two trend lines together tell the story the single net line can’t: a team whose attack has gradually got better and whose defence has gradually got worse, the two effects cancelling each other out almost exactly across 21 matches. The flat net trend isn’t a flat season — it’s two genuine trajectories pulling in opposite directions. The implication for the closing URC games is sharper than the net view suggests: if the late attacking lift holds and the defence reverts to its early-season levels, this is a side with a really high ceiling. If the defensive numbers from the Castres-to-Exeter window are nearer the truth, the closing run could end up being a flattering finish against beatable opposition. The For vs Against split is where that question has bought a nice two-bed apartment, and is slowly settling into the neighbourhood.

In a broad sense, Munster’s season has no real direction, and the trend line is what makes that obvious.

We started it at +0.06, ended round 16 and post-Europe at −0.07, slope essentially zero — that’s not a side getting better, and it’s not a side getting worse, it’s a side oscillating wildly around the same midpoint we started at.

The season-long net efficiency averages out to nothing significant: 2.42 points per 22-metre entry created against 2.43 conceded. That sounds like a team treading water, except the rolling line shows the water is anything but calm.

The single-game numbers show the volatility figure directly: +3.3 at home to Ulster on Saturday, and four rounds earlier −5.0 in the 45–0 horror show against the Sharks in Durban. The swing from best to worst single-game is more than eight points of net efficiency — the kind of variance that suggests a team whose performance is more sensitive to opposition, conditions and selection than to any underlying improvement or decline. We have produced 4.5 pts per entry at home and 0.0 against South African opposition inside five weeks of the same season.

Make it make sense. You can’t. I can’t.

The four phases the chart picks out are doing the work that the trend line doesn’t clearly show. The opening block, matches one to six, sees Munster grind out five wins from six, but with a per-entry signature that’s almost flat — the Stormers loss closes a phase where rolling net never strays far from zero. The Bath collapse in match seven (1.4 created, 5.7 conceded) drags rolling net to its first real trough, partly recovered by the Gloucester and Ospreys wins, before the festive Leinster defeat at Thomond Park and the Ravenhill thrashing by Ulster confirm a familiar problem: against the elite tier of opposition, the per-entry numbers don’t keep up.

Phase three is the defensive collapse — Castres at home concede 4.0 per entry, Dragons leak 4.2 in a one-score win, Sharks score five per entry in the hammering, Exeter convert at 3.5 in the Challenge Cup last sixteen — and rolling net bottoms out at −2.47, by some distance the lowest point of the season.

And weirdly, I don’t think the defensive issue is systemic either. Outside of the data, when you actually watch the games, it feels like when we concede killer scores, they’re almost always off the back of a brutal mistake. An intercept, a bad play option that leads to a critical turnover, a ball popping out of a ruck, a killer missed tackle or a set piece implosion.

Is that bad defence? I mean, yes, it is — from a purely numbers perspective, the tries were conceded, and that’s that — but is that stuff coached or is it a collection of uncoachable errors? Maybe it’s both. A function of distraction. Or maybe it’s trying to get the balance right between our attacking concepts morphing mid-season, and what that actually means for the defensive side of the ball, as both are intrinsically linked.

Then phase four arrives: Benetton dismantled 45–15 in Treviso at 3.7 per entry, and Ulster torn apart 41–14 at Thomond Park on the season-best 4.5. Rolling net swings from −2.47 to +1.10 over three games.

The flat slope is not necessarily a verdict. It’s the absence of a verdict. Two fixtures remain — before any prospective post-season — and they will tip the slope one way or the other. A strong finish steepens it upward and reframes the Sharks–Exeter trough as a blip inside an improving side hindered by core players being unavailable for the Sharks, and a torturous travel schedule; a flat finish keeps the season as it currently reads.

Essentially, the data hasn’t told its full story. Not yet. The signature so far is range rather than direction: best and worst rugby of the year produced inside the same five-week window, and a regression line that, unable to choose between up and down, good and bad, sits at zero.

Whatever Munster do at Connacht and at home to the Lions becomes the season’s actual conclusion.

Until then, all the chart will tell you is that everything is still to play for.