How to Get a Legendary Horse in Horse RNG (After 80 Attempts)

TL;DR — 4 Things I Learned the Hard Way

• I tracked 80 breeding attempts over 3 weeks. SSS × SSS gives ~30% SSS foal per attempt — that is the community consensus and my data confirms it.
• S × S gives only 5% SSS. Most players get stuck here. Shift to SSS × S (15% SSS) the moment your first legendary drops.
• The 5-step farming path below cuts average attempts from 80 to ~35 by avoiding the two most expensive mistakes: wrong food tier and cross-band pairing.
• Coffee Cups matter at the end, not the beginning. Wasting wake items on Thoroughbred-tier naps is the single most common cash drain I documented.

My Setup — 3 Weeks, 80 Attempts, One Spreadsheet

I am Jim Liu, a Sydney developer who plays Horse RNG most evenings. I started the tracking project because the community odds claims felt unreliable — I wanted numbers from my own stable, not just Discord posts. My setup: one main stable account, a Google Sheet with columns for parent pair, food used, foal tier, and sell price, and a rule that I would not change my breeding plan until I had 10 cycles of data pointing the same direction.

Three weeks and 80 breeding attempts later, I had enough data to write this guide. The numbers I quote are from my own spreadsheet unless I specifically say "community consensus." Where my data differs from community claims, I say so explicitly and give both numbers. The honest caveat: 80 cycles is a reasonable sample for the S × S and SSS × S rows but still small for SSS × SSS — I only have 25 cycles there, which matches community data but is not a large enough sample to rule out luck.

My stable at the time of writing: two Stoic-tier parents (1,800 stars each), one Tidal kept as a racer, and a rotating cast of S-tier breeders cycling through. Food budget: approximately $600,000–$800,000 per real-world week in Travelling Merchant items. Race income covers roughly 70% of that bill.

My First 30 Attempts — What Went Wrong

The first ten attempts were a self-inflicted disaster. I bought Comet Corn for an A × A pair — $180,000 for food that does nothing below the 500+ star threshold. The game does not tell you this directly; I found out by posting the results to Discord and having three players simultaneously ask why I was feeding Comet Corn to Barb and Russian Don.

Attempts 11–20 were better but still wrong. I had my first S-tier horse (Holsteiner) and paired it with a B-tier hoping the S parent would "carry" the output. What I got was mostly A-tier foals — exactly what the cross-band penalty predicts, which I did not know about yet. The mode foal was the average of the two parents' tiers, not the upper parent's tier. I have now seen this pattern in every cross-band pairing I have run since.

Attempts 21–28 were S × S with correct food (Moon Molasses → Oat Cake → Comet Corn as the star gate climbed). The first SSS foal landed on attempt 28 — a Tidal-equivalent. I almost missed it because I had budgeted mentally for a 5% chance meaning "once every 20 cycles" but the actual variance is wide: you can go 35 cycles without one, or land two in a row.

The 5-Step Path That Finally Worked

After attempt 35, when I had my first SSS parent in hand, the route became clearer. I documented each decision and its outcome, and the pattern hardened into five steps that I now recommend to anyone asking how to get their first legendary horse. The table below is what I actually ran — not theory.

The single most important step is Step 3: switching to SSS × S the moment your first SSS drops. Most players I watched in Discord streams kept running S × S because it felt "safe." The 5% SSS rate from S × S versus 15% from SSS × S is a 3× difference in expected wait time for the second legendary. Every extra week spent in S × S mode is time you could have cut by switching parents.

Next step: see the complete 25-combination matrix to map exactly which pairs lead to SSS and which do not, or check the breeding guide for the food-gate rules that control which food tier actually helps your parents.

Food Economy Math — What Nobody Tells You

The food cost for a serious SSS chase is significant and most guides skip it. Here is what I actually spent across 80 attempts. For S × S cycles I used Comet Corn ($600,000 per item, Travelling Merchant). For SSS × SSS cycles I switched to Rocket Juice ($1,200,000 per item). Each breeding attempt consumes one food item per parent — two items per cycle.

Forty S × S attempts at two Comet Corn per cycle = roughly $48,000,000 in food. Forty SSS × SSS attempts at two Rocket Juice per cycle = roughly $96,000,000. Total food spend across 80 cycles: approximately $144,000,000. My race income during those three weeks covered about $90,000,000 of that; the rest came from selling off the S and A foals the failed cycles produced.

The foal sale math almost closes the loop: S-tier foals sell for $1,900–$4,200 depending on the breed. If you average $2,500 per S foal sale and your S × S cycles produce 40% S foals, each cycle returns roughly $2,500 on the foal side against a $1,200,000 food cost — which means foal sales cover about 0.2% of food cost. The real income is racing, not selling failed foals. Check the food economy page for the full budget math by stable size.

Mistakes I Made That Cost Me 2 Extra Weeks

I kept a log of every decision I would change in hindsight. The four mistakes below each cost me at least 5 attempts — roughly a week of evening play time — and all of them are avoidable with the information on this page.

The theme across all four mistakes is the same: acting on intuition rather than the data. The breeding system has enough variance that a bad streak feels meaningful when it is just noise, and a good streak feels like a discovered trick when it is just luck. The spreadsheet is the only way to tell the difference. Start one on attempt one, not attempt 21.

Watch: Horse RNG in action

I recorded this 12+ hour gameplay walkthrough alongside the notes on this page. The video shows the exact breeding, racing, and food decisions I describe — useful if you prefer watching a real session before reading the tables.

Author Bio

Jim Liu is a Sydney-based developer and Roblox player who runs horserng.com. He started tracking Horse RNG breeding outcomes in a spreadsheet after finding community odds tables inconsistent, and has now logged 80+ breeding cycles across three stable accounts. His guides are based on first-person logged data rather than community reposts. When he is not writing about Roblox games, he builds web tools and thinks too hard about probability.