Action Research - Math Fluency

I think it has been close to seven weeks of implementing daily math games with my students. So far, we have spent a lot of time with the game called addition top-it. I have made variations to the game using dice, dominoes, and number cards. We also have created games using make 10 combinations and double combinations. I am noticing my students are becoming more fluent with a few combinations, but there are combinations that many struggle with and seem to slow them down. Those seem to be the bigger single digit numbers such as adding six through nine together with other larger numbers.
This past week I noticed my struggling students making progress with their combinations. Before they were using their fingers on everything. However, I noticed they were able to recall some double combinations, make 10 combinations and a few others with little or no hesitation. This is exciting to see. Although I have seen considerable progress in this area with many students, there written and oral assessments are not showing the growth I would have expected. I am wondering if it is still too early to tell. Students have been making progress on their timed tests, but I am wondering what is holding students back from showing more progress on the assessments?
I have been comparing written vs. oral timed tests and trying to figure out if I can draw any conclusions from them. One thing I have noticed is in regards to students who seem to be more fluent than their peers. Those students are making larger gains on their oral than that of the written timed tests.
The information below shows the breakdown of how students scored on their oral vs. written timed tests. The first breakdown shows students who showed a decrease to a plus 2 on the oral compared to written. The next one shows students who had plus 3 or plus 4. The last column shows those students who had plus 5 or more facts on the oral vs. written.
Tests       neg to +2           +3/+4            +5 (+)
Test 1          8  44%           4  22%           6  33%
Test 2          9  50%           2  11%           7  39%
Test 3         14 82%           0   0 %           3  18%
Test 4          8  50%           5  31%           3  19%

I think taking a test orally or in written form really does not matter unless you are more fluent with facts. My students who are more fluent with their facts are the ones who are making plus five or more facts correct on their oral timed tests. I wonder if I will continue to see the same results when comparing written vs. oral timed tests. If students are becoming more fluent, shouldn't I see an increase in the tests and have a larger amount of students fall into the plus five category. I am curious to find out how this information will change in the coming weeks.