Bovada Pick'em Poker - old and new analyzed after 1.5 million spins

Was hesitant to make this thread, but after a just horrendous run at Bovada Pick'em poker, I think it needs to be made. How bad was this recent run? 61,100 hands, -13,625 points for a 95.54 return on a 99.45% payback game. Playing .25 and .50 cent machines (at max of course) this equated to -$4,615 gross loss with $1121 in bonuses for a net loss of apx $3500. I know 61,000 hands is not enough to tell much, but now that there is close to 600,000 hands recorded since the payout change, the results are a little disturbing. But we need to disclose the complete data and let people decide for themselves.

Back years ago when I joined Bovada (to play poker after PokerStars left) I noticed they had many 100% match bonuses (with rollover requirements obviously) and Pick'em Poker that had a 99.95% payback. Since this was an advantage play for the player, I figured it was worth trying out, and I also doubted the veracity, so I decided to record every session, including every premium paying hand (RF,SF,Q,FH, Fl, St) and also the result of every live 3 of a kind that was dealt (obviously you cant make quads if they deal quads so that wouldn't count as a live dealt trips). This was a few years worth of playing. Here are the results when the game was set to pay 99.95% :

Total Hands : 874,551
chips played : 4,372,755
chips returned 4,379,864
payback 100.16%

Here are the premium hand totals, with the expected totals in parenthesis:

Royal Flush 1 (2.6)
Straight Flush 31 (22.7)
4 of a kind 362 (371)
Full House 2128 (2060)
Flush 2876 (2789)
Straight 4461 (4430)

Since it is too laborious to record 3 of a kinds, 2 pairs, and pairs, we simply subtract the premium hits from the total and we get what I call "COMMON", which is total payback of non-premium hands. Here is the results for COMMON :

3,466,920 for 79.28 % (79.31)

We ran really hot for the last 150,000 hands, and one of the high-lites was 5 straight flushes in only 12,000 hands (50,000-1 odds of that occurring). One of the cold streaks was 154 dealt 3 of a kinds in a row without a single 4 of a kind conversion (0.15% chance). After all the bonuses, I made a net profit of $19,471. ( I bet a lot of horses and sports too, so it's not like I wasn't giving them fair action to get it back). THEN THEY CHANGED THEY PAYOUT FROM 99.95% to 99.45%.

I had no problem with them changing the payout, because with bonuses i understood it was too much of a liability for them, and I enjoyed the game so I continued to play. Right off the bat, we were cold, I mean, strangely cold, and as I kept playing, it felt a little odd, but I knew the effects of variance, but to make this short and sweet, here are the stats after straights were reduced from 55 to 50, which made it a 99.45% payback:

Total hands 596,978
chips played 2,984,890
chips returned 2,946,029
payback 98.70%

Royal Flush 1 (1.8)
Straight Flush 11 (15.5)
4 of a kind 245 (253)
Full House 1353 (1406)
Flush 1880 (1904)
Straight 2883 (3024)

common 79.50% (79.31)

Here's my concern. To be .75% off expected payout after apx 600,000 hands and to be under expectation for EVERY premium hand makes me wonder is this just variance? Or is it possible that when they changed the payouts, they altered the game in other ways?
None of the numbers are that far out of line (the only one a little borderline being straights, which has only a 0.49% chance of being under 2884), but losing so often for so many sessions, many just getting destroyed, and not getting the occasional long hot streaks we used to get before has me either rightly concerned, or unjustly paranoid.

When we combine the stats of old and new, the game is really close to where it should be. Since straights payouts were changed, I added the 5 points per straight to the reduced payout stats to make the combined stats a uniform 99.95% payout for both.

Total Hands 1,471,529
chips played 7,357,645
chips returned 7,340,388
payback 99.76%

Royal Flush 2 (4.4)
Straight Flush 42 (38)
4 of a kind 607 (624)
Full House 3481 (3467)
Flush 4756 (4693)
Straight 7344 (7455)
common 79.37 (79.31)

So if anyone is left reading this (that hasn't left out of boredom) who is REALLY sharp with numbers, is this .75% shortage and being under expectation for every premium hand after 600,000 hands worthy of concern? Or is this just a natural occurrence with variance, and no big deal at all? Because to be honest, I love Bovada, and have complete trust in them, and have been paid by them enough to recommend them highly, but man, It's not even fun to play this game anymore when you get buried over and over again. They've gotten rid of many of the bonuses you can use, and rarely offer match bonuses like they used to, so when you're way below expectation, it stings a bit.

To be fair, I need to reiterate that even with reduced bonuses, the .75% under performance in payback and the reduced payout change, I am still net profit a couple thousand with the new game, as $19,101 in bonuses still is more than the $16,751 in gross losses. I feel a bit silly questioning a game that I've made a total profit of over 20k with bonuses over the last 3+ years, but real bad runs do something to your mind.

I'd be curious to hear opinions from the sharp math people, and thanks for reading, and I apologize for the length of this thread post.

Pretty sure it's still within variance at that point, but someone better versed in statistics than me may be able to give you more raw numbers.

Sorry about the negative variance lately, if that's what's going on. Really kind of you to track that many hands and share the numbers here. Very interesting, but I'm not a math guy so I have no answers for you.