Recent message board discussion raised the question of how we sim players breed new horses. Are we emphasizing scratchbreds and breeding fewer babies using existing pedigrees? Are we using more retired mares as more of us gain access to better pedigrees among "our" retired mares? Are there really any noticeable patterns in sim breeding?

Using Zapspride's weekly horse pool and breeding stat reports (available in the Black Type Bugler), this article is an attempt to better understand how simplayers are using the various breeding options. This could not have been done without Zapspride's weekly reports.

Statistical information:

We have stats going back to August 26, 2005. There are four data issues that must be addressed before getting to the conclusions. First, due to duplicate data publication, we have no data for the second week of October, 2005. Second, the data for the third week of September, 2005 may be in error. The values for the horse pool in total increase as they should over the previous week, but the breakdown of new additions exactly matches that of the previous week. I have left the data for that week as-is, but it is likely that the values are somewhat off. Third, while I would have liked to report the breedings by sim-week, that became problematic due to conversion of calendar dates into sim-weeks, matching stats articles with identical headlines to different weeks, etc. Finally, keep in mind that as of the first week of October each year, unbred retired mares become fair game for general use.

In addition, the differences noted on the charts and in text are not necessarily statistically significant. I may go back and run some appropriate tests to determine if and when differences are significant, but for the moment a visual examination of the data is appropriate.

There were a number of ways we could examine breeding trends. The total of new breeds each week does vary, sometimes by a few horses, sometimes by a few hundred horses. To even things out a bit, all graphs shown below consider the percentages of breedings so that weeks could be directly compared.

Overall

The big question, of course, is what is the overall pattern. To display this information, I examined the breakdown of breedings between the following four possible types: scratch bred (creation of a new horse ID), scratch bred (eliminating an existing horse but not using its pedigree), breeding to an active horse (using the pedigree of a mare or sex-changing a he-mare), and breeding to a retired mare. The following table shows the trends since Zapspride began publishing his stats:

That can be a bit confusing to interpret, so let's look at the same data simplified into scratchbreds (either type) vs using an existing pedigree (either type):

We can see that during the period for which data are available, scratchbreds have outnumbered babies based on existing horses at a 3:2 clip. Noticeable changes in the average occured when retired mare "open season" began in October of both 2005 and 2006, and at the new year when retired mares were once again available for new 2yos. scratchbreds, on the other hand, were more common at the end of each year. This may be due to owners breeding 3yos to obtain horses breedable the next year, an accelerated future broodmare program, if you will. Then again, it may be due to other factors, such as players breeding for residency contests that start with the New Year. NWRC has one such contest, other residencies may as well.

2005

Looking at those huge trend charts may make spikes seem more pronounced than they really are, with the data compressed as it is. So, I also looked at data by year. Of course, we have incomplete data for 2005, and 2007 is not even half over. Still, the data by year shows some interesting things. We'll begin with 2005.

During the last quarter of 2005, just under half of the new horses were bred from scratch with new ID numbers created. Given the extra 50 points charged to breed these horses, simplayers were certainly paying a premium. Retired mare breeding greatly increased in comparison to breeding using an active mare or he-mare in October. This is likely attributable to the start of open season on retired mares. Note the spike in scratchbreds created using existing numbers (so-called breedovers) at the end of the year. This may be due to players giving up on horses and clearing stall space in anticipation of the start of the 2006 breeding season. Breedings to active horses were also at their lowest point at the end of the year. Perhaps players chose to wait to breed 2yos instead of firing on 3yos; this is particularly likely if the active horses were males facing sex-changes.

Collapsing the data to show completely new horses versus those created using an existing horse (active or retired), we can see the strong preference of sim players to create brand new horses in 2005. Despite the data primarily covering open season on retired mares, only about 30 percent of new horses were created using either retired or active pedigrees.

2006

Was 2006 any different? It's the only year for which we have complete breeding data, so there should be differences in breeding patterns compared to 2005. 

Notice that spike in scratchbreds created with new ID numbers in early April 2006? That week, of 1068 new babies, 644 fit into that category. That week, incidentally, was the biggest breeding week in terms of new horses for the year, but there was a week in 2007 that saw even more horses bred. I really can't explain why that week was so popular for breedings. The beginning of April is the start of 2yo races carded by the CPU, so perhaps the large number of new babies is a result of the creation of 2yos for the CPU pool. Or maybe one of the sim's magnates opened two dozen new stables and needed to fill them.

Again, let's see what happens when we simplify the data.

Scratchbreds still outnumbered breedings using existing horses, but the gap narrowed a bit. Notable points include the large percentage of new babies bred using existing pedigrees to open the year, and in October at the start of retired mare open season.

2007

So what's happened so far this year?

Scratchbreds with new ID creation are still the most frequent type of horse creation, but it's dropped drastically in comparison to 2005 and 2006. Breeding to retired mares is at about the same level as 2006, but breedings using existing horses has escalated to about quarter of all new babies. The largest number of new babies were bred the last week of January. I suppose that owners primed to breed the first few weeks using their retired mares found themselves out of desirable mares yet still itching for some action, so they went for some scratchbreds. This time, though, they eliminated a lot of horses. perhaps this can be explained by finally giving up on some of last years 2yos. Note the first week of breeding in January: this week's breedings are more properly attributed to 2006, but resetting the data and recreating the charts is an effort I don't wish to go to.

Let's collapse the data again.

In comparison to 2005 and 2006, the number of scratchbreds is nearly equal to the number of horses bred using existing pedigrees. The primary spike, at the far left of the chart, again really belongs to 2006. So far we really don't see any major spikes for the rest of the year, outside some minor changes in the proportions at the end of January.

Overall

So let's look at the trends by year, in a slightly different way this time.

The above chart takes out those weekly spikes and shows us trends year by year. The trend in scratchbreds is toward retiring the bredover horses immediately so that no one else can retrieve it from the horse pool.  Oddly, over time we're seeing a smaller proportion of new breedings with a reused ID. This is despite the small point incentive to give horses back to the CPU pool when breeding by eliminating a horse. Perhaps a greater incentive would reverse that trend.

And let's collapse that data one last time:

Clearly the trend is toward breeding horses using an existing pedigree, be it by the use of a retired mare or by using an active mare or sex-changed male. The scratchbreds still are a slim majority, however. It'll be interesting to see if that pattern continues as the year continues.

In conclusion

So what have we learned? Primarily, that we simplayers are changing how we play the breeding game. We're taking advantage of retired mares, and using our active horses, at rates never before seen in sim breeding. We're also breeding a heck of a lot of horses, over 34 thousand in 2006. We're on pace to far exceed that with just over 15 thousand already bred in 2007. Despite the changing of our ways, we remain on the whole a group of gamers who enjoy creating brand-new horses. Maybe it's the lure of a G1 winner created entirely on our own that entices us to continue breeding by scratch. Maybe it's the prospect of creating sim dynasties that leads us to use those retired mares at record levels. Most likely, we're all looking for that next sim superstar and taking every possible path to getting it.