per game starting pitchers, HAL override?

Had something come up that I had never seen before. I am using the per game starting pitchers to match up my starters against opponents. I have 6 possible starters and one is a "star" rated which allows me some flexibility. Twice (game 10 and game 15) Hal overrode my per game selection and started an RP only guy (Danforth).

I am 100% certain I had an eligible pitcher selected for those games and of course, it could not have been Danforth as the menu won't let you select a relief only guy to start. I have not had any pitcher injuries which would have necessitated this.
Danforth pitched in relief in game 14, the day before HAL had him start. He also pitched in game 8 and no "s" rated pitcher pitched in relief in game 9. It might make sense if Mike Marshall or Ken Clay (the two who I think got scratched) had pitched a ton of innings in relief a game or two before, but that was not the case in either game (I'm not really sure if that is how the logic works, but the point is it would not be in play here looking at the boxscores).

Any explanations on this?

My first guess was that one of your S/R types threw a long relief outing 2 days prior. But you said that wasn't the case--though how are you certain in Bullpen v3?

I am not positive you can be certain, because the PCF algorithm is not transparent and neither is the suggested pitch count determination for starters and relievers, which I think is like a rolling average approach--I have no idea, but it is not based strictly on innings nor is it always the same.

So what I mean is, a few scenarios come to mind:
1. A S/R type heavily exceeded suggested pitch count over multiple appearances days prior to your desired start assignment and did not have enough available suggested pitch count to cover a start

2. Danforth had a higher suggested pitch count than either of your S/R types. This could be possible if they were S5.

3. A bug. The game engine tweaks to the online version of the code is not baselined anywhere else to my knowledge--it is unique to SOM365. While performance over the years suggests SOM has done (and invested) multiple orders of magnitude better than TSN did at hardware and software CM, it is not impossible to have a rare bug that evaded a test case, especially if changes are made via scripts and test case artifacts are not as formal.

That sounds like as good as explanation as any. It is probably what you describe tho--whatever the look back period is Hal determined Danforth had more innings available, the ironic thing is that Danforth seems to be getting overused in comparison to my S pitchers. Danforth is arguably better than the guys that would have started so I guess I'm not complaining (although he did lose both games it wasn't because of him per se). Maybe he's just competing for a starting spot.

Any chance that fatigue rating would be a factor here? I was checking my settings and I have several of my S rated pitcheres to get yanked at F6 and Danforth I left blank.

Hmmm...I wouldn't think so, because the fatigue setting (ie Do Not relieve before...) is only in play after the game starts. However, we have no way of knowing...scenario 4...

This is a long shot, but are Marshall and Clay set as your closers? If Closer Usage is set on Aggressive, it seems that Hal doesn't like to use them in any other roles.

I have no settings for either aside from fatigue limits. Additionally, Red Faber was more than rested and not assigned to a game or made any appearances in the same stretch. It really makes no sense other than there must be some curious algorithm that privileged Danforth over Clay, Marshall and Faber.