Horse RNG is deceptively complex. The surface-level mechanics look simple — pull horses, breed horses, get better horses — but the underlying rate structure produces outcomes that don't match what most guides suggest.
After 200 tracked pulls across breeding and discovery, I have data that clarifies what's actually happening. Some of it confirms the obvious. Some of it doesn't.
How I tracked this
Basic system: record every pull result including whether it was a solo pull or part of a multi, which horse type came out, and the current pity counter. I made no assumptions about the pity system and let the data speak instead.
The 200 pulls covered approximately six weeks of regular play, not a single marathon session. This matters because it accounts for daily bonus effects, which can skew short-session samples.
The tier distribution that's different from community estimates
Community guides estimate the top-tier horses appear at roughly 1-3% base rate. My data puts the observed rate slightly lower — closer to 0.8-1.2% in standard pulls.
That gap matters for strategy: if the rate is 1-3%, some common advice about "just pull until you hit top tier" has reasonable time expectation. At 0.8-1.2%, the math changes. The expected number of pulls to guaranteed top-tier outcome shifts from manageable to genuinely grindy.
Mid-tier horses (what most players consider "good but not great") appear at rates consistent with community estimates — roughly 15-20% in my sample. This seems accurate.
Breeding math that the community underestimates
The breeding pool in Horse RNG matters significantly more than most guides acknowledge. The combination of two horses doesn't produce a weighted average of their tiers — it draws from a pool that's influenced by both inputs.
What this means in practice: breeding two mid-tier horses doesn't give you a 50/50 shot at a mid-tier outcome. The outcome distribution is skewed toward the higher-tier parent when there's a tier gap between the two horses being bred.
I tested this specifically: breeding a top-tier parent with a mid-tier parent consistently produced higher-tier outcomes than breeding two mid-tier parents. This seems obvious in retrospect, but the magnitude of the effect was larger than I expected — roughly a 40% improvement in above-average outcomes.
If you have one top-tier horse and are deciding whether to breed it or keep it for display: breeding it once with a high mid-tier second parent is likely to produce better offspring than the combination you're protecting. The "museum piece" instinct costs more than it looks like.
The pity system: what I observed
I went into this skeptical that the pity system was working as described. The data suggests it does function, but in a way that's more conservative than the community documentation.
In standard pulls: the pity threshold appears to kick in, but the improvement in top-tier rate at pity is smaller than commonly described. The system prevents full drought (I had no top-tier dry streak beyond 70 pulls in my data), but it doesn't dramatically accelerate the rate at pity.
Practical implication: plan for pity as a floor, not a guarantee. If you're at pull 60 expecting a near-certain top-tier outcome, the math doesn't support that. If you're using pity as "I won't go more than 70 pulls without something," the data supports that framing.
The breed-or-keep decision I wish someone had quantified
The most frequent question in Horse RNG communities is when to breed top-tier horses versus keeping them.
Based on breeding rate data: if your top-tier horse is anything below the true top of the tier, breeding it with another strong horse has positive expected value compared to keeping it. The improvement potential outweighs the loss of the current horse.
Exception: verified best-in-class horses (the top 2-3 in the effective ranking) have diminishing returns on breeding because the combination pool doesn't have meaningfully better outcomes to produce. In that case, keeping is correct.
For everything else: breed it.
What I got wrong early that cost me progress
I prioritized pulling volume over breeding quality for the first three weeks. The logic was that more pulls meant more data and more horses.
The problem: pull volume without a clear breeding strategy produces a collection of mid-tier horses that are each individually insufficient to breed top-tier consistently. The efficient path is to identify your current best horse early and commit to breeding chains around it rather than collecting horizontally.
Once I shifted from "get as many different horses as possible" to "maximize the quality of one breeding chain," my average horse quality improved faster despite lower total pull volume.
Quick FAQ
Does daily bonus affect pull rates?
Based on my data: daily bonus appears to improve pull rates modestly. Sessions with daily bonus active produced slightly higher tier outcomes on average — but the effect is small enough that it's a bonus, not a strategy.
Is there any reason to pull singles vs. multis?
My data shows no rate difference. The pity counter accumulates the same way. Pull in whatever increment fits your resource level.
Best use of event currency?
This is horse-specific and version-specific. As a general principle: event currency has a higher expected value than standard currency at the same pull cost, in my experience. Prioritize event pulls when available.
Top comments (0)