TL;DR

  • Season simulators reward repeatable points, not highlight-reel Sundays. A “consistent P5–P7” profile often beats a “win-or-bust” profile across 24 races.
  • The F1 points curve is steep at the top and then flattens. Zero is still zero — and a few zero-point weekends erase a surprising number of wins.
  • Consistency is more than “not crashing.” It’s variance control across the whole weekend: qualifying, starts, strategy, and conversion.
  • Build two driver profiles with different peak-vs-consistency tradeoffs in the Drivers tool and see how often each one wins the simulated championship.

A lot of F1 debate gets stuck on a false choice. Peak pace — who can win on raw speed — versus consistency — who can bank points when the weekend isn’t perfect.

A good simulator doesn’t pick a side. It turns both ideas into distributions, runs them across a calendar, and lets the points system do the scoring. And when you do that, the pattern that keeps showing up is stubborn: drivers who finish near their expected position most weeks quietly outperform drivers who spike higher but fall off a cliff more often.

This isn’t a philosophical claim. It’s what the math does. If you want to understand why — and model it without pretending to predict the future — start by stress-testing driver profiles in the Drivers tool.

Why simulations tend to “like” consistency

A season simulation is an accounting system. It doesn’t care how iconic a win looks on TV. It cares how many points land on the spreadsheet across 20-plus races. That’s why a consistent driver — lower week-to-week variance, fewer zero-point outcomes — often beats a peakier driver even when the peakier driver has the higher ceiling.

The core reason is simple. The F1 points structure is steep at the top and then flattens, but zero is still zero. A driver who alternates between P1 (25) and DNF (0) averages 12.5 points per race. A driver who mostly lives around P4–P5 (12–10) with very few DNFs can match or beat that average while looking less impressive in highlight packages. Over a season, the simulator favours the profile with fewer catastrophic outcomes — not because the model is conservative, but because championships accumulate, not peak.

Make this practical. Model the driver as a distribution: typical finishing range when nothing unusual happens, and tail risk when things go wrong (contact, penalties, reliability, strategy traps, Q1 exits). Compare drivers on expected season points, not on number of wins. Put the assumptions side by side in the Drivers tool and the numbers answer the debate faster than an argument can.

How the points curve prices volatility

Every finishing position is a discrete payout. That means your finishing-position distribution matters more than your best lap or your best race. Volatility isn’t automatically bad — it’s uncertainty, and the points table prices that uncertainty in a specific way.

The tradeoff: the upside of a peak weekend is capped at 25 points for the win. The downside of a bad weekend is not “a few points less” — it can be zero, plus the knock-on effects of grid penalties, component usage, and the dent in qualifying confidence that a messy weekend leaves behind. Run a season simulator and those low-end outcomes stack up faster than people expect. You don’t need many DNFs or P15s to erase the advantage of a couple of wins.

This is why “win probability” is a misleading headline metric. A driver can have a higher chance to win on any given Sunday and still be a worse championship bet if their floor is low. Treat win likelihood as one slice of the distribution in the Drivers tool, not the whole story.

Consistency is more than “not crashing”

When fans say “consistent,” they usually mean “finishes races.” That’s only one dimension — DNF probability. The broader effect is variance control across the full weekend pipeline: qualifying position → start → first-stint traffic exposure → strategy freedom → finishing position.

A driver with slightly slower peak pace can still score more over a season if they repeatedly avoid the high-variance states that force extreme strategies. Think of the classic Saturday difference. A driver who is reliably P5–P7 in qualifying is less likely to start in the midfield pack, which cuts lap-one incident exposure, which cuts the need for desperate undercuts. Less desperation means fewer errors, fewer penalties, fewer compromised tyre plans, fewer “we had to try something” Sundays. Simulators reward that because it pushes probability mass into the solid-points region.

You don’t need to predict every grid slot. You need a reasonable spread — qualifying variance, race conversion variance. Run it as a distribution and look at the season totals in the Drivers tool, especially how often each profile lands in the 0–4 point range versus the 10–18 range.

The hidden championship weapon

Zero-point weekends are championship poison. They create a gap normal strong finishes can’t easily close, especially when your rival keeps banking 10–18 points. This is the nonlinear part season simulators expose most clearly: the points you lose finishing P12 instead of P7 are real, but the points you lose by DNFing out of a P4-capable weekend are enormous.

Splitting outcomes into two buckets makes the mechanism obvious.

A consistent profile has a high share of races in the “bankable” bucket — P3 to P8. A peaky profile has more races in the extremes — podiums and low/zero scores. Across enough races, the bankable bucket usually wins unless the peaky profile’s upside is genuinely dominant (a genuinely best-in-class win rate, not a highlight reel).

Compare two profiles with the same average pace but different DNF and variance assumptions in the Drivers tool and check how many simulated seasons are won by each. If the consistent driver wins more seasons despite fewer race wins, you’ve just learned something actionable: championships are frequently decided by floors, not ceilings.

Why “fewer wins” can be the optimal championship profile

Wins are valuable, but they aren’t the only way to outscore rivals. The math often favours drivers who turn strong-but-not-maximum weekends into points with high reliability.

An illustrative example across 24 races (rough numbers for structure, not prediction):

  • Driver A (peakier): 6 wins (150), 6 podiums averaging 15 (90), 6 DNFs, and 6 low-points finishes averaging 2 (12). Total ≈ 252.
  • Driver B (steadier): 1 win (25), 10 podiums averaging 15 (150), 13 strong points finishes averaging 10 (130), zero DNFs. Total ≈ 305.

Driver A feels faster because of the wins. Driver B wins the championship because they keep the season out of the ditch. Your exact numbers will change, but the structure is robust. The simulator isn’t biased toward boring. It’s biased toward compounding.

The practical move is to avoid cherry-picking a single scenario. Build multiple plausible distributions and check whether the conclusion holds. Use the Drivers tool to run sensitivity checks — if the result flips when you nudge DNF rate or qualifying variance a little, your story is fragile and any confident prediction is overreach.

Assumptions that quietly decide the output

Season simulators look precise because they output clean standings tables. The output is only as credible as the assumptions going in — and most disagreements are about hidden inputs, not bad math.

Define what “peak pace” actually means. Is it one-lap pace (qualifying)? Race pace (degradation management)? Situational pace (clean air, tyre warm-up, traffic)? Those are different levers, and a driver can be elite at one and average at another.

Isolate DNF and incident risk. Some of it is driver, some is team, some is environmental. In a calculator you rarely need the exact cause. You need the rate and the variance, and you need to see how the standings respond to each.

Honour the current rules. From 2025 onwards there’s no fastest-lap bonus. That subtly reduces the value of late-race risk-taking for P10-to-P2 runners and removes one bonus pathway for peaky outcomes — which generally increases the relative value of consistent finishing positions, because there’s one fewer mechanism for chaotic single-point swings.

Read results as ranges. A good simulation output is not “Driver X finishes P2.” It’s “Driver X’s most likely band is P2–P4, with a long tail toward P6 if DNFs cluster.” Keep the uncertainty attached to the number and your decisions stay honest.

How to use the Drivers tool for this

If you want real analysis instead of a Twitter thread, compare drivers as distributions rather than single ratings. Open the Drivers tool and treat it like a modelling workspace.

Build two profiles that reflect the debate you’re having. One with higher upside (better peak finishing potential) but more variance (qualifying spread, incident risk, error rate). One with a slightly lower ceiling but a higher floor (tighter variance, fewer zeros, better conversion). Run the season logic repeatedly and look past the mean — check medians, percentile bands, and how often each profile wins the championship despite winning fewer races.

If you take one habit from this: when a simulation likes a consistent driver, it’s usually because your assumptions create too many zero-point weekends for the peaky one. That isn’t a flaw. That’s the model telling you where the championship is actually being won and lost.

The takeaway

Peak pace wins races. Consistency wins seasons — often without looking spectacular week to week. The right way to evaluate that isn’t to pick a narrative. It’s to model finishing distributions, DNF risk, and conversion variance, then let the points system do the scoring.

Run the comparison you care about in the Drivers tool. If the conclusion survives realistic uncertainty, it’s robust. If it flips on a small assumption change, it was never a prediction — it was a preference dressed up in numbers.